Skeptics Guide to the Universe Forums

General Discussions => Tech Talk => Topic started by: Teslaborscht on April 13, 2015, 09:37:39 AM

Title: Self-Driving Cars
Post by: Teslaborscht on April 13, 2015, 09:37:39 AM
Rather than have the general public adopt self-driving cars in the next decade or so, I suspect that those who have the most to gain will adopt self-driving cars: taxi, bus, and truck companies. Car companies will push (and get government handouts for) fleet sales first. So it is quite likely that we'll be riding in self-driving taxis (with a driver) before we see consumer-level self-driving cars.
Title: Re: Self-Driving Cars
Post by: Andrew Clunn on April 13, 2015, 09:43:05 AM
I doubt it.  Liability is a huge issue, and because of the autonomous nature of vehicles, a bottom up approach works just as well.  I think this is going to more likely follow the luxury item curve, where wealthy (aging) individuals will buy in first.  Well aside from the military, which is already on this bandwagon.
Title: Re: Self-Driving Cars
Post by: The Latinist on April 13, 2015, 10:21:43 AM
I understand that all Tesla vehicles will soon come with a self-driving feature for highways only.  I've also seen advertisements for vehicles that will automatically brake to avoid collisions.  I imagine such things will soon be common at least in higher-end vehicles.  Over time, it will trickle down.  I imagine that will be the sort of piecemeal approach we'll see, rather than a system-wide rollout like the rogues seemed to be envisioning.
Title: Re: Self-Driving Cars
Post by: superdave on April 13, 2015, 11:31:15 AM
I doubt it.  Liability is a huge issue, and because of the autonomous nature of vehicles, a bottom up approach works just as well.  I think this is going to more likely follow the luxury item curve, where wealthy (aging) individuals will buy in first.  Well aside from the military, which is already on this bandwagon.

This is pretty much true already.  Certain Mercedes cars can already theoretically drive autonomously on the highway but the software requires the driver can't take his hands off the wheel for my than 5 seconds or it shuts off. 
Title: Re: Self-Driving Cars
Post by: frothy on April 13, 2015, 12:09:59 PM
What do you mean by "Liability is a huge issue"?
Title: Re: Self-Driving Cars
Post by: TheIrreverend on April 13, 2015, 12:13:52 PM
If a self-driving car gets in an accident, who is to blame?  The person behind the wheel?  The software developer?  The vehicle manufacturer?  Until those problems are solved, almost definitely by legislative intervention, widescale adoption is unlikely.
Title: Re: Self-Driving Cars
Post by: Andrew Clunn on April 13, 2015, 12:15:25 PM
What do you mean by "Liability is a huge issue"?

There's no real legal notion of what to do in situations where self driving cars get into accidents.  Large companies are going to wait for personal cases to flesh that out before putting themselves on the hook.  This could be mitigated by insurance companies, but they'll likely want to test the waters at smaller scales to get data for determining risk and costs before scaling up for large corporate policies.

EDIT - Yeah, what he said above.
Title: Re: Self-Driving Cars
Post by: PANTS! on April 13, 2015, 12:48:54 PM
If a self-driving car gets in an accident, who is to blame?  The person behind the wheel?  The software developer?  The vehicle manufacturer?  Until those problems are solved, almost definitely by legislative intervention, widescale adoption is unlikely.

Who get sued when a train derails?  When the solar powered monorail speeds out of control?  When an airplane crashes?

I mean I get that it hasn't been solved for cars, but I do not think it is too insurmountable.  There are transportation models we can draw upon now.  Seems to me like the adjusters and the police will need some training, and some new laws would need to be drafted.  But hell, if anything our current system of determining liability is far far more ambiguous when determining fault.  Not to mention there will be FAR fewer crashes, so in general each crash will be investigated more thoroughly.
Title: Re: Self-Driving Cars
Post by: frothy on April 13, 2015, 12:53:46 PM
But why is this a huge issue?

Just make some self driving cars, eventually an accident will happen, the courts will figure out how to apportion blame, and then the world continues. I don't see how it's a roadblock to progress.
Title: Re: Self-Driving Cars
Post by: amysrevenge on April 13, 2015, 01:00:22 PM
But why is this a huge issue?

Just make some self driving cars, eventually an accident will happen, the courts will figure out how to apportion blame, and then the world continues. I don't see how it's a roadblock to progress.

Well, if one anticipated apportioning of blame is "manufacturer sued for hundreds of millions of dollars" I can understand some reluctance to become one.
Title: Re: Self-Driving Cars
Post by: PANTS! on April 13, 2015, 01:07:27 PM
But why is this a huge issue?

Just make some self driving cars, eventually an accident will happen, the courts will figure out how to apportion blame, and then the world continues. I don't see how it's a roadblock to progress.

Well, if one anticipated apportioning of blame is "manufacturer sued for hundreds of millions of dollars" I can understand some reluctance to become one.

I think that happens quite often even in today's climate.
Title: Re: Self-Driving Cars
Post by: frothy on April 13, 2015, 01:11:28 PM
Couldn't you use that argument for any product?
Title: Re: Self-Driving Cars
Post by: Andrew Clunn on April 13, 2015, 01:18:22 PM
Here's the thing.  If the first accident that results in a lawsuit is with a private car owner, then if the owner of the self driving vehicle ends up on the hook, it will be for far less since they're just some person.  If the first lawsuit is against a large company with several of these, then it's likely to come back higher, and the first court case sets at lot for future ones.  This is what we libertarians calls, impeding progress through fear of litigation and anti-corporate sentiment, or "fuck lawyers" for short.
Title: Re: Self-Driving Cars
Post by: amysrevenge on April 13, 2015, 01:20:28 PM
Couldn't you use that argument for any product?
You could.  It boils down to some suit's cost-benefit analysis.  Some projects have a higher potential return and/or lower anticipated risk than others. 

You could argue that the estimation of either is being done incorrectly, but that's what it's about.
Title: Re: Self-Driving Cars
Post by: The Latinist on April 13, 2015, 01:35:10 PM
I have another concern with self-driving cars: icy and snowy roads.  I can readily believe that current technology is capable of driving a car on a dry road, but driving on ice or in snow is a significantly different task, one which requires a fair amount of counter-intuitive action.  I have my doubts that the sensor technology used in such cars is capable of detecting and responding to winter road conditions.
Title: Re: Self-Driving Cars
Post by: PANTS! on April 13, 2015, 01:39:59 PM
I have another concern with self-driving cars: icy and snowy roads.  I can readily believe that current technology is capable of driving a car on a dry road, but driving on ice or in snow is a significantly different task, one which requires a fair amount of counter-intuitive action.  I have my doubts that the sensor technology used in such cars is capable of detecting and responding to winter road conditions.

Hmmm...  Interesting objection.  Wet roads too might be an issue.  However, there most certainly are traction control technologies available.  How / if they are integrated with the self-driving part is unknown to me.
Title: Re: Self-Driving Cars
Post by: amysrevenge on April 13, 2015, 01:45:40 PM
I have another concern with self-driving cars: icy and snowy roads.  I can readily believe that current technology is capable of driving a car on a dry road, but driving on ice or in snow is a significantly different task, one which requires a fair amount of counter-intuitive action.  I have my doubts that the sensor technology used in such cars is capable of detecting and responding to winter road conditions.

Heh.  I've been assuming all along that when they say "self driving cars" they really mean "self driving cars in California".  With California as a stand-in for places where you can, for instance, count on having lines stay on the road without being obscured for months at a time by snow and ice, and scraped/corroded off by various snow removal procedures (I still remember marveling at the sticky-outie reflective thingies incorporated into the lane dividing lines when we moved to Seattle).  Places where you can count on 350+ days of driving impacted by nothing more onerous than a bit of rain in any given year.
Title: Re: Self-Driving Cars
Post by: The Latinist on April 14, 2015, 11:43:02 AM
Hmmm...  Interesting objection.  Wet roads too might be an issue.  However, there most certainly are traction control technologies available.  How / if they are integrated with the self-driving part is unknown to me.

In the winter months, I frequently drive on roads where there are no visible lines or pavement, and even no distinct edge to the road.  I also frequently drive when visibility is in the 20-30' range.  I have no idea how self-driving car would even keep its lane in such circumstances.  As for traction control, I have it; but it is not a panacea.  Indeed, there are times when one must disable it to drive effectively in snow or on ice.

How does a self-driving car in white-out conditions know what do do when its wheels start spinning?  Should it slow the engine, should it brake, or should it turn the wheels sharply into the spin and accelerate?  I don't think it does.
Title: Re: Self-Driving Cars
Post by: amysrevenge on April 14, 2015, 11:53:27 AM
I don't think it's an insurmountable problem - I just think that self-driving in normal* conditions in not-California is at least an order of magnitude more difficult than self-driving in even abnormal conditions in California.

You just need the engineers to spend waaaay more winters in Canada or Wisconsin or what have you than they are likely to want to.  :)

*Where "normal" includes conditions such as The Latinist mentions above - these are not abnormal conditions, these are expected, regular, long-duration conditions.
Title: Re: Self-Driving Cars
Post by: superdave on April 14, 2015, 12:25:56 PM
It's not like humans do so well in winter conditions either.  Why couldn't a computer do as good or better?
Title: Re: Self-Driving Cars
Post by: amysrevenge on April 14, 2015, 12:34:28 PM
The problem is one of detection and synthesis of external conditions.  The kind of things people are traditionally good at and computers haven't been as good at.

It's easy to, say, figure out where you are and how fast you're going on a proper road with visible lane demarcations and distinct edges and long visibility.  I got close to the math of it in an image processing graduate course a few years ago (it's a very similar problem to adding a virtual first down line to the TV display of a football game - we did work out that problem and it was pretty fun).

It is a lot harder when the visual cues are irregular and transient.  If you can add in deeper "visual" data with additional unhuman senses or whatever (radar, for instance), you can mitigate it. 

Order of magnitude stuff, not impossible stuff.  If you can solve it, then you're gonna do it better than humans (just like with California conditions).
Title: Re: Self-Driving Cars
Post by: The Latinist on April 14, 2015, 01:30:10 PM
It's not like humans do so well in winter conditions either.  Why couldn't a computer do as good or better?

Skilled winter drivers do quite well, thank you.  I'm not saying it's an impossible problem but, as amysrevenge said, it is orders of magnitude more difficult than driving a car on dry roads.  We're looking at cars being able to drive in ideal conditions quite soon, but I would not be at all surprised if the ability for a car to self-drive in winter conditions were decades away.
Title: Re: Self-Driving Cars
Post by: kvuo75 on April 14, 2015, 07:16:25 PM
Hmmm...  Interesting objection.  Wet roads too might be an issue.  However, there most certainly are traction control technologies available.  How / if they are integrated with the self-driving part is unknown to me.

In the winter months, I frequently drive on roads where there are no visible lines or pavement, and even no distinct edge to the road.  I also frequently drive when visibility is in the 20-30' range.  I have no idea how self-driving car would even keep its lane in such circumstances.  As for traction control, I have it; but it is not a panacea.  Indeed, there are times when one must disable it to drive effectively in snow or on ice.

How does a self-driving car in white-out conditions know what do do when its wheels start spinning?  Should it slow the engine, should it brake, or should it turn the wheels sharply into the spin and accelerate?  I don't think it does.

i'm pretty sure it can be figured out..

driving a car is not even rocket science.. and we can do rocket science.

Title: Re: Self-Driving Cars
Post by: The Latinist on April 14, 2015, 09:05:28 PM
Let me once again make very clear that I'm not saying cars that self-drive on winter roads can't be made.  I'm just not, and I wish people would stop responding as if I were.  What I'm saying is that I don't think it's anywhere near to being realized.  We already have cars driving independently on dry roads, and the tech's likely to come to consumer vehicles very soon; but I think we're at least 10 years and perhaps quite a bit longer away from a car that can drive in snow and ice.

Remember, in order for a car to be able to handle snow and ice effectively we're going to have to be able to create a computer abstraction of all the factors that go into the problem, something that to my knowledge nobody is even working on. Indeed, I don't think that we even have a basic theoretical understanding of the problems involved. I think everyone here except Ambious is significantly underestimating the difficulty of the problem.
Title: Re: Self-Driving Cars
Post by: kvuo75 on April 14, 2015, 10:19:02 PM
i think i understand it.. you're basically talking about different amounts of friction between tire and road surface.

even rudimentary ABS's already do that.. so we've pretty much got the stopping figured out.

the vehicle stability control system in my inexpensive car already uses vehicle speed, yaw, steering input, etc. to simply apply braking on one side or the other to correct for over/understeer (which is caused by human error to begin with).  i could only imagine if it took control of steering itself, how much more effective that would be.

as for being able to see a roadway -- i REALLY dont trust a person.. computers can use visibile light, ir, microwave to see roads AND traffic better than we can.. not to mention GPS and INS for positioning.

you're probably right that it's probably 10+ years away from reality, but nevertheless.  i think the technology already exists and i doubt ice and snow is much more a challenge than driving on wet roads...


edit:

we're also assuming the current self-driving cars can't do it already. why are we assuming that?
Title: Re: Self-Driving Cars
Post by: arthwollipot on April 15, 2015, 02:20:28 AM
I think an appropriately-programmed autonomous car will handle ice and snow better than most people.
Title: Re: Self-Driving Cars
Post by: brilligtove on April 15, 2015, 08:54:33 AM
This article (http://www.autoconnectedcar.com/2015/03/tesla-mercedes-benz-delphi-audi-winning-self-driving-race/) mentions the radar and lasers used to see in snow and rain, but I could not find anything indicating how the most recent cross country test drive handled wet or snowy conditions.

If a self-driving car gets in an accident, who is to blame?  The person behind the wheel?  The software developer?  The vehicle manufacturer?  Until those problems are solved, almost definitely by legislative intervention, widescale adoption is unlikely.

Who get sued when a train derails?  When the solar powered monorail speeds out of control?  When an airplane crashes?

I mean I get that it hasn't been solved for cars, but I do not think it is too insurmountable.  There are transportation models we can draw upon now.  Seems to me like the adjusters and the police will need some training, and some new laws would need to be drafted.  But hell, if anything our current system of determining liability is far far more ambiguous when determining fault.  Not to mention there will be FAR fewer crashes, so in general each crash will be investigated more thoroughly.

I don't think these are equivalent to private drivers driving their privately owned vehicles. Or rather private drivers being driven in their privately owned vehicles. Can I be sued if my car kills your dog? Can the software developers be sued? Their employer?

I own the vehicle but I have no responsibility for how it behaves. How can I be sued when its behavior is murderous? If I owned a plane and hired a pilot to fly it and the pilot did something insane like crashed into a mountainside am I responsible for the pilots actions? In the case of an airline, the answer may be yes. Not because I have direct control of the pilot actions, but because am airline is regulated to take extraordinary measures to ensure that the pilots behaviour will remain in a safe zone. as the passenger and owner in an autonomous vehicle, I have no way of conforming to that kind of regulation so I cannot be held responsible. Except I chose to buy that car, just like I chose to get that dog, and when the dog bites a kid that's the end of that dog and its my fault. Of course in the case of a dog I have significantly more control over the "programming" of the animal.

I guess what I'm saying is I think we are entering new territory in terms of assigning blame and responsibility. It is going to be interesting.
Title: Re: Self-Driving Cars
Post by: The Latinist on April 15, 2015, 10:02:04 AM
I think an appropriately-programmed autonomous car will handle ice and snow better than most people.

Sure.  And an appropriately-programmed computer will be able to write poetry better than most people.  The question is, how close are we to being able to create that programming?  Right now we have programming that can turn sports stats into news articles, but I don't think we're anywhere near a convincing poetry-writing computer.  I think the difference between driving on dry roads and driving on snowy and icy roads is on the same order of magnitude.
Title: Re: Self-Driving Cars
Post by: Johnny Slick on April 15, 2015, 02:46:06 PM
I sincerely doubt that an appropriately programmed computer will be able to write poetry well at all because that requires lateral thinking of the variety that computers are not good at.

With the self-driving cars issue, I can't help but think of the classic XKCD comic...

(http://imgs.xkcd.com/comics/tasks.png)

Driving strikes me as something that's stealthy-hard. You have to account for lots and lots of variables, and account for them so quickly that you don't consciously know that you accounted for them until you're after the fact. Figuring out what all of them are is going to be tough and I agree with TheLatinist that except in very specific instances (for instance, if you were to create a stretch of road that was fully automated for all vehicles involved), we're not particularly close to the point to where we've accounted for enough of said variables that it'll outdo humanity.
Title: Re: Self-Driving Cars
Post by: superdave on April 15, 2015, 05:06:05 PM
I sincerely doubt that an appropriately programmed computer will be able to write poetry well at all because that requires lateral thinking of the variety that computers are not good at.

With the self-driving cars issue, I can't help but think of the classic XKCD comic...

(http://imgs.xkcd.com/comics/tasks.png)

Driving strikes me as something that's stealthy-hard. You have to account for lots and lots of variables, and account for them so quickly that you don't consciously know that you accounted for them until you're after the fact. Figuring out what all of them are is going to be tough and I agree with TheLatinist that except in very specific instances (for instance, if you were to create a stretch of road that was fully automated for all vehicles involved), we're not particularly close to the point to where we've accounted for enough of said variables that it'll outdo humanity.

The comic isn't wrong, but we really have come a long way in this field.  The IEEE has an entire weekly newsletter dedicated to this space.
http://spectrum.ieee.org/blog/cars-that-think (http://spectrum.ieee.org/blog/cars-that-think)
Title: Re: Self-Driving Cars
Post by: brilligtove on April 15, 2015, 05:12:39 PM
I sincerely doubt that an appropriately programmed computer will be able to write poetry well at all because that requires lateral thinking of the variety that computers are not good at.

With the self-driving cars issue, I can't help but think of the classic XKCD comic...

(http://imgs.xkcd.com/comics/tasks.png)

Driving strikes me as something that's stealthy-hard. You have to account for lots and lots of variables, and account for them so quickly that you don't consciously know that you accounted for them until you're after the fact. Figuring out what all of them are is going to be tough and I agree with TheLatinist that except in very specific instances (for instance, if you were to create a stretch of road that was fully automated for all vehicles involved), we're not particularly close to the point to where we've accounted for enough of said variables that it'll outdo humanity.

The comic isn't wrong, but we really have come a long way in this field.  The IEEE has an entire weekly newsletter dedicated to this space.
http://spectrum.ieee.org/blog/cars-that-think (http://spectrum.ieee.org/blog/cars-that-think)
You may be right, Johnny Slick. Then again, I can do that search for a bird in my Google Photos right now, and it did come back with a picture of a bird I took last month. The XKCD comic is #1425 (http://xkcd.com/1425/) from six months ago, BTW.

I'm sure that there are many specific conditions that will need to be dealt with using specific algorithms or tweaks. At the same time, the car probably doesn't need to use the kind of algorithms that humans use for a lot of that. It seems like the folks building these things are getting pretty good at building a system that can interpret realtime sensory data, and from an array of senses that are much broader and better than our own.

In any case, I'm in the camp that thinks these will become part of our culture the way mobile phones did. The adoption curve may be slower presuming that people keep cars for longer than they keep phones. I'm actually more concerned about hacking and emergent behaviours than I am about whether individual cars are safe drivers. 
Title: Re: Self-Driving Cars
Post by: SnarlPatrick on April 15, 2015, 08:52:09 PM
This can't come soon enough for me. I find driving to be enormously stressful, especially at night, when my vision is not particularly good. I don't normally do this, but "trigger warning, car accident"

When I was a teenager, I was the passenger in a car, when we struck a old women who was wandering in the road, in a dementia induced fog. We arrived at the top of a hill, and there she was right in front of us. I saw her with just enough to time scream "STOP", the driver managed to break, but only very slightly. Her head struck the windshield directly in front of my face, with a sound I will not describe, and sent cracks throughout the glass, she was thrown several meters down the road. 

I heard someone screaming, and then realized it was me. For a few moments, I simply stared ahead, at the skull shaped dent in the glass, the pieces of hair and gore, maybe 6 inches from my face. Then everything went blank, I completely failed to provide first aid, and ran several blocks away, and curled up into a quivering dissociative ball until a police officer found me.

Fortunately, there was a bus stop nearbye, with a nurse present, who was able to provide first aid, and she lived. But I see this image over and over and over again when I am at the wheel. If I had a self-driving car, I would go many more places than I do now.
Title: Re: Self-Driving Cars
Post by: Kay-Ole on April 17, 2015, 06:58:11 PM
I wrote this in the thread for episode #509, but it's relevant for this thread too, so I will do a copy/paste.

"There is one important point about driverless cars that has yet to be discussed in the SGU. It was brought to my attention by none other than Jeremy Clarkson, who usually has the scientific sophistication of an uneducated orangutang.

Anyway, the ethical questions that will be raised in programming these cars is quite substantial. In the unlikely event that five children suddenly finds their way into the path of the car, and all other measures for noticing them in time has failed; what will the car "choose" to do? Say the only option for saving these children is swaying off the road, but this will certainly lead to the death of the person in the car. This is of course a variation over the old, well known thought experiments on human ethics, but the fact of the matter is someone will actually have to program the cars computers for this eventuality."
Title: Re: Self-Driving Cars
Post by: kvuo75 on April 17, 2015, 09:26:05 PM
I wrote this in the thread for episode #509, but it's relevant for this thread too, so I will do a copy/paste.

"There is one important point about driverless cars that has yet to be discussed in the SGU. It was brought to my attention by none other than Jeremy Clarkson, who usually has the scientific sophistication of an uneducated orangutang.

Anyway, the ethical questions that will be raised in programming these cars is quite substantial. In the unlikely event that five children suddenly finds their way into the path of the car, and all other measures for noticing them in time has failed; what will the car "choose" to do? Say the only option for saving these children is swaying off the road, but this will certainly lead to the death of the person in the car. This is of course a variation over the old, well known thought experiments on human ethics, but the fact of the matter is someone will actually have to program the cars computers for this eventuality."


i think as current law usually allows. roads are dangerous..  the pedestrians get the benefit of the doubt, unless they are doing something unlawful.

eg. if you are in a car and someone is standing in the road, you have to stop.  if someone jumps in front of your car, they cant expect to not get hit. 

you could just program it to behave like a typical reasonable motorist, and plow into the pedestrians.. status quo.

i might not be thinking it through enough, but i dont see it as a big thing.  a computer could certainly make whatever decision it needed to quicker, so already we're ahead of a human driver.






Title: Re: Self-Driving Cars
Post by: SnarlPatrick on April 20, 2015, 03:51:45 AM
Yeah idk.... i think the idea that the car sensors would be so precise, that they could weigh the relative risks of different actions... i mean... you'd need sensors that counted the number of occupants of cars, and then weighed the relative value of people, injuries to pedestrians vs occupants of vehicles with restraints? I think thats hugely optimistic for the near future. People don't do a great job of this. Computers should be able to do about as well. There might well be room for an ethicist in the design process, but I'm not sure it is going to be that difficult a problem, just because we're going to have limited functionality at first. When in doubt... err in the the side of utilitarianism i suppose?
Title: Re: Self-Driving Cars
Post by: Kay-Ole on April 21, 2015, 08:36:11 PM
Well, the computers will be as good at this as the humans that programme them. And that is the point. There is no computer code for "Asses human value correctly" that will make computers magically do the right thing. But these issues will need to be adressed when the computer for the cars behaviour are going to be programmed.

There are allready driverless cars to some degree, and I'm sure there are plenty of prototypes of computers allready out there, so it would be interesting to know how these issues are adressed.
Title: Re: Self-Driving Cars
Post by: brilligtove on April 22, 2015, 09:11:05 AM
Yeah idk.... i think the idea that the car sensors would be so precise, that they could weigh the relative risks of different actions... i mean... you'd need sensors that counted the number of occupants of cars, and then weighed the relative value of people, injuries to pedestrians vs occupants of vehicles with restraints? I think thats hugely optimistic for the near future. People don't do a great job of this. Computers should be able to do about as well. There might well be room for an ethicist in the design process, but I'm not sure it is going to be that difficult a problem, just because we're going to have limited functionality at first. When in doubt... err in the the side of utilitarianism i suppose?

It's not that the sensors will be so precise. Heck, our human sensors are crap wrapped up in a cloudy bag. The computers don't have to make great moral judgements either. Accidents happen when
The control systems in driverless cars will be far less prone to driver error in part because of better sensors than humans have (such as radar, lidar, and IR cameras), and in part because they are much less likely to be distracted or overwhelmed.

There will always be environmental and vehicle conditions that a car can't handle, whether driven by a human or by itself. Some of these can be posed as moral questions with absolute answers. This is not a fair way to go about it, since there is no right answer to program into the car. We use avoidance and mitigation strategies as best we can; asking 'hit the child or hit the adult' can't be addressed, but minimize acceleration to anything with human body temperature and size - that could be taught.

Well, the computers will be as good at this as the humans that programme them. And that is the point. There is no computer code for "Asses human value correctly" that will make computers magically do the right thing. But these issues will need to be adressed when the computer for the cars behaviour are going to be programmed.

There are allready driverless cars to some degree, and I'm sure there are plenty of prototypes of computers allready out there, so it would be interesting to know how these issues are adressed.

There is no factual basis to assert that a computer can only be as good as it programmers at some cognitive task. Deep Blue. Google search. Watson. Humans make tools that are better than us in specific domains all the time. (Your typewriter had better handwriting than you.) It's part of what makes us human.
Title: Re: Self-Driving Cars
Post by: kvuo75 on April 22, 2015, 09:56:48 AM
There is no computer code for "Asses human value correctly" that will make computers magically do the right thing.

people don't magically do the right thing, either. in fact, they do the exact wrong thing, constantly.
Title: Re: Self-Driving Cars
Post by: brilligtove on April 22, 2015, 10:33:03 AM
There is no computer code for "Asses human value correctly" that will make computers magically do the right thing.

people don't magically do the right thing, either. in fact, we they do the exact wrong thing, constantly and consistently.

FTFY. ;) Good point to keep in mind.
Title: Re: Self-Driving Cars
Post by: PANTS! on April 22, 2015, 11:21:17 AM
Yeah - I think this is a non existent problem.  No way will the computer ever be given the choice to swerve off the road or otherwise put the passengers in danger.  It certainly will otherwise try to avoid any collision.  I mean what human realistically in the moment does anything else anyway.  I reject the idea that a driver when faced with a collision that is such a surprise it can not be avoided has time to asses the moral weight of thier life vs. the worth of the objects/beings they are about to strike.
Title: Re: Self-Driving Cars
Post by: teethering on April 22, 2015, 11:47:48 AM
There are already driverless subway and metro systems in the world functioning today.  So I'm not convinced legal or liability barriers are really that big of a deal.  I think the key thing is proving the car technology to be just as safe, but as a category it should be fine as we have similar precedents.

However, apparently technology is proving to be really hard.  Harder than people estimated originally.  On the one hand that's not too surprising, given that it's in the sort of area (machine intelligence) where we've been terrible at predicting things on the too optimistic side.  On the other hand, you'd think we'd learn by now.
Title: Re: Self-Driving Cars
Post by: Jeremy's Sea on April 22, 2015, 03:28:17 PM
I realize it's no cold comfort to a parent whose child has been plastered by a self driving car, but while there will still be accidents, there will be tens of thousands of lives saved by sleepy/drunk/anxious/inattentive drivers not having to operate a vehicle they have no business being in. Every year an elderly person plows into a crowd at a mall because they think they're braking while hitting the accelerator. With a multiple system redundancy this will be so vanishingly small of a chance as to not have to worry about it.
It seems to me any moral dilemma presented needs to look at the larger picture. Don't let perfection get in the way of good enough, or better?
Title: Re: Self-Driving Cars
Post by: Morvis13 on April 22, 2015, 03:32:08 PM
Don't forget that a computers reaction time is significantly better than a human. I think the trade off of eliminating human error is well worth it. 
Title: Re: Self-Driving Cars
Post by: Jeremy's Sea on April 22, 2015, 03:35:05 PM
Don't forget that a computers reaction time is significantly better than a human. I think the trade off of eliminating human error is well worth it.
I agree too. I still encounter people though who resist this.
Title: Re: Self-Driving Cars
Post by: Morvis13 on April 22, 2015, 03:48:58 PM
Human error accounts for like 80% of crash statistics.
Plus the car can be programmed to reduce speed in areas where children play and even during hours when they would be out.
I think stop lights and speed limits would go extinct once the humans are removed from behind the wheel.
Title: Re: Self-Driving Cars
Post by: Johnny Slick on April 22, 2015, 03:52:50 PM
I'm guessing that a lot of it would continue to work like airplane autopilot, wherein the driver would have the option to take over for the instruments in case of failure.

I think the biggest issue people have with this stems from the lack of autonomy. There's this sense that yeah, maybe a self-driving car will create fewer accidents but at least when I get into an accident I know it was my fault and not some dumb luck caused by a system crash or an eventuality someone never got around to programming into the system. I don't particularly agree with this argument but then I don't get freaked out when I go on planes because the pilot might crash it.
Title: Re: Self-Driving Cars
Post by: brilligtove on April 22, 2015, 06:35:00 PM
I'm guessing that a lot of it would continue to work like airplane autopilot, wherein the driver would have the option to take over for the instruments in case of failure.

I think the biggest issue people have with this stems from the lack of autonomy. There's this sense that yeah, maybe a self-driving car will create fewer accidents but at least when I get into an accident I know it was my fault and not some dumb luck caused by a system crash or an eventuality someone never got around to programming into the system. I don't particularly agree with this argument but then I don't get freaked out when I go on planes because the pilot might crash it.

I can't remember where I heard this, but I think there was a problem with the 'option to take over': people ceased to pay any attention at all to driving once they were comfortable with what the car could do by itself - and they were completely unprepared to do anything when it encountered conditions it couldn't deal with.
Title: Re: Self-Driving Cars
Post by: arthwollipot on April 23, 2015, 02:11:42 AM
I'm guessing that a lot of it would continue to work like airplane autopilot, wherein the driver would have the option to take over for the instruments in case of failure.

I think the biggest issue people have with this stems from the lack of autonomy. There's this sense that yeah, maybe a self-driving car will create fewer accidents but at least when I get into an accident I know it was my fault and not some dumb luck caused by a system crash or an eventuality someone never got around to programming into the system. I don't particularly agree with this argument but then I don't get freaked out when I go on planes because the pilot might crash it.

I can't remember where I heard this, but I think there was a problem with the 'option to take over': people ceased to pay any attention at all to driving once they were comfortable with what the car could do by itself - and they were completely unprepared to do anything when it encountered conditions it couldn't deal with.


I'm pretty sure that if I were in a self-driving car playing a game on my iPad and something leaped out at the car, there would be absolutely zero chance I could notice, assess the situation and take control in time to do the right thing.


We use self-driving cars to take the burden off the driver. If you're paying close attention all the time, you might as well be driving.
Title: Re: Self-Driving Cars
Post by: SkeptiQueer on February 10, 2016, 12:50:51 PM
http://www.reuters.com/article/us-alphabet-autos-selfdriving-exclusive-idUSKCN0VJ00H?utm_source=loopinsight.com&utm_medium=referral&utm_campaign=Feed%3A+loopinsight%2FKqJb+%28The+Loop%29&utm_content=FeedBurner

In boost to self-driving cars, U.S. tells Google computers can qualify as drivers

Excerpts:

"NHTSA will interpret 'driver' in the context of Google's described motor vehicle design as referring to the (self-driving system), and not to any of the vehicle occupants," NHTSA's letter said."We agree with Google its (self-driving car) will not have a 'driver' in the traditional sense that vehicles have had drivers during the last more than one hundred years."



TL:DR as I understand it, this could set the standard for not having the driver at fault so long as the computer was in control. My dream of napping or reading on the way to work without having to take three times as long for a bus to pick me up may come true.
Title: Re: Self-Driving Cars
Post by: PANTS! on February 10, 2016, 01:03:40 PM
I CAN NOT WAIT!

Title: Re: Self-Driving Cars
Post by: Ah.hell on February 10, 2016, 03:11:02 PM
Insurance companies are really going to push this once they are confident it works.   Pretty much all accidents are caused by people.

So, to some extent I agree with the op but mostly because fleet operators tend to upgrade to new vehicles more quickly than the public as a whole.   The same is true of high end consumers. 

ETA, lots of squirrels and other small animals will die.  No computer will swerve to miss a cat.  On the other hand, people are stupid so maybe it won't effect animal deaths while not swerving.
Title: Re: Self-Driving Cars
Post by: PANTS! on February 10, 2016, 03:20:59 PM
Insurance companies are really going to push this once they are confident it works.   Pretty much all accidents are caused by people.

Yeah - pretty much.  Although my sister was hit by a landslide in the NC mountains early this year.  I do not know if Self-Driving cars can avoid that.....
Title: Re: Self-Driving Cars
Post by: SkeptiQueer on February 11, 2016, 01:22:29 AM
Something I didn't highlight from the article: google specifically does not want to have the ability for a human to override. Like not even having a steering wheel or gas and brakes.  I'm sure people will still object, but frankly it's for the best. Remove the ability for fuckheads to decide that no really, we need to be able to do 95mph in a snowstorm.

I still wonder if it's going to lead to a ban on manually operated cars on some or most roads. Motorcycles too. Insurance rates for manually operated vehicles would skyrocket pretty quickly as the pool shrank.
Title: Re: Self-Driving Cars
Post by: PANTS! on February 11, 2016, 02:03:10 PM
Something I didn't highlight from the article: google specifically does not want to have the ability for a human to override. Like not even having a steering wheel or gas and brakes.  I'm sure people will still object, but frankly it's for the best. Remove the ability for fuckheads to decide that no really, we need to be able to do 95mph in a snowstorm.

I still wonder if it's going to lead to a ban on manually operated cars on some or most roads. Motorcycles too. Insurance rates for manually operated vehicles would skyrocket pretty quickly as the pool shrank.

There are two great 99% invisible episodes on this acceptance of giving up control.  Automation Paradox pt. 1 & pt. 2.  Look it up!

Somewhere out there the is also a podcast on the transition from human operated elevators to automated ones, and how society handled that change.  Only I can't remember what podcast put it out.  Sorry.   :'(

EDIT: Ah - here it is, Plant Money.

http://www.npr.org/2015/07/31/427990392/remembering-when-driverless-elevators-drew-skepticism

Title: Re: Self-Driving Cars
Post by: amysrevenge on February 11, 2016, 02:08:52 PM
As someone who does automation engineering for a living, I can tell you that we include any number of unnecessary human interfaces into the process, all the time.  A lot of it seems dumb, but it's required by customers.

I would be all for the ability for a human to override the overall choices, but not the mechanics of how to get there.  "Hey car avoid that squirrel" and then let the car figure out the best way to do that.  "Hey car, I gotta barf pull over" and let the car figure out the best way to do that.  If the "best way" involves a squirrel splat or barf inside the car, then so be it.
Title: Re: Self-Driving Cars
Post by: arthwollipot on February 11, 2016, 07:08:54 PM
Everyone who gets into the passenger seat of a car gives up control.
Title: Re: Self-Driving Cars
Post by: brilligtove on February 12, 2016, 10:11:13 AM
Everyone who gets into the passenger seat of a car gives up control.

My understanding is that the loss of control is one of the reasons some people are afraid of flying, and of being passengers.
Title: Re: Self-Driving Cars
Post by: Ah.hell on February 12, 2016, 11:30:14 AM
Everyone who gets into the passenger seat of a car gives up control.
But we don't give up the illusion of control, however slight.
Title: Re: Self-Driving Cars
Post by: Johnny Slick on February 12, 2016, 12:35:00 PM
Maybe people would be less afraid of flying if we put a toy steering wheel behind every seat.
Title: Re: Self-Driving Cars
Post by: brilligtove on February 12, 2016, 03:00:06 PM
Maybe people would be less afraid of flying if we put a toy steering wheel behind every seat.

That makes total sense. Because I want the wootastic possibility that all the idiots on this plane might accidentally actually have control.

#FearOfFlyingIntensifies
Title: Re: Self-Driving Cars
Post by: Friendly Angel on February 12, 2016, 03:39:28 PM
Maybe people would be less afraid of flying if we put a toy steering wheel behind every seat.

My wife told me her dad used to tell her that he didn't need to wear a seat belt as the driver because the steering wheel would protect him.

He was wrong then.  He's dead now. 
Title: Re: Self-Driving Cars
Post by: Mormegil on February 19, 2016, 12:05:34 PM
There are two great 99% invisible episodes on this acceptance of giving up control.  Automation Paradox pt. 1 & pt. 2.  Look it up!

Somewhere out there the is also a podcast on the transition from human operated elevators to automated ones, and how society handled that change.  Only I can't remember what podcast put it out.  Sorry.   :'(

EDIT: Ah - here it is, Plant Money.

http://www.npr.org/2015/07/31/427990392/remembering-when-driverless-elevators-drew-skepticism

I recall that's why we have the "STOP" button on elevators.  That and for dramatic moments in movies.

I guess the idea is to have a similar button in an autonomous car, without having the full controls.
Title: Re: Self-Driving Cars
Post by: Friendly Angel on February 22, 2016, 05:10:41 PM
Bar raised.

Quote
The European Commission has announced that the northern Spanish city of San Sebastian has been chosen as the location for initial tests of the CityMobil2 self-driving bus. The trial will last for three months with the 12-person capacity electric bus connecting public transport with the business area of San Sebastian’s Science and Technology Park. It will be the first time a self-driving bus has been tested with real passengers in Spain. The fleet of three vehicles will run 24 hours a day, offering free trips as well as free Wi-Fi and 4G connectivity to the 4,000 workers in the industrial park.
Title: Re: Self-Driving Cars
Post by: MikeHz on February 22, 2016, 07:54:03 PM
I have a semi-autonomous vehicle, in that it follows the car in front of it, and will self-brake to prevent collisions. Works great, most of the time. However, if the weather gets too bad, such as heavy rain or snow, it gives a message to the effect of, "Hell-I can't handle this. You're on your own, buddy!"

Now, this is fine for me, since I've been driving for decades. What happens when most people let their cars self-drive most of the time, and then have to suddenly take over in the worst driving conditions? Say, I-85 in the Columbia Gorge during a snowstorm with black ice and high wind. "Me? You expect me to take over? I haven't driven in five years!" 
Title: Re: Self-Driving Cars
Post by: PANTS! on February 22, 2016, 11:07:49 PM
Well, thats a problem.  One that we are having with pilots and auto-pilot failures. 

And the solution is to stull give the car autonomy, and never give it to tje driver.  Up to the point whete the car "refuses" to drive in terrible conditions.  Or if there is an emergency, it would have to drive at safe speeds.  Which is what I would expect a human to do.  And even then, its still going to do better than a human would in the same conditions.  Why should it ever turn control back to the humans, especially in your scenareo where they have all but forgotten how to drive.
Title: Re: Self-Driving Cars
Post by: arthwollipot on February 27, 2016, 07:18:27 PM
An interesting article about why self-driving cars should never have steering wheels:

Why Self-Driving Cars Should Never Have Steering Wheels (http://www.gizmodo.com.au/2016/02/why-self-driving-cars-really-shouldnt-ever-have-steering-wheels/)

Quote
The head of Google’s self-driving car division made headlines recently for asking US federal regulators to allow a vehicle without human-facing features like a steering wheel. Now he’s made a very good case for why no autonomous vehicle on the road should have these things at all...
Title: Re: Self-Driving Cars
Post by: phooey on February 28, 2016, 12:31:47 PM
Ah hel,

I think you just highlighted one of the many myriad problems that automated cars are going to have to deal with, and that we aren't even close to tackling.  Ok, the car doesn't swerve for dogs and cats, what about for deer?  Which deer, all deer?  Kids playing by the side of the road, do you slow for all of them?  How slow?  What if a ball is rolling into the street, do you stop?

But that's not even the tough part.  I come up to a self driving car and I stand in front of it, because I want to make you stop.  What does the car do, drive on the grass to get away from me?  Back up? 

Oh, but its even worse than this.  Two people are in the car, a husband and wife.  they have been drinking.  They are arguing, the husband says take me home.  The wife says, I am not going home with you, let me out of the car right now.  The husband says like hell you will, drive faster car.  the wife yells take me to my mothers car!  Wait, I changed my mind, stop at the liquor store.  No, wait, that is not a short cut, that road is closed.  Hey, who is that guy walking up to our car with a baseball bat, drive away quick!  And why isn't that car in front of us moving, go around him!  What do you mean you can't go around him, just drive over the trash can! 

Next, what do we do about large potholes, swerve?  When do we swerve and when do we hold tight?  What about debris?  How large does it have to be to force a panic stop?  What if it was just a balloon?

And what about that car in front which is not autonomous, it sure is creeping close to my lane, what should my car do?  is that guy looking over his shoulder because he wants to turn, or because he is yelling at his kids.  Did he forget to signal?  Did he forget to turn off his signal? What's that bicyclist about to do?   

People think they are going to be able to nap in their cars and let the car drive?  Sorry, not in this lifetime. If it is going to require the driver to be paying attention at all times, why would anyone bother. 
Title: Re: Self-Driving Cars
Post by: brilligtove on February 28, 2016, 12:43:49 PM
People think they are going to be able to nap in their cars and let the car drive?  Sorry, not in this lifetime.

Since self driving cars already exist for the simpler conditions of highway driving, and Teslas can already find their way to you from a parking spot, you're on the wrong side of the evidence on this one.

Title: Re: Self-Driving Cars
Post by: phooey on February 28, 2016, 12:50:32 PM
Sorry, but no.  Yes a self driving car could deal with the very limited requirements of a restricted highway, with a driver at the ready to deal with any issues out of the ordinary that pop up.  This is hardly the challenge.

Now what about all of the other scenarios I just listed?  Until you can come up with solutions for these, sorry, no napping in the car. 
Title: Re: Self-Driving Cars
Post by: brilligtove on February 28, 2016, 01:04:48 PM
You're so cute, the way you think angrily moving goalposts is the same as being right. Where you are right that there isn't a car that can handle absolutely any condition of the road or traffic. It is also true that there is no human who can handle absolutely any condition of the road or traffic.

Is there a fallacy called "reductio ad perfecticum"?

In any case, hands-free highway driving is here today, and Elon Musk is predicting they'll have fall-asleep quality AI in 2018 in production vehicles. He could be wrong, of course, but I'm expecting to be alive for another few years.
Title: Re: Self-Driving Cars
Post by: phooey on February 28, 2016, 01:09:13 PM
What goalpost did I move?? 

There are cars that can park themselves right now.  So does that mean self driving cars already exist? 

What I said was, there won't be self driving cars that will let the driver forget about driving and do other things, for all the reasons I explained. You can disagree if you like (although you do so without explaining how to overcome the problems I mentioned), but that is certainly not changing any goalposts.  Are you getting your sports mixed up?   
Title: Re: Self-Driving Cars
Post by: SkeptiQueer on February 28, 2016, 01:10:29 PM
There are literally self-driving cars on the road right now.
Title: Re: Self-Driving Cars
Post by: phooey on February 28, 2016, 01:29:49 PM
Yep, there sure are.  And they all require a person to be sitting right in front of the wheel, watching what is happening, so it can deal with everything the computer can't. Who is suggesting that we can't get this far?

Now, how do we get the driver out of the drivers seat?  Well, we can't.  And we won't, for quite some time.  And I just described about 20 of the 2000 reasons why we can't in the near future.

Just dealing with pot holes is going to be a nightmare.  Slow down, swerve, go over the bumpiest part of it, swerve to catch it in the least disruptive way....the computer has no idea which is best. 

You know autopilots could theoretically land an airplane, and yet they don't.  Airlines don't even allow pilots to let the plane do it. And that is in a much more controlled environment than a busy road that lots of other people use.   But yeah, sure, autopilot can fly a plane in a straight line.  That is all autopilot can do in a car. It ain't dealing with the tough stuff. 

Title: Re: Self-Driving Cars
Post by: PANTS! on February 28, 2016, 02:42:35 PM
Yep, there sure are.  And they all require a person to be sitting right in front of the wheel, watching what is happening, so it can deal with everything the computer can't. Who is suggesting that we can't get this far?

Now, how do we get the driver out of the drivers seat?  Well, we can't.  And we won't, for quite some time.  And I just described about 20 of the 2000 reasons why we can't in the near future.

Just dealing with pot holes is going to be a nightmare.  Slow down, swerve, go over the bumpiest part of it, swerve to catch it in the least disruptive way....the computer has no idea which is best. 

You know autopilots could theoretically land an airplane, and yet they don't.  Airlines don't even allow pilots to let the plane do it. And that is in a much more controlled environment than a busy road that lots of other people use.   But yeah, sure, autopilot can fly a plane in a straight line.  That is all autopilot can do in a car. It ain't dealing with the tough stuff.

This is one big argument from ignorance.  Are you really saying that there is no way in this lifetime an AI will be built that can handle pot holes.  Dude, as soon as on network enabled car detects it (maybe by going through it, but more likely via some sort of radar, or visual display reader), all the other cars will know about it.  And then they can work on avoiding it.  And they will be better at it than humans.

When the technology can provably in production stop at impromptu stop signs, avoid road debris, notice road cones, etc. why would a pothole stymie them.  The engineers have just not started to work on that problem yet.
Title: Re: Self-Driving Cars
Post by: brilligtove on February 28, 2016, 02:43:51 PM
Goalpost: A car that will drive while I am asleep behind the wheel will not exist in this lifetime. I presume you mean a span of say the next 20-50 years.

Fact: The technology for nap time cars is very close, and expected to be in production vehicles in two years.

New Goalpost: You claim that there are virtually insurmountable engineering obstacles to overcome, such as potholes, and that I must now argue against a vehicle that can navigate perfectly in all conditions, no matter how complex.

I won't do that because
A) the fact noted above, and
B) the list of engineering challenges is both infinite and irrelevant in light of (A).
Title: Re: Self-Driving Cars
Post by: arthwollipot on February 28, 2016, 07:35:19 PM
Ok, the car doesn't swerve for dogs and cats, what about for deer?  Which deer, all deer?  Kids playing by the side of the road, do you slow for all of them?  How slow?  What if a ball is rolling into the street, do you stop?
Who says the car doesn't swerve for dogs and cats? Why are your dogs and cats roaming about unleashed on the roads? Self-driving cars are designed to deal with hazards - it doesn't distinguish what those hazards are. A deer is as much of a hazard as a person is - why do you think they would not be programmed to deal with them?

But that's not even the tough part.  I come up to a self driving car and I stand in front of it, because I want to make you stop.  What does the car do, drive on the grass to get away from me?  Back up?
It takes action appropriate to the situation. That may mean stopping, that may mean slowing down and going around. Why do you think a self-driving car would not be able to deal with a situation like this?

Oh, but its even worse than this.  Two people are in the car, a husband and wife.  they have been drinking.  They are arguing, the husband says take me home.  The wife says, I am not going home with you, let me out of the car right now.  The husband says like hell you will, drive faster car.  the wife yells take me to my mothers car!  Wait, I changed my mind, stop at the liquor store.  No, wait, that is not a short cut, that road is closed.  Hey, who is that guy walking up to our car with a baseball bat, drive away quick!  And why isn't that car in front of us moving, go around him!  What do you mean you can't go around him, just drive over the trash can!
This is just as equally a problem with today's GPS navigation. You change your mind, you enter a new destination into the GPS - hopefully at the moment you pull over as you do this rather than doing it while driving. Oh but wait - in a self-driving car, you don't need to pull over to do it. Again, why on earth would you think this is a problem?

Next, what do we do about large potholes, swerve?  When do we swerve and when do we hold tight?  What about debris?  How large does it have to be to force a panic stop?  What if it was just a balloon?
Again, it is programmed to take appropriate action to deal with hazards. Why do you think this would be a problem?

And what about that car in front which is not autonomous, it sure is creeping close to my lane, what should my car do?  is that guy looking over his shoulder because he wants to turn, or because he is yelling at his kids.  Did he forget to signal?  Did he forget to turn off his signal? What's that bicyclist about to do?
The ideal situation for self-driving cars, as has been pointed out earlier in the thread, is to make all cars self-driving. That way the human error component of accidents (which is currently something like 96%) is reduced to zero. In all of the accidents that self-driving cars have been involved in during their testing, what percentage of them do you think were caused by other human drivers? That's right - one hundred percent. One hundred percent. Those damn humans are making it much much harder to program the self-drivers because human drivers are less predictable.

But again, self-driving cars are programmed to appropriately react to hazards. What's the difference between a deer, a human, and a human driving a car? They're all hazards. Why do you think this will be a problem?

People think they are going to be able to nap in their cars and let the car drive?  Sorry, not in this lifetime. If it is going to require the driver to be paying attention at all times, why would anyone bother.
The whole point is that requiring a driver to be paying attention at all times is significantly more hazardous than not having any facility for a driver at all. Computers react faster and more appropriately than a potentially panicky human who acts with limited knowledge and bad judgement.
Title: Re: Self-Driving Cars
Post by: phooey on February 28, 2016, 07:37:54 PM
Pants,

The pothole problem is just the simplest of the issues it will have to deal with, I gave you a whole list of issues that it won't be able to handle well in my opinion. 

You are suggesting when other cars go over bumps, it can relay that information to my car, which can then prepare for it.  Well, just think about how a human deals with a big bump.  You see a hole, or maybe just a slightly raised sewer cover, and you drive so it goes either between the wheels, or just on the outside of your wheels (depending on where other cars are and how much you don't won't  to be jostled), or you slow because you have no choice but to just bump over it, but its the least preferable option.  Now, how does the computer know how to deal with it, when there is a truck on my right, another car on my left, but its kind of slow, and if I sped up just a touch, I could swerve slightly left, and I won't annoy the guy on the left because I will get there just before he does, and be back in my lane in time, but I will still probably catch a bit of the hole, but it looks not too bad, so I don't think I will slow too much, oh but wait now the guy on my left is speeding up a bit more, but I think I better just hold my ground now, because I am almost there, its ok if I am a little close to the guy on the left, I think he sees me, do you think that bicyclist is turning though, maybe better hit the brakes, Oh wait, that guy behind me is too close to hit the brakes now, oh, its two potholes! 

Sometime you should think about all the split second decisions you are forced to make, in the middle of one small commuter through a large city.  Most of the decisions are not based on one best decisions, but a compromise of many. No computer in the world exists right now that can decide for me which I think is best.

Heck, if it was so easy, why don't airlines let planes land themselves?  Because we don't have computers that are smarter than the human mind right now, that's why.  Does the computer know the difference between a frightened look on a kids face who is about to run in front of traffic, and a smiling kid who doesn't appear to be in any hurry, but is just hanging out with his skateboard? 

People who think it will happen in a few years haven't thought about the problem much I contend. 



Title: Re: Self-Driving Cars
Post by: brilligtove on February 28, 2016, 07:42:23 PM
Pants,

The pothole problem is just the simplest of the issues it will have to deal with, I gave you a whole list of issues that it won't be able to handle well in my opinion. 

You are suggesting when other cars go over bumps, it can relay that information to my car, which can then prepare for it.  Well, just think about how a human deals with a big bump.  You see a hole, or maybe just a slightly raised sewer cover, and you drive so it goes either between the wheels, or just on the outside of your wheels (depending on where other cars are and how much you don't won't  to be jostled), or you slow because you have no choice but to just bump over it, but its the least preferable option.  Now, how does the computer know how to deal with it, when there is a truck on my right, another car on my left, but its kind of slow, and if I sped up just a touch, I could swerve slightly left, and I won't annoy the guy on the left because I will get there just before he does, and be back in my lane in time, but I will still probably catch a bit of the hole, but it looks not too bad, so I don't think I will slow too much, oh but wait now the guy on my left is speeding up a bit more, but I think I better just hold my ground now, because I am almost there, its ok if I am a little close to the guy on the left, I think he sees me, do you think that bicyclist is turning though, maybe better hit the brakes, Oh wait, that guy behind me is too close to hit the brakes now, oh, its two potholes! 

Sometime you should think about all the split second decisions you are forced to make, in the middle of one small commuter through a large city.  Most of the decisions are not based on one best decisions, but a compromise of many. No computer in the world exists right now that can decide for me which I think is best.

Heck, if it was so easy, why don't airlines let planes land themselves?  Because we don't have computers that are smarter than the human mind right now, that's why.  Does the computer know the difference between a frightened look on a kids face who is about to run in front of traffic, and a smiling kid who doesn't appear to be in any hurry, but is just hanging out with his skateboard? 

People who think it will happen in a few years haven't thought about the problem much I contend.

To level set, what is your expertise in the field of machine intelligence and in the field of robotics? I'm well read and was a programmer, but I'm not an engineer and don't have degrees in the field.

I ask because your pronouncements seem to be profoundly ignorant of basic technologies, like expert systems.
Title: Re: Self-Driving Cars
Post by: phooey on February 28, 2016, 07:49:52 PM
arthwollipot,

Yes, it becomes significantly easier when all cars are automatic, but how long is that going to take, when some are and some aren't?  You are going to suddenly make a billion people buy new cars? 

Secondly, you question is why are dogs and cats running around unleashed?  Have you ever lived in the real world, this happens. So what do you do for a dog, swerve or not swerve?  What do you do for a kid that is kind of running towards an intersection, but he looks like he is just jogging and he will stop when he gets to the edge?  Do you move over a little to your left, do you slow, do you stop?  What is the guy behind you thinking you are doing, you are having car trouble?  Does he go around you? 

You clearly have not thought about this at all, as your GPS comment suggests.  You are driving along and you see a quick stop on your right and you suddenly realize you need a soda.  You don't reach over and reprogram a GPS, if that was the system, it would be ridiculous.  Oh wait, what is the name of this 7-11, oh...too late already past...  Should I make a U turn...wait, car go back to that 7-11, no honey there is no time to stop right now, car keep going..No car listen to me, I want to stop...but its my car and I don't want it to stop...but we have already past...it will just take a second, turn around....

SiriCar to passengers "Whose fucking car is this, you guys get it straight so I can know what to do, you are driving me crazy, I use voice commands, get it" 
Title: Re: Self-Driving Cars
Post by: arthwollipot on February 28, 2016, 07:53:48 PM
arthwollipot,

Yes, it becomes significantly easier when all cars are automatic, but how long is that going to take, when some are and some aren't?  You are going to suddenly make a billion people buy new cars? 

Secondly, you question is why are dogs and cats running around unleashed?  Have you ever lived in the real world, this happens. So what do you do for a dog, swerve or not swerve?  What do you do for a kid that is kind of running towards an intersection, but he looks like he is just jogging and he will stop when he gets to the edge?  Do you move over a little to your left, do you slow, do you stop?  What is the guy behind you thinking you are doing, you are having car trouble?  Does he go around you? 

You clearly have not thought about this at all, as your GPS comment suggests.  You are driving along and you see a quick stop on your right and you suddenly realize you need a soda.  You don't reach over and reprogram a GPS, if that was the system, it would be ridiculous.  Oh wait, what is the name of this 7-11, oh...too late already past...  Should I make a U turn...wait, car go back to that 7-11, no honey there is no time to stop right now, car keep going..No car listen to me, I want to stop...but its my car and I don't want it to stop...but we have already past...it will just take a second, turn around....

SiriCar to passengers "Whose fucking car is this, you guys get it straight so I can know what to do, you are driving me crazy, I use voice commands, get it"

"Car, I feel like a soda. Take me to the nearest 7-11."

Well that was easy.
Title: Re: Self-Driving Cars
Post by: phooey on February 28, 2016, 08:05:09 PM
Right, except I just explained a few posts back about why voice commands are also going to be a problem in a car, when you have two people in the car, both are drunk and they are arguing.  NOW what does the computer do??
Title: Re: Self-Driving Cars
Post by: SkeptiQueer on February 28, 2016, 08:17:37 PM
Point of order: Some airplanes do land themselves, and even more could. Right now liability remains a key sticking point. It's also likely that the airlines worry about passengers not trusting a robotic plane, given the already disproportionate distrust of airlines and pilots in general.

In truth, autoland programs do most landings, with the pilot supervising. That's true for everything from 747's to F18's landing on aircraft carriers to UAVs carrying wounded IDF soldiers. The tech exists, and it works.

So, there's one major objection/analogy shot full of holes.
Title: Re: Self-Driving Cars
Post by: arthwollipot on February 28, 2016, 08:19:48 PM
Right, except I just explained a few posts back about why voice commands are also going to be a problem in a car, when you have two people in the car, both are drunk and they are arguing.  NOW what does the computer do??

Well, for a start, the computer certainly doesn't react to every voice that is heard in the car. That would be ridiculous. What if you're just talking with someone about where you want to have lunch. You don't want the car to be reacting to all of the restaurants you're considering, do you? That would be dumb.

Voice print security is technology that we already have. If you like, you could add a fingerprint scanner as well.

Seriously, these things you are bringing up as giant massive problems that will prevent self-driving cars from being autonomous in our lifetimes have already been solved.
Title: Re: Self-Driving Cars
Post by: phooey on February 28, 2016, 08:25:39 PM
So wait, you mean my wife can't get out of the car unless I tell the car to allow her?

Wait, maybe this technology is not so bad....Take me to the golf course! 
Title: Re: Self-Driving Cars
Post by: arthwollipot on February 28, 2016, 08:30:53 PM
So wait, you mean my wife can't get out of the car unless I tell the car to allow her?
To the extent that she can't get out of the car while it's moving, yes.

If you're willing to endure a brawl in your car as you physically struggle with each other, then you could set it up so that it recognises either your fingerprint or your wife's. You know, just like iPhones were able to two years ago.
Title: Re: Self-Driving Cars
Post by: brilligtove on February 29, 2016, 12:12:02 AM

To level set, what is your expertise in the field of machine intelligence and in the field of robotics? I'm well read and was a programmer, but I'm not an engineer and don't have degrees in the field.

I ask because your pronouncements seem to be profoundly ignorant of basic technologies, like expert systems.

bump
Title: Re: Self-Driving Cars
Post by: SkeptiQueer on February 29, 2016, 12:42:11 AM
So wait, you mean my wife can't get out of the car unless I tell the car to allow her?
To the extent that she can't get out of the car while it's moving, yes.

If you're willing to endure a brawl in your car as you physically struggle with each other, then you could set it up so that it recognises either your fingerprint or your wife's. You know, just like iPhones were able to two years ago.
Chip and PIN is pretty hard to dupe, yeah? I've been using some form of chip and pin card now for almost a decade, between DOD, hospital, and industrial security and it seems fairly ubiquitous. How hard would it be to provide and reasonable level of security by replacing car keys with a chip and PIN reader, so that the car only recognized input from the badgeholder/holders?

I agree that the notion that a driver would be able to sort out three disagreeing drunks any better than the car to be a bit dishonest. The only difference is that the car won't be pressured into siding with any one person by the prospect of sex or distracted by the shouting the way a human driver would. At worst, what, it sits there waiting for a clear consensus, eternally patient.
Title: Re: Self-Driving Cars
Post by: phooey on February 29, 2016, 03:07:30 AM
So does it also sit there being eternally patient while you are being carjacked?  Or while a semi truck is barreling towards you in the fog? 

Seems not many people have thought through this much.
Title: Re: Self-Driving Cars
Post by: phooey on February 29, 2016, 03:15:59 AM
arthwollipot,

I think you are still not getting problem #7 of 1231.  Does the car take commands from one person or two? If its one person, does that person need to be awake?  The insurance issues alone on this problem are going to overwhelm the technology costs. 

Title: Re: Self-Driving Cars
Post by: SkeptiQueer on February 29, 2016, 03:17:57 AM
In what scenario are two drunks arguing about where to go with a semi barreling down on them from the fog?

Of course, "ownership" solutions have already been mentioned. There's no reason, aside from technophobia, to presume that the UI in your car would be susceptible to the same faults as already $250 Xbox One. He car doesn't need to have a final destination to detect a hazard and move out of the way.

Again, a chip and pin or fingerprint would pretty well end carjacking. No point in stealing a car if you can't go anywhere in it.

Protip: install a decent firewall, stop visiting risky websites, and don't automatically click "okay" to download bloatware. My laptOP has been running through two OS generations, its second battery now no longer holds more than a 10 minute charge, and the number keys no longer work because it's been dropped too often and I don't feel like soldering. Hasn't crashed in two years, AFAIK, since I tried running Photoshop CS4.

Your computer problems aren't a result of stoopid scientists what can't do make right, it's a result of the engineers who wrote your software, hardware, and decided to allow you to have enough control to fuck it up.

Don't blame SCIENCE! for your end-user failures.
Title: Re: Self-Driving Cars
Post by: phooey on February 29, 2016, 03:43:04 AM
You just said that the car will stop and wait patiently for two people to decide where the car is going to go.  Surely your imagination is capable of conjuring two people on a dark country road arguing about where to go.  The wife says, take me to my mothers.  The husband says, no way, we are going back to the pub to hang out with the guys.  Finally their friend in the back seat says, look, let's just turn around and go back home.  The man tells the car, turn around right now. What does the car do?  Was that hard for you to imagine.  Because you need to look out the semi is coming and there is no shoulder. 

Pro Tip Question:  What's a risky website?  Does that mean anything related to Emma Watson?  Is that the level of caution I am going to need to eat safely in the future world of Monsanto that you envision?   
Title: Re: Self-Driving Cars
Post by: arthwollipot on February 29, 2016, 03:53:23 AM
You just said that the car will stop and wait patiently for two people to decide where the car is going to go.  Surely your imagination is capable of conjuring two people on a dark country road arguing about where to go.  The wife says, take me to my mothers.  The husband says, no way, we are going back to the pub to hang out with the guys.  Finally their friend in the back seat says, look, let's just turn around and go back home.  The man tells the car, turn around right now. What does the car do?  Was that hard for you to imagine.  Because you need to look out the semi is coming and there is no shoulder.

You insist on conjuring up these wacky scenarios that are designed specifically in order to make the paradigm fail.

What if a meteor falls from the sky into the road in front of the car? What will happen then? What if Al-qaeda sets off a bomb in the road while two people are having sex in the back seat? What if the car is struck by lightning while two people are having a conversation about politics - one of them is eating an ice cream and the other is looking something up on their iPad? What happens then, huh? What about if a tsunami hits and the car is swept out to sea? Will it cope with that? Huh? Will it?

I'll tell you what will happen in any of those scenarios. An appropriately programmed self-driving car will deal with the situation better and more effectively than any human could.
Title: Re: Self-Driving Cars
Post by: SkeptiQueer on February 29, 2016, 04:16:33 AM
You just said that the car will stop and wait patiently for two people to decide where the car is going to go.  Surely your imagination is capable of conjuring two people on a dark country road arguing about where to go.  The wife says, take me to my mothers.  The husband says, no way, we are going back to the pub to hang out with the guys.  Finally their friend in the back seat says, look, let's just turn around and go back home.  The man tells the car, turn around right now. What does the car do?  Was that hard for you to imagine.  Because you need to look out the semi is coming and there is no shoulder. 

Pro Tip Question:  What's a risky website?  Does that mean anything related to Emma Watson?  Is that the level of caution I am going to need to eat safely in the future world of Monsanto that you envision?
In your situation, a driver would be more distracted than an AI. The self-driving car won't turn around and look around around passenger or get involved in the argument. If it sensed a semi coming, it's pretty simple to move out of the way. Of course if the semi also is being driven by a computer instead of a fatigued human relying on caffeine, them you now have two completely alert robotic drivers both working to avoid collision. Volo already has semis that will commence emergency braking in the time it would take a human to react, better preventing collisions.

I've never gotten into an argument about where to go in the middle of a deserted road at night where a semi could present a threat to me. Again, even your convoluted scenarios don't break the system. Maybe if there was an EMP cause by micronukes to bring down a skuscraper, though...

If you spent a fraction of the time learning what's going on with your Internet use, what you're putting in your computer, as you did worrying about evil Monsanto, you wouldn't have as much of a virus issue.

If you had 100,000,000 people actively trying to poison you, then yes you'd need to be that careful with your food. You don't, though, so there's no analogy to food here. It's more like being careful who you take food or drink from at a party, or who you let handle your credit card.
Title: Re: Self-Driving Cars
Post by: phooey on February 29, 2016, 07:49:43 AM
Oh you have never gotten into an argument about where to drive?  Well, then I guess it wouldn't happen would it? Because its so uncommon for people to argue. 

Unfortunately your imagination seems to be limited to so few scenarios, that I can see why you don't think many situation could trip up a computer.  It will just drive away from dangerous situations....until it doesn't.  And then who is going to pay, Google?  Tesla?  How large do you think their insurance policy is going to need to be?

 Here is something else to test your imagination.  Computers make decisions based on yes or no.  Humans make decisions based on compromises.  A computer is told, in this situation do this, in that situation, this.  Should I drift slightly out of my lane, because the driver next to me looks distracted?  Or do I follow the rule that the middle of the lane is the best place to drive? Do I do something slightly illegal, like cross a solid line, because well, I just don't think that bicyclist is totally in control.  What about motorcycle lane splitters.  And did you see that person walking in between two cars, I think they aren't looking at me.  That patch of road ahead looks like it has a little bit of ice, I better avoid it.  There is no fricking way on this planet that we have computers that know how to deal with all of this right now.  If you think so, hows about strapping one of those systems onto a Kawasaki Ninja and letting it take you from Manhattan to Queens in as fast a time as possible.  Lets see how far you get.   

If you think a computer can do that better than a human, good luck to you. 

And computers don't even land planes better than humans, you also have that wrong.  Your quick Wikipedia check was inaccurate. 
Title: Re: Self-Driving Cars
Post by: PANTS! on February 29, 2016, 09:29:36 AM
So does it also sit there being eternally patient while you are being carjacked?  Or while a semi truck is barreling towards you in the fog? 

Seems not many people have thought through this much.

Who is going to carjack a constantly networked car?  What would they use it for?  They certainly won't take it home or to work, or to a chop shop.  Why even carjack when there will almost certainly be car sharing available?

Why is the semi-truck able to overcome it's own programming, and run into other cars?
Title: Re: Self-Driving Cars
Post by: PANTS! on February 29, 2016, 09:32:01 AM

You insist on conjuring up these wacky scenarios that are designed specifically in order to make the paradigm fail.


We call that special pleading.
Title: Re: Self-Driving Cars
Post by: phooey on February 29, 2016, 09:32:31 AM
You mean you think it will start off with all cars being self driving?  How is that going to happen? 
Title: Re: Self-Driving Cars
Post by: phooey on February 29, 2016, 09:33:36 AM
Oh yea, its so whacky to think that someone may want to stop someones car and mug them.  That would never happen! 
Title: Re: Self-Driving Cars
Post by: SkeptiQueer on February 29, 2016, 09:58:30 AM
Commercial vehicles will eliminate drivers ASAP.

Note I didn't say I'be never stopped to argue about directions, but that it's never happened in the dark in the fog out in the middle of nowhere in the path of semi.

At best, you might come up with a few scenarios where an automated vehicle would be a liability. There's already thousands of known scenarios where human drivers cause deaths. Can you demonstrate that automated cars will cause more deaths?

Computers have FAR more than just yes/no options. They can operate conditionally, and do so in real time constantly. I'm sure any of the programmers here can explain that far more, but suffice it to say that I'm playing against fairly intelligent and reactive AI right now using a relatively weak computer (playstation) and it still manages to react to what I'm doing in real-time, fractions of a second.
Title: Re: Self-Driving Cars
Post by: SkeptiQueer on February 29, 2016, 10:00:15 AM
It's whacky to assume that a carjacker will be able to stop an expert machine in a way that puts the occupants at more risk than they would be while driving normally.
Title: Re: Self-Driving Cars
Post by: phooey on February 29, 2016, 10:09:58 AM
So the self driving car will just run them over?

Or it will be smart enough to know the difference between a beggar and someone who is going to mug them?  Wow, quite a smart computer. 
Title: Re: Self-Driving Cars
Post by: phooey on February 29, 2016, 10:11:32 AM
Skeptiqueer, you need to get over to the Microsoft Sucks thread and tell all of them how the problem isn't Microsoft, its that they don't know how to use computers. 
Title: Re: Self-Driving Cars
Post by: PANTS! on February 29, 2016, 10:27:41 AM
You know what, you are 100% right.  We are all living in a fantasy.  Now can you please let us go back to talking about that fantasy, and the implications of it, instead of shooting down your gish gallop?
Title: Re: Self-Driving Cars
Post by: brilligtove on February 29, 2016, 11:02:47 AM

To level set, what is your expertise in the field of machine intelligence and in the field of robotics? I'm well read and was a programmer, but I'm not an engineer and don't have degrees in the field.

I ask because your pronouncements seem to be profoundly ignorant of basic technologies, like expert systems.
bump
rebump
Title: Re: Self-Driving Cars
Post by: Johnny Slick on February 29, 2016, 11:08:17 AM
So does it also sit there being eternally patient while you are being carjacked?  Or while a semi truck is barreling towards you in the fog? 

Seems not many people have thought through this much.

Who is going to carjack a constantly networked car?  What would they use it for?  They certainly won't take it home or to work, or to a chop shop.  Why even carjack when there will almost certainly be car sharing available?

Why is the semi-truck able to overcome it's own programming, and run into other cars?
Haven't you seen the documentary Maximum Overdrive, dumas? ::)
Title: Re: Self-Driving Cars
Post by: phooey on February 29, 2016, 11:09:24 AM
I see no evidence that you understand the complex issues computers face in dealing with real world traffic, so I am not sure why you think this post needs a response. 
Title: Re: Self-Driving Cars
Post by: SkeptiQueer on February 29, 2016, 11:15:47 AM


Skeptiqueer, you need to get over to the Microsoft Sucks thread and tell all of them how the problem isn't Microsoft, its that they don't know how to use computers.

What qualifies you to judge the analytical ability of a modern expert machine?
So the self driving car will just run them over?

Or it will be smart enough to know the difference between a beggar and someone who is going to mug them?  Wow, quite a smart computer.
So mow it's a coordinated carjacking, with a team blocking the car in on all sides all others breaking into the car, right? We don't see that typically. What makes you think this TTP is reasonable?
Title: Re: Self-Driving Cars
Post by: Johnny Slick on February 29, 2016, 11:19:53 AM
The irony, of course, is that voice recognition plus the lack of a steering wheel would make a classic carjacking hard to impossible to do. You'd need to coerce the "driver" into telling the car to drive wherever themselves (and at that, I imagine the first thing they'd add in that voice recognition software is a "safeword" that causes the car to drive to the nearest police station the same way that typing "911" on an ATM keypad locks it down and eats your card).
Title: Re: Self-Driving Cars
Post by: phooey on February 29, 2016, 11:48:24 AM
The irony of course is that it may not be the car the person with evil intent wants to steal. 

Gee, imagine that.  And since your car has no steering wheel, I guess, your computer is going to have to be pretty good at knowing when its time to break some rules. 
Title: Re: Self-Driving Cars
Post by: phooey on February 29, 2016, 11:54:11 AM
Skeptiqueer,

What in the world are you talking about? 

Why would it need to be a gang of people?  I walk right next to your car, I am 3 inches away from it, is the computer going to allow the car to drive away and run over my foot?  How does the computer know what my intention is?  Maybe I am working for the Fireman's benevolent charity? 

Please take a little time to think through some of these problems before you just attack your computer with unfettered chaotic aggression.  Take your time, think about the real world. 
Title: Re: Self-Driving Cars
Post by: brilligtove on February 29, 2016, 12:09:45 PM
I see no evidence that you understand the complex issues computers face in dealing with real world traffic, so I am not sure why you think this post needs a response.

I ask because your pronouncements seem to be profoundly ignorant of basic technologies, like expert systems.

Also, I don't need to understand the "complex issues computers face in dealing with real world traffic" because - and this is an important point so we'll slow down the narrative to give you a chance to absorb it - because real cars are self-driving now, and the manufacturers of self driving cars say they will have naptime cars in two years.

You can argue that Tesla Motors' engineers and Elon Musk don't know what they're talking about if you'd like. They seem like credible experts in the "complex issues computers face in dealing with real world traffic" though, so good luck with that.

Now, I'd love to know what qualifications you have to say that
a) the earth is flat self-driving cars are not self driving now, and
b) Tesla's plans for 2018 production of naptime cars is wrong.

So far you keep yelling "it's too hard" as if you have a four hour priapism and need a doctor.
Title: Re: Self-Driving Cars
Post by: phooey on February 29, 2016, 05:26:40 PM
Do you wanna bet?

http://www.huffingtonpost.com/entry/google-driverless-car-accident_us_56d48effe4b0871f60ec3988
Title: Re: Self-Driving Cars
Post by: arthwollipot on February 29, 2016, 06:07:29 PM
Oh you have never gotten into an argument about where to drive?  Well, then I guess it wouldn't happen would it? Because its so uncommon for people to argue.

And yet we somehow all still get to where we want to go.
Title: Re: Self-Driving Cars
Post by: arthwollipot on February 29, 2016, 06:14:21 PM
Do you wanna bet?

http://www.huffingtonpost.com/entry/google-driverless-car-accident_us_56d48effe4b0871f60ec3988

Yes indeed. The self-driving car was travelling at less than 2mph when it collided with the bus, which was travelling at 15mph, causing a minor dent. And this is the only time that a self-driving car has collided with another vehicle (as opposed to the human-controlled vehicle colliding with it) in over a million miles already driven.

That's a better driving record than I have. And I have an excellent driving record.
Title: Re: Self-Driving Cars
Post by: Johnny Slick on February 29, 2016, 06:28:35 PM
The irony of course is that it may not be the car the person with evil intent wants to steal. 

Gee, imagine that.  And since your car has no steering wheel, I guess, your computer is going to have to be pretty good at knowing when its time to break some rules.
CAMON LIBES A SELF DRIVING CAR CAN'T POSSIBLY PREVENT AGAINST NINJA VIOLENCE WHITEHOUSE IS NOT THE EXCEPTION
Title: Re: Self-Driving Cars
Post by: PANTS! on February 29, 2016, 07:26:25 PM
Can a self-driving car do this!!!!!!

(http://i.imgur.com/7A1QgFs.gif)
Title: Re: Self-Driving Cars
Post by: MikeHz on February 29, 2016, 07:41:02 PM
Possibly as time passes roads will be configured with the idea of driverless vehicles. For example, maybe roads will be designed with guide cables in them to make it easier for cars to "see" the road.
Title: Re: Self-Driving Cars
Post by: arthwollipot on February 29, 2016, 07:51:38 PM
Possibly as time passes roads will be configured with the idea of driverless vehicles. For example, maybe roads will be designed with guide cables in them to make it easier for cars to "see" the road.

Don't need 'em.
Title: Re: Self-Driving Cars
Post by: Captain Video on February 29, 2016, 09:16:29 PM
Skeptiqueer, you need to get over to the Microsoft Sucks thread and tell all of them how the problem isn't Microsoft, its that they don't know how to use computers.

I dont need to go to that thread, ill say it here, the problem isn't Microsoft and you don't know how to use computers. Its fairly obvious at this point how technologically inept you are.
Title: Re: Self-Driving Cars
Post by: phooey on March 01, 2016, 05:16:20 AM
Do all skeptics have such problems with comprehension?  It wasn't me who said Microsoft sucks, it was others who started a whole thread about it!!  Get it? 

I guess they are all stupid computer illiterates and actually Microsoft is perfect.  Go tell them! 
Title: Re: Self-Driving Cars
Post by: SkeptiQueer on March 01, 2016, 05:31:50 AM
Do all skeptics have such problems with comprehension?  It wasn't me who said Microsoft sucks, it was others who started a whole thread about it!!  Get it? 

I guess they are all stupid computer illiterates and actually Microsoft is perfect.  Go tell them!
None of them are complaining about a computer loaded with bloatware and crashing after two years. They're talking about problems with higher-end tweaking, you mentioned two user-caused problems and extrapolated from that to scientists being incompetent at systems design.

You're making the "why are there still monkeys" argument of intelligent conveyance.
Title: Re: Self-Driving Cars
Post by: brilligtove on March 01, 2016, 10:20:55 AM
A Google Self-Driving Car Got Into a Crash With a Bus (And That's Okay) (http://gizmodo.com/a-google-self-driving-car-got-into-a-wreck-with-a-bus-1762007421)

Quote
In a California DMV report discovered by Mark Harris, one of Google’s autonomous Lexus cars was stopped at a Mountain View intersection on February 14 when it needed to maneuver around sandbags placed in the right-hand lane. The car assumed—as a human driver might have—that as it nudged back into traffic, a slowing city bus was allowing the car to merge. It wasn’t, and the car struck the bus.

Three. Two. One.

Phooey!
Title: Re: Self-Driving Cars
Post by: phooey on March 01, 2016, 10:27:09 AM
Going three miles an hour and it can't avoid hitting something.  Impressive.

We still have trains that are on a track that have head on collisions, and you think we are going to have nap ready self driving cars in 2 years.  Haha. 

Did you mean 2 Pluto years? 
Title: Re: Self-Driving Cars
Post by: SkeptiQueer on March 01, 2016, 10:46:04 AM
Going three miles an hour and it can't avoid hitting something.  Impressive.

We still have trains that are on a track that have head on collisions, and you think we are going to have nap ready self driving cars in 2 years.  Haha. 

Did you mean 2 Pluto years?
1 collision per 1 million road miles. Human drivers are more like 200 crashes per million road miles.

Trains, btw, that are put on tracks by....(drumroll please) humans. Who's brakes and speed are controlled by...humans.

Go ahead. Ask why there are still monkeys.
Title: Re: Self-Driving Cars
Post by: Anders on March 01, 2016, 10:57:18 AM
Why are there still monkeys?
Title: Re: Self-Driving Cars
Post by: Captain Video on March 01, 2016, 11:50:42 AM
Ha, Brilligtove  got to it first. I heard the news on the way to work and immediately thought of this thread. Oddly enough one of my windows machines just crashed on set. Fuckin Microsoft!
Title: Re: Self-Driving Cars
Post by: brilligtove on March 01, 2016, 11:56:02 AM
I'm starting to enjoy having a troll on a leash. This may be the first time I've baited one like this, and intentionally taken a generally insulting tone with anyone online.

I presume Phooey didn't read the part where the human driver made exactly the same mistake as the AI?

#OnceYouGoBlackDarkYouNeverGoBack
Title: Re: Self-Driving Cars
Post by: TheIrreverend on March 01, 2016, 02:18:11 PM
Administrator Comment This is a warning that the mods have a drone in the area.
Title: Re: Self-Driving Cars
Post by: The Latinist on March 01, 2016, 02:18:49 PM
Administrator Comment This is a warning that the mods have a drone in the area.

Self-driving drone?
Title: Re: Self-Driving Cars
Post by: Captain Video on March 01, 2016, 03:35:46 PM
I don't think he is a troll. I have been enjoying the argument. This would be a weird thread to troll in. Am I missing something?

Or Are we being overly mean? If so I apologize and encourage Phooey to stick around.

Brill is not usually mean (in a direct way LOL) Maybe he is breaking out of his shell and I would encourage that just a little 
Title: Re: Self-Driving Cars
Post by: phooey on March 01, 2016, 07:30:47 PM
Going three miles an hour and it can't avoid hitting something.  Impressive.

We still have trains that are on a track that have head on collisions, and you think we are going to have nap ready self driving cars in 2 years.  Haha. 

Did you mean 2 Pluto years?
1 collision per 1 million road miles. Human drivers are more like 200 crashes per million road miles.

Trains, btw, that are put on tracks by....(drumroll please) humans. Who's brakes and speed are controlled by...humans.

Go ahead. Ask why there are still monkeys.

Wrong again.

Self-Driving Cars Are More Accident-Prone, Study Finds

The study, by researchers Brandon Schoettle and Michael Sivak at the University of Michigan’s Transportation Research Institute, found that self-driving cars are in accidents at five times the rate of human-controlled cars.


http://time.com/4098303/self-driving-cars-accident/


And the cars are limited to very controlled environments and a maximum speed of 25 miles per hour!   

Spectacularly wrong. 
Title: Re: Self-Driving Cars
Post by: Captain Video on March 01, 2016, 08:11:08 PM
Ok I guess I was missing something, I started to read Phooeys posts in other threads, Yea Kinda troll like >:D

You tell us man, what are you here for? Ill give you the benefit of doubt, I probably would have been considered a troll when I first got here had it not been for Mafia.
Now about that article, You also seem to have missed the bold part in the article that said
Quote

But there are several caveats to the findings

and

Quote
However, the new research comes with a laundry list of caveats that indicates that transportation experts are still struggling to ascertain just how safe self-driving cars actually are.

What I get from the article is  " The numbers say this but in truth... we don't know" 
Title: Re: Self-Driving Cars
Post by: brilligtove on March 01, 2016, 08:16:14 PM
I'll tone it down. Just got a little giddy there.

I wouldn't be surprised if the numbers have changes substantially in the 4 months since the article (and the longer time since the study was done). This tech is advancing rapidly.
Title: Re: Self-Driving Cars
Post by: phooey on March 02, 2016, 04:34:32 AM
Simon Jester,

I don't even know what being a troll means to be honest, but I am fairly sure it doesn't simply mean having opinions that are not shared by some others. 

As far as the study, heck sure I can acknowledge the caveats.  However that is in direct conflict with skeptiqueers claim that self-driving cars get into 5 times LESS accidents than human controlled cars. 

The fact is that computer controlled cars do not perform well in comparison to their human counter-parts at this time.  Can you imagine how overwhelmingly bad they would perform if they were NOT in controlled environments and if they were allowed to drive at full speed?  People would be getting hit, bicyclist would be getting hit, they would be plowing head on into semi trucks, they would be driving off the edges of roads-I don't think there is any doubt this would happen. And we haven't even begun to consider what happens when the computers fail, or when they get hacked, or when the whole system suddenly is knocked out. 

And who do you think is going to pay for all this mess, auto-insurance companies?  They are going to just say, Ok,, we trust you Google, I am sure you will make it safe, here is a 10 trillion dollar insurance policy in case it all fucks up some day. 


Now, if my observations don't jibe with a bunch of people who started calling themselves skeptics, but are in fact just another group of people who believe everything they are told by their skeptical Gods, well, then I am fine with being called a troll.  Since the word "skeptic" has been hijacked by a bunch of people who refuse to be skeptical at all about a whole host of scientific claims, I definitely would not be happy being called a skeptic. 
 
Title: Re: Self-Driving Cars
Post by: SkeptiQueer on March 02, 2016, 05:35:28 AM
Simon Jester,

I don't even know what being a troll means to be honest, but I am fairly sure it doesn't simply mean having opinions that are not shared by some others. 

As far as the study, heck sure I can acknowledge the caveats.  However that is in direct conflict with skeptiqueers claim that self-driving cars get into 5 times LESS accidents than human controlled cars. 

The fact is that computer controlled cars do not perform well in comparison to their human counter-parts at this time.  Can you imagine how overwhelmingly bad they would perform if they were NOT in controlled environments and if they were allowed to drive at full speed?  People would be getting hit, bicyclist would be getting hit, they would be plowing head on into semi trucks, they would be driving off the edges of roads-I don't think there is any doubt this would happen. And we haven't even begun to consider what happens when the computers fail, or when they get hacked, or when the whole system suddenly is knocked out. 

And who do you think is going to pay for all this mess, auto-insurance companies?  They are going to just say, Ok,, we trust you Google, I am sure you will make it safe, here is a 10 trillion dollar insurance policy in case it all fucks up some day. 


Now, if my observations don't jibe with a bunch of people who started calling themselves skeptics, but are in fact just another group of people who believe everything they are told by their skeptical Gods, well, then I am fine with being called a troll.  Since the word "skeptic" has been hijacked by a bunch of people who refuse to be skeptical at all about a whole host of scientific claims, I definitely would not be happy being called a skeptic.
"Most of the self-driving vehicles involved in accidents were hit in the rear when they were traveling 5 miles per hour or slower. None were involved in very serious accidents, such as head-on collisions. "

I'm talking about accidents caused by the car. So far this study points to human drivers still causing the accidents. They're not running over bicycles, pedestrians, or running off roads, they're getting hit by human drivers.

Thanks for proving my point though!
Title: Re: Self-Driving Cars
Post by: arthwollipot on March 02, 2016, 05:56:11 AM
You'd have to be seriously disconnected from the literature to believe that self-driving cars are involved in more accidents than human-driven cars.
Title: Re: Self-Driving Cars
Post by: phooey on March 02, 2016, 06:38:07 AM
If there are more accidents that happen, then with human drivers, its not much consolation claiming that it was the humans fault.  If a car suddenly makes a panic stop where one wouldn't expect it, and a driver rear ends it, that is not much comfort to know that a human probably wouldn't have done that. 

And this is at 25 miles per hour tops! 

When they pull in front of a bicyclist going 55 miles an hour because they couldn't detect it, and the bicyclist slams into it, he is probably not going to feel so great knowing skeptiqeer says its his fault.  The insurance companies will probably feel the same too. 
Title: Re: Self-Driving Cars
Post by: Guillermo on March 02, 2016, 11:45:57 AM
So does it also sit there being eternally patient while you are being carjacked?  Or while a semi truck is barreling towards you in the fog? 

Seems not many people have thought through this much.
How is this any different from the current situation? This situation will not deter any considerations for autonomous cars at all. In the case that someone is using the Autonomous Car like standing in front and behind to make the car sit still. Well, I don't think that can change. It is quite easy to program something for that particular scenario. Quite likely a Lock all doors and windows and call the cops type of mechanism. This will happen rarely, it's easier to capture a human driven car. PS. I live in a country that this things happen every once in a while. An autonomous car will not be any different than a regular car.
Title: Re: Self-Driving Cars
Post by: bimble on March 02, 2016, 02:09:17 PM
If there are more accidents that happen, then with human drivers, its not much consolation claiming that it was the humans fault.  If a car suddenly makes a panic stop where one wouldn't expect it, and a driver rear ends it, that is not much comfort to know that a human probably wouldn't have done that. 

Though totally the following driver's fault for not leaving enough room to be able to react to any actions that the car ahead is doing


When they pull in front of a bicyclist going 55 miles an hour because they couldn't detect it, and the bicyclist slams into it, he is probably not going to feel so great knowing skeptiqeer says its his fault.  The insurance companies will probably feel the same too.

A cyclist, sorry, bicyclist doing 55mph?!? Is this one of the unrealistic situations mentioned earlier??
Title: Re: Self-Driving Cars
Post by: phooey on March 02, 2016, 08:40:50 PM
Oh my goodness. The car!  When the car is going fifty five miles an hour! 

Right now the cars are limited to only 25 miles an hour and already they cause more accidents than human driven cars. 

No wonder you all think they are going to be here next year. 
Title: Re: Self-Driving Cars
Post by: SkeptiQueer on March 02, 2016, 08:50:10 PM
Why do you think your bicycle scenario is possible?

You seem to be arguing mainly that automated automobiles are bad because they follow the rules of the road too well, but in that case, it's not their fault is it?

I mean by your logic, we shouldn't start teaching the next generaton to drive better because the shitty drivers will run into them and it'll be the better drivers' fault.

You're not making any sense, and based on the data we should be incentivizing Johnny Cabs as much as possible in the interest of getting as many human drivers off the road as possible.
Title: Re: Self-Driving Cars
Post by: phooey on March 02, 2016, 08:50:17 PM
Guillermo,

I suspect if given a choice, most people would prefer the drive the fuck outta there as fast as possible in a human powered car, to the sit locked in your automated car shivering waiting for a cop to hopefully arrive in 15 minutes while the guy with the bat and the tattooed knuckles is taking aim at your windows. 

But hey, at least we will still have a choice for a while. 
Title: Re: Self-Driving Cars
Post by: phooey on March 02, 2016, 09:04:11 PM
No skeptiqueer, you don't get it at all. 

Following the rules of a pre-thought out script that some programmer thinks is the best algorithim is NOT the safest way to travel, when you have a complex world out there of kids, animals, drunks, debris, skateboarders, old ladies, drunk drivers, new drivers, road construction, ice, people who skipped their meds, and stickballers- that either don't or can't always follow a pre-programmed script.   

If you set out to simply follow every rule of the road, you know like stay in the middle of your lane, even when someone looks like they are drifting dangerously close to yours, and they seem to be looking backwards, then you are going to have more accidents than the guy who says, I don't think I am going to wait to find out how this one will play out. 

You mind is in a fairyland, and yet amazingly you don't seem to have the slightest imagination for how things actually happen in the real world.  How many times did you have to make a decision about 12 different things going on near the road, in you short little jaunt to the grocery store?  Computers don't make decisions, they say either yes, or no.  They don't understand the concept of slight compromises, each with their own inherent disadvantages and advantages. 

What kind of silly justification is it that you are making, that well, if people were just smarter than we wouldn't have this problem.  Yes, but there are still going to be people out there!! 

And guess who is going to design the programs that drive these cars???? Use your imagination. Those same dumb people!!!

Title: Re: Self-Driving Cars
Post by: SkeptiQueer on March 02, 2016, 09:19:41 PM
You understand that it's not a simple script, right? These things have a sensor array slightly less complex than my MRA did (I don't think they have some of the shit I can't talk about, but they do have color vision) and are programmed to react to other things on the road.

People keep asking about your background in programming because you keep talking about these this like they're programmed on a Galaxy arcade and running on totally closed automation.

It has nothing to do with being smarter, it's about driving safely. Google cars don't mash a hamburger in their pie hole while changing lanes at 15 over the limit, they don't get drunk or tired, they don't road rage, they don't gun it to make a light or run red lights. They can't get distracted by a text message or an argument and run a stop sign. They won't look left and miss the cyclist crossing the intersection on the right, because they look in all directions at once.

Your argument, once the rote misunderstanding of the technology is removed, is still that people are dangerous drivers. That's a better argument for pushing autonomous transit than not though.
Title: Re: Self-Driving Cars
Post by: phooey on March 02, 2016, 09:26:04 PM
And yet a computer still can't land a plane as well as a person, imagine that. 
Title: Re: Self-Driving Cars
Post by: Caffiene on March 02, 2016, 09:27:57 PM
Following the rules of a pre-thought out script that some programmer thinks is the best algorithim is NOT the safest way to travel, when you have a complex world out there of kids, animals, drunks, debris, skateboarders, old ladies, drunk drivers, new drivers, road construction, ice, people who skipped their meds, and stickballers- that either don't or can't always follow a pre-programmed script.

Humans are terrible at predicting these sort of things too; and when they do occur, computers can react much faster and in much more optimal ways.

Most humans faced with (say) an animal that might potentially run on the road dont do the safe thing, theyre just as likely to barely notice it and then panic at the last minute and swerve into oncoming traffic. Or write-off their car by rolling it trying to swerve around a snake on the road like an acquaintance of mine did.
Title: Re: Self-Driving Cars
Post by: SkeptiQueer on March 02, 2016, 09:33:02 PM
And yet a computer still can't land a plane as well as a person, imagine that.
[CITATION NEEDED]
Title: Re: Self-Driving Cars
Post by: brilligtove on March 02, 2016, 10:11:06 PM
And yet a computer still can't land a plane as well as a person, imagine that.
[CITATION NEEDED]

Ask the Captain: When Planes Land Themselves (http://www.usatoday.com/story/travel/columnist/cox/2014/02/09/autoland-low-visibility-landings/5283931/)

Quote
Question: Can large commercial airliners be landed with only computers and no pilot intervention?

-- submitted by reader Raymond, Charlotte, N.C.

Answer: Many modern airplanes can autoland. These new generation jets can fly the approach, flare, touch down, track the centerline of the runway, apply brakes and deploy the ground spoilers. They cannot apply thrust reverse and cannot turn off the runway.

It requires significant pilot input to set up the autoflight systems for the approach, and to arm the proper modes of the systems. I would not say there is no pilot intervention, but it is accurate to say the pilots are not manually flying.

There's another question in the article.

ETA: I googled "can a computer land a plane" and clicked the first link to find this. It might have taken ten seconds to find.
Title: Re: Self-Driving Cars
Post by: wastrel on March 02, 2016, 10:20:39 PM
And yet a computer still can't land a plane as well as a person, imagine that.

The immediate refutes aside, you make it sound as though driving is more complex than landing a plane.  What an odd approach to this argument.
Title: Re: Self-Driving Cars
Post by: phooey on March 02, 2016, 11:10:37 PM
You know brilligtove, you are just such a typical "skeptic".  Wait, wait, let me look up real quick on google,  This is the level of your thought. 

First its not that a computer CAN'T land a plane, its that they don't do it better than humans. But in your frenzy to be a good skeptic comrade, you thought, Ok, I can find out something on Google, then I can show how brilliant I am at arguing.   Nevermind that you missed the point completely, and that you can't even use Google very well.  Autopilot functions are not as good as human control, and a good pilot would never choose that over landing it himself.  Pilots are sometimes forced to use autopilot when they can't see, but it is definitely not their preferred option.

Instead of sitting there using Google to try to show how smart you are, the other options is to actually be in the world and interact with pilots.  Maybe learn WHY it is that humans can control the plane better than a computer. 

Aspire to something more than being a self-professed search expert.   
Title: Re: Self-Driving Cars
Post by: SkeptiQueer on March 02, 2016, 11:12:11 PM
You know brilligtove, you are just such a typical "skeptic".  Wait, wait, let me look up real quick on google,  This is the level of your thought. 

First its not that a computer CAN'T land a plane, its that they don't do it better than humans. But in your frenzy to be a good skeptic comrade, you thought, Ok, I can find out something on Google, then I can show how brilliant I am at arguing.   Nevermind that you missed the point completely, and that you can't even use Google very well.  Autopilot functions are not as good as human control, and a good pilot would never choose that over landing it himself.  Pilots are sometimes forced to use autopilot when they can't see, but it is definitely not their preferred option.

Instead of sitting there using Google to try to show how smart you are, the other options is to actually be in the world and interact with pilots.  Maybe learn WHY it is that humans can control the plane better than a computer. 

Aspire to something more than being a self-professed search expert.
So you've got some evidence for your claim that human pilots are better at landing, right?

Right?
Title: Re: Self-Driving Cars
Post by: phooey on March 02, 2016, 11:20:50 PM
Yes.  But since not all of life's knowledge can be obtained by looking at the first five links to a Google search, you will probably not learn about it. 
Title: Re: Self-Driving Cars
Post by: SkeptiQueer on March 02, 2016, 11:37:58 PM
My brother in law is a pilot. Before his current employ, he was a flight instructor for Delta, teaching new pilots to operate everything from single-engine up to Boeing 747s. I grew up with two pilots as teachers. My gym budy on one deployment was our Air Officer, a former F22 pilot doing a rotation in an Infantry Battalion as a leg up to his first star. I work at a place where planes are built. One of my polyamorous friends and another of our close friends worm for the same company, and within the extended circle we all talk with the engineers who design the planes and their automation systems. One of the topics we discuss over locally brewed oatmeal stout is the Boeing vs Airbus attitude toward autpilot/autoland command and control programming.

I'm not going off google, so if you could kindly point me to the evidence you have for the claim that contradicts all the firsthand accounting I have of autoland being as capable or more capable, I'd love to hear it.

Particularly since you seem think pilots can't land without visibility.
Title: Re: Self-Driving Cars
Post by: phooey on March 03, 2016, 12:20:59 AM
Then go ask your brother-in-law if he would prefer to use the autopilot to land the plane (which his company might not even allow) or land it himself, and if he tells you he feels it safer for the plane to land itself, then your story is untrue.   
Title: Re: Self-Driving Cars
Post by: SkeptiQueer on March 03, 2016, 12:22:54 AM
Either I agree with you or I'm a liar. Yep, that's a reasonable position.

So where, again, is your evidence?
Title: Re: Self-Driving Cars
Post by: Alex Simmons on March 03, 2016, 01:51:55 AM
If there are more accidents that happen, then with human drivers, its not much consolation claiming that it was the humans fault.  If a car suddenly makes a panic stop where one wouldn't expect it, and a driver rear ends it, that is not much comfort to know that a human probably wouldn't have done that. 

And this is at 25 miles per hour tops! 

When they pull in front of a bicyclist going 55 miles an hour because they couldn't detect it, and the bicyclist slams into it, he is probably not going to feel so great knowing skeptiqeer says its his fault.  The insurance companies will probably feel the same too.
Google's self drive car has already encountered cyclists, and this story is an example we in the cycling community discussed at the time:

http://www.bicycling.com/culture/safety-etiquette/trackstanding-cyclist-perplexes-googles-self-driving-car

and Google has been working on the issue of improving safety for cyclists:
http://www.bicycling.com/culture/news/google-s-new-self-driving-car-could-mean-safer-roads-cyclists

I can assure you, removing the hot headed idiots from making driving decisions that regularly threaten cyclists lives, this looks like a really positive innovation. Bring it on.

https://www.youtube.com/watch?v=TsaES--OTzM
Title: Re: Self-Driving Cars
Post by: SkeptiQueer on March 03, 2016, 02:42:12 AM
 Germanwings 9525 could only have happened with a manual controlled aircraft. Robots don't get depressed.

The Airbus vs Boeing debate is about whether the aircraft should be able to refuse to do something like auto-descend into a mountain on the pilot's command, or if it should flip the pilot a bird and ascend to a safe height. Pilots will stab each other over it, or cite the water landing in NY as evidence that real pilots are just better, but when it comes down to it pilot error has downed at least six planes in the last five years. Pilot heroics has purportedly saved one.

Similarly, even if we presume that carjacking is a real threat (in the US half are successful, and most involve a weapon and occur when the vehicle is stopped such as entry or exit or in as confined space where flooring it will be unsuccessful) we would weigh about 38,000 carjackings against 500,000 reported (important word, reported) accidents, 30,000 fatal crashes and 32,000-35,000 fatalities every year.

It would be interesting to sit down and talk with the engineers about solving issues like emergency driving (I need to get to the hospital NOW, does my car send out an emergency signal that clears a path? Or does JohnnyCab send a special vehicle? Or will a UAV perform a Foley extraction?) and responding to threats (will JohnnyCab spray OC at the assailant like the hire cars in SA? Will armed security replace taxi drivers?) with expert systems, but it's intellectually dishonest to keep inventing scenarios as reasons why the tech can't work, especially when they're not true (planes vs pilots) or already addressed (seeing bicycles).

I'm really interested to see what incentives will arise, whether fleets of hire cars will replace POVs, or if there will just be incentives for people to replace dumb cars with expert drivers.

I for one would happily transition to a JohnnyCab system as long as it's as cheap or cheaper than a car payment+insurance. Take a commuter to and from work, rent out a station wagon to take me to the big rifle range outside the city for a day, and then call something a little spacier for a trip with SO to the train station, then call a JohnnyCab to run from the station in Chicago to our hotel.

Stop killing my dreeeeeams.
Title: Re: Self-Driving Cars
Post by: arthwollipot on March 03, 2016, 02:57:57 AM
Right now the cars are limited to only 25 miles an hour and already they cause more accidents than human driven cars.

No, they don't. They cause far, far fewer accidents than human driven cars. I have no idea where you're getting your information from but it is so far wrong it's actually backwards.
Title: Re: Self-Driving Cars
Post by: phooey on March 03, 2016, 08:04:11 AM
Alex,

You mean going 25 miles per hour, with a driver in the seat to keep it from going off course, in controlled environment roads they haven't caused any cycling accidents yet?  Gee, I am so glad to hear that (I am also a cyclist). 

Now let me ask you, if you get into an accident with a self driving car because of something it did wrong, and all your teeth got knocked out, would you sue do you think? Who would you sue?   
Title: Re: Self-Driving Cars
Post by: phooey on March 03, 2016, 08:25:29 AM
SkeptiQueer,

If you think I wouldn't like to have self driving cars, that's not true.  I think it would be great, and I am sure one day, about 100 years from now, we will. 

In the meantime all of this talk of having them in two years, ...nah.  We are just not that smart yet.  We always think we are, but we are not.  We still can't provide safe tap water for likely millions of people.  We have no clue what our next energy source is.  We have devastated the middle class.  Our highways and bridges are crumbling.  Our food is killing people by the pound.  We have a tough time keeping credit card information from being stolen.  The health care system in America has probably gotten worse not better in the last 20 years.  And Donald Trump might even get elected as President. 

If someone is going to come up with a clever system for transportation, maybe America isn't the one who is going to do it.   
Title: Re: Self-Driving Cars
Post by: SkeptiQueer on March 03, 2016, 08:40:55 AM
That's a giant non-sequitur.

You still haven't provided your evidence that pilots land better than autoland.
Title: Re: Self-Driving Cars
Post by: phooey on March 03, 2016, 08:55:48 AM
Ask your brother in law.

BTW:

Most Drivers Say They Wouldn’t Feel Safe in a Self-Driving Car (75% in fact.  Whoa that's high!)

http://finance.yahoo.com/news/most-drivers-wouldn-t-feel-224841978.html
Title: Re: Self-Driving Cars
Post by: phooey on March 03, 2016, 09:57:56 AM
"The ability to judge intention and respond accordingly is also central to driving. From determining whether a pedestrian is going to jaywalk to slowing down and avoiding a driver who seems drunk or tired, we do it constantly while behind the wheel. Self-driving cars can’t do this now. They likely won’t be able to do it for years. But this isn’t just about routine-but-confusing interactions like that between the Google self-driving car and the Mountain View bus.

Even the best AIs are easy to fool. State-of-the-art object recognition systems can be tricked into thinking a picture of an orange is really an ostrich. Self-driving cars will be no different. They will make errors—which is not so bad on the face of it, as long as they make fewer than humans. But the kinds of errors they make will be errors a human would never make. They will mistake a garbage bag for a running pedestrian. They will mistake a cloud for a truck.

All of this means that self-driving cars will be incredibly easy to troll. Sticking a foot out in the road or waving a piece of cloth around might be enough to trigger an emergency stop. Tapping the brakes of your car could trigger a chain reaction of evasive maneuvers. Perhaps a few bored 12-year-olds could shut down L.A. freeways with the equivalent of smiley faces painted on balloons."


http://www.slate.com/articles/technology/future_tense/2016/03/google_self_driving_cars_lack_a_human_s_intuition_for_what_other_drivers.html


I guess I am not the only one who has thought of this. 
Title: Re: Self-Driving Cars
Post by: SkeptiQueer on March 03, 2016, 10:08:52 AM
Why should I ask my BIL? You already told me either he agrees with you or I'm lying. He used the autoland the times I flew with him in a duly equipped aircraft, but that may have been company policy.

Do you have evidence or not?

A whole bunch of Americans don't trust all kinds of stuff. You just called them all stupid, so why should we listen to what all these people you think are stupid think?

That Slate article points out that people are terrible drivers, just like you did. Isn't that a great reason to get as many human drivers off the road as possible?
Title: Re: Self-Driving Cars
Post by: phooey on March 03, 2016, 10:23:18 AM
If that's what you think the article said, I leave it to others to see for themselves.  Some might disagree with your imaginative interpretation.
Title: Re: Self-Driving Cars
Post by: Morvis13 on March 03, 2016, 10:35:43 AM
Nah, I say program the robot to kill anything that crosses in front of it. Then people will learn not to jaywalk.
We need as many human removed from driving as possible. Once the robots take over we remove 80% of the accidents.
Once the idiots are no longer driving erratically we can rev up the robot speed.
Title: Re: Self-Driving Cars
Post by: brilligtove on March 03, 2016, 10:53:19 AM
You know brilligtove, you are just such a typical "skeptic".

You keep using that word, "skeptic"...

(http://www.prdaily.com/Uploads/Public/i-do-not-think-it-means-what-you-think-it-means.jpg)

And yet a computer still can't land a plane as well as a person, imagine that.
Wait, wait, let me look up real quick on google,  This is the level of your thought. 

Since your assertions are asinine and also trivial to debunk it isn't like we have a hard time showing you're wrong. A simple, direct response to your assertion that a computer can't land a plane as well as a person seemed appropriate, as was a quick demonstration of how easy it is to find facts related to your incorrect assertion.

(click to show/hide)

First its not that a computer CAN'T land a plane, its that they don't do it better than humans.

It is clear that you can read, so the problem must be with your capcity to comprehend. I'll use small words for you: all the facts show you are wrong.

Autopilot functions are not as good as human control, and a good pilot would never choose that over landing it himself.  Pilots are sometimes forced to use autopilot when they can't see, but it is definitely not their preferred option.
[SNIP]
Maybe learn WHY it is that humans can control the plane better than a computer. 

 ::)

But in your frenzy to be a good skeptic comrade, you thought, Ok, I can find out something on Google, then I can show how brilliant I am at arguing.   Nevermind that you missed the point completely, and that you can't even use Google very well.
[SNIP]
Instead of sitting there using Google to try to show how smart you are, the other options is to actually be in the world and interact with pilots.
[SNIP]
Aspire to something more than being a self-professed search expert.

Hrm. I wonder where the Mod Drone is now...? ;)
Title: Re: Self-Driving Cars
Post by: phooey on March 03, 2016, 11:08:58 AM
"The ability to judge intention and respond accordingly is also central to driving. From determining whether a pedestrian is going to jaywalk to slowing down and avoiding a driver who seems drunk or tired, we do it constantly while behind the wheel. Self-driving cars can’t do this now. They likely won’t be able to do it for years. But this isn’t just about routine-but-confusing interactions like that between the Google self-driving car and the Mountain View bus.

Even the best AIs are easy to fool. State-of-the-art object recognition systems can be tricked into thinking a picture of an orange is really an ostrich. Self-driving cars will be no different. They will make errors—which is not so bad on the face of it, as long as they make fewer than humans. But the kinds of errors they make will be errors a human would never make. They will mistake a garbage bag for a running pedestrian. They will mistake a cloud for a truck.

All of this means that self-driving cars will be incredibly easy to troll. Sticking a foot out in the road or waving a piece of cloth around might be enough to trigger an emergency stop. Tapping the brakes of your car could trigger a chain reaction of evasive maneuvers. Perhaps a few bored 12-year-olds could shut down L.A. freeways with the equivalent of smiley faces painted on balloons."


http://www.slate.com/articles/technology/future_tense/2016/03/google_self_driving_cars_lack_a_human_s_intuition_for_what_other_drivers.html


I guess I am not the only one who has thought of this.

Read this first brillingtrove, then get back to me. 
Title: Re: Self-Driving Cars
Post by: Friendly Angel on March 03, 2016, 02:54:48 PM
If there are more accidents that happen, then with human drivers, its not much consolation claiming that it was the humans fault.  If a car suddenly makes a panic stop where one wouldn't expect it, and a driver rear ends it, that is not much comfort to know that a human probably wouldn't have done that. 

And this is at 25 miles per hour tops! 

When they pull in front of a bicyclist going 55 miles an hour because they couldn't detect it, and the bicyclist slams into it, he is probably not going to feel so great knowing skeptiqeer says its his fault.  The insurance companies will probably feel the same too.
Google's self drive car has already encountered cyclists, and this story is an example we in the cycling community discussed at the time:

http://www.bicycling.com/culture/safety-etiquette/trackstanding-cyclist-perplexes-googles-self-driving-car

and Google has been working on the issue of improving safety for cyclists:
http://www.bicycling.com/culture/news/google-s-new-self-driving-car-could-mean-safer-roads-cyclists

I can assure you, removing the hot headed idiots from making driving decisions that regularly threaten cyclists lives, this looks like a really positive innovation. Bring it on.

https://www.youtube.com/watch?v=TsaES--OTzM

Fook yeah!  On top of the sensors; when patience is just a programming feature the streets will be safer for everybody.
Title: Re: Self-Driving Cars
Post by: Captain Video on March 03, 2016, 03:24:04 PM
If all the cars are automatic why would (non)drivers facial expressions mater? The computers would speak to each other or the road or whatever system is in place. You keep complaining about how these cars will react with humans when the real question is how will they react with each other. Again the answer is much better than humans, there will be no prediction, the car in front will tell the car in back "i'm about to turn"  lights and signals can be changed to warn pedestrians rather than other cars.  Cyclists and joggers can use the new "road safety phone app" I just came up with in my head so cars will be notified of their presence (note Cyclists must now be responsible with electronic turn signalling, get over it, you should be doing it now). Highway Speeds can be increased safely. There is no end to the improvements, go to the track if you want to put your life on the line and drive.

We can stop wasting money on cops patrolling the traffic and small townships will have to find another source of income instead of raping drivers all day

The one thing we should do however is put some kind of legislation in place so that the authorities can not abuse the power of taking over your cars function, only for safety or with a warrant, etc even then, the Apple/FBI argument will apply here as well, no you cant have a key to all the cars. They will be able to track you though and i have never felt that was a violation of my freedom, its how they use that data that can be the problem
Title: Re: Self-Driving Cars
Post by: Alex Simmons on March 03, 2016, 04:08:36 PM
Alex,

You mean going 25 miles per hour, with a driver in the seat to keep it from going off course, in controlled environment roads they haven't caused any cycling accidents yet?  Gee, I am so glad to hear that (I am also a cyclist). 

Now let me ask you, if you get into an accident with a self driving car because of something it did wrong, and all your teeth got knocked out, would you sue do you think? Who would you sue?
Did you even watch the video? Or other such examples? The drivers never control the vehicle to keep it on course, only monitor.

And in many cities where there is greatest level of cyclist-motor vehicle driver conflict, guess what the average speeds are? In and around my city, I'm lucky to average 25km/h in a car, let alone 25mph. Between the city and my home, my max average speed in the car has been 24km/h. In peak hour it's generally ~13km/h.

I can assure you, it seems to me this tech is evolving the the stage where on the whole I would be far safer with self driven cars than human driven ones.

The argument about who I might sue in an accident isn't relevant - it's a complete red herring.

What is relevant is whether the rate of collision or crash inducing incident is reduced significantly.

I'd much rather Google or whomever had to deal with 10 claims per year, than the 10,000 that the insurance companies may need to deal with at present. And guess what would happen to insurance premiums when the rate of accident falls dramatically? Guess what happens to the cost of emergency medical and trauma care when the number of such people reduces dramatically?

Drivers who seemingly are looking right into your eye still cause problems. I've experienced this when such a driver stopped at an intersection that we were both approaching (I had right of way), he was looking right at my eyes and appeared to yield (I don't assume they will just because I have right of way) so I continued. Yet despite that he then took off anyway and t-boned me with his 4WD bull bar and sent me flying into the oncoming traffic lane. Yet here we have a self drive car that's detecting a rider making minute movements due to an unusual riding technique (track standing) and the car is applying a high degree of caution. More so than a large number of drivers would.

I'm leaving my current location because it's become too hostile to cycle. Such tech may one day resolve that. Bring it on.
Title: Re: Self-Driving Cars
Post by: PANTS! on March 03, 2016, 04:35:54 PM
It's almost as if phooey's blind assistance to press onward, thinking he is right, thinking he is better suited than any other option, and ignoring outside evidence and input.  Speeding ahead towards the crash with righteous blinders on.  Not able to see the risk despite it being right there in front of him.

Its like.... its almost analogous to some situation.  It's on the tip of my tongue what would be a very similar situation for this behavior. 

Title: Re: Self-Driving Cars
Post by: Friendly Angel on March 03, 2016, 05:00:05 PM

Its like.... its almost analogous to some situation.  It's on the tip of my tongue what would be a very similar situation for this behavior.

It's almost like the trolley problem... switch to self driving cars and avoid the hundreds of thousands of fatal accidents caused by bad drivers, knowing that a few folks are going to die because of your decision.
Title: Re: Self-Driving Cars
Post by: brilligtove on March 03, 2016, 06:33:23 PM
"The ability to judge intention and respond accordingly is also central to driving. From determining whether a pedestrian is going to jaywalk to slowing down and avoiding a driver who seems drunk or tired, we do it constantly while behind the wheel. Self-driving cars can’t do this now. They likely won’t be able to do it for years. But this isn’t just about routine-but-confusing interactions like that between the Google self-driving car and the Mountain View bus.

Even the best AIs are easy to fool. State-of-the-art object recognition systems can be tricked into thinking a picture of an orange is really an ostrich. Self-driving cars will be no different. They will make errors—which is not so bad on the face of it, as long as they make fewer than humans. But the kinds of errors they make will be errors a human would never make. They will mistake a garbage bag for a running pedestrian. They will mistake a cloud for a truck.

All of this means that self-driving cars will be incredibly easy to troll. Sticking a foot out in the road or waving a piece of cloth around might be enough to trigger an emergency stop. Tapping the brakes of your car could trigger a chain reaction of evasive maneuvers. Perhaps a few bored 12-year-olds could shut down L.A. freeways with the equivalent of smiley faces painted on balloons."


http://www.slate.com/articles/technology/future_tense/2016/03/google_self_driving_cars_lack_a_human_s_intuition_for_what_other_drivers.html


I guess I am not the only one who has thought of this.

Read this first brillingtrove, then get back to me.

I did read it before posting, but there wasn't anything meaningful to respond to.

Tell us why you think you're credible in this discussion and I'll see if getting back to you is worth the effort.
Title: Re: Self-Driving Cars
Post by: DG on March 03, 2016, 07:31:38 PM

Moderator Comment Dear all

 :ninja:

That is all.
(click to show/hide)
Title: Re: Self-Driving Cars
Post by: Guillermo on March 04, 2016, 09:06:16 AM
Guillermo,

I suspect if given a choice, most people would prefer the drive the fuck outta there as fast as possible in a human powered car, to the sit locked in your automated car shivering waiting for a cop to hopefully arrive in 15 minutes while the guy with the bat and the tattooed knuckles is taking aim at your windows. 

But hey, at least we will still have a choice for a while.
Not really an option. The chances of actually happening to you, by the time you realize that you are being assaulted in a normal car is when you have a shotgun pointed at you through the window. This happens often enough here for me to know that you rarely have a chance to escape. In other words, there really is no much difference, it just changes the assaulter's tactics. I don't think this is an issue.
Title: Re: Self-Driving Cars
Post by: phooey on March 04, 2016, 10:21:36 AM
But if you read the article from Slate, he talks about exactly how easy it would be to make a car stop, that you could do for all kinds of nefarious reasons.  All a person would likely need to do if start to stick their leg out, and how is the computer going to know what to do other than stop? 

Think how easy it would be to mug someone in a car if this was the case.  Sit, by the side of the road, where you know someone is going to be, force the computer to stop the car, and...voila. 

How do you propose they will resolve this? 
Title: Re: Self-Driving Cars
Post by: Anders on March 04, 2016, 11:05:33 AM

Moderator Comment Dear all

 :ninja:

That is all.
(click to show/hide)

https://www.youtube.com/watch?v=OMOGaugKpzs
Title: Re: Self-Driving Cars
Post by: Morvis13 on March 04, 2016, 11:05:52 AM
But if you read the article from Slate, he talks about exactly how easy it would be to make a car stop, that you could do for all kinds of nefarious reasons.  All a person would likely need to do if start to stick their leg out, and how is the computer going to know what to do other than stop? 

Think how easy it would be to mug someone in a car if this was the case.  Sit, by the side of the road, where you know someone is going to be, force the computer to stop the car, and...voila. 

How do you propose they will resolve this?

better security.
Title: Re: Self-Driving Cars
Post by: Dan I on March 04, 2016, 12:21:40 PM
"The ability to judge intention and respond accordingly is also central to driving. From determining whether a pedestrian is going to jaywalk to slowing down and avoiding a driver who seems drunk or tired, we do it constantly while behind the wheel. Self-driving cars can’t do this now. They likely won’t be able to do it for years. But this isn’t just about routine-but-confusing interactions like that between the Google self-driving car and the Mountain View bus.

Even the best AIs are easy to fool. State-of-the-art object recognition systems can be tricked into thinking a picture of an orange is really an ostrich. Self-driving cars will be no different. They will make errors—which is not so bad on the face of it, as long as they make fewer than humans. But the kinds of errors they make will be errors a human would never make. They will mistake a garbage bag for a running pedestrian. They will mistake a cloud for a truck.

All of this means that self-driving cars will be incredibly easy to troll. Sticking a foot out in the road or waving a piece of cloth around might be enough to trigger an emergency stop. Tapping the brakes of your car could trigger a chain reaction of evasive maneuvers. Perhaps a few bored 12-year-olds could shut down L.A. freeways with the equivalent of smiley faces painted on balloons."


http://www.slate.com/articles/technology/future_tense/2016/03/google_self_driving_cars_lack_a_human_s_intuition_for_what_other_drivers.html


I guess I am not the only one who has thought of this.

You and that article make an awful lot of presumptions that seem to accept as given that there is no room for improvement in the recognition systems. That what we have now is exactly what will go into the cars 5-10 years from now.
Title: Re: Self-Driving Cars
Post by: Alex Simmons on March 04, 2016, 05:17:06 PM
But if you read the article from Slate, he talks about exactly how easy it would be to make a car stop, that you could do for all kinds of nefarious reasons.  All a person would likely need to do if start to stick their leg out, and how is the computer going to know what to do other than stop? 

Think how easy it would be to mug someone in a car if this was the case.  Sit, by the side of the road, where you know someone is going to be, force the computer to stop the car, and...voila. 

How do you propose they will resolve this?

Simple. Permit the occupant to resume control of the vehicle.
Title: Re: Self-Driving Cars
Post by: Johnny Slick on March 04, 2016, 05:21:38 PM
But if you read the article from Slate, he talks about exactly how easy it would be to make a car stop, that you could do for all kinds of nefarious reasons.  All a person would likely need to do if start to stick their leg out, and how is the computer going to know what to do other than stop? 

Think how easy it would be to mug someone in a car if this was the case.  Sit, by the side of the road, where you know someone is going to be, force the computer to stop the car, and...voila. 

How do you propose they will resolve this?

Simple. Permit the occupant to resume control of the vehicle.
Well, to be fair, Google wants to get rid of steering wheels altogether. I still don't think that the threat of roving bands of gangs sticking their legs out at cars passing by in the hopes that the driver won't be a "wheelie" and pull a Bonfire of the Vanities on them is a high-risk eventuality that we need to plan for. In related news I also don't think that cars of the future should necessarily be programmed to account for giant, iced-over balls of poop that fall from airplanes a la Joe Dirt, or a random assortment of baseball players who want to fix the 1919 World Series a la Eight Men Out. No, I for one accept that these are items that are known as "movies", which by and large are not "real" even though the things that happen in them are things that could happen in real life.
Title: Re: Self-Driving Cars
Post by: Friendly Angel on March 04, 2016, 05:22:18 PM
But if you read the article from Slate, he talks about exactly how easy it would be to make a car stop, that you could do for all kinds of nefarious reasons.  All a person would likely need to do if start to stick their leg out, and how is the computer going to know what to do other than stop? 

Think how easy it would be to mug someone in a car if this was the case.  Sit, by the side of the road, where you know someone is going to be, force the computer to stop the car, and...voila. 

How do you propose they will resolve this?

Simple. Permit the occupant to resume control of the vehicle.

I was thinking more along the lines of a panic button, maybe to activate flame throwers and spew tear gas.
Title: Re: Self-Driving Cars
Post by: amysrevenge on March 04, 2016, 05:54:28 PM
But if you read the article from Slate, he talks about exactly how easy it would be to make a car stop, that you could do for all kinds of nefarious reasons.  All a person would likely need to do if start to stick their leg out, and how is the computer going to know what to do other than stop? 

Think how easy it would be to mug someone in a car if this was the case.  Sit, by the side of the road, where you know someone is going to be, force the computer to stop the car, and...voila. 

How do you propose they will resolve this?

Simple. Permit the occupant to resume control of the vehicle.

I was thinking more along the lines of a panic button, maybe to activate flame throwers and spew tear gas.

Oh yeah, hot shot?  What about when the gangs of leg-sticker-outers wear those crinkly metal flame-resistant suits and gas masks?  Huh?  SLAM DUNK IN YOUR FACE SUCKA checkmate.
Title: Re: Self-Driving Cars
Post by: arthwollipot on March 04, 2016, 06:03:59 PM
Seriously - diverting to avoid someone's stuck out leg wouldn't even take the car out of its lane.

This is a silly argument.
Title: Re: Self-Driving Cars
Post by: SkeptiQueer on March 04, 2016, 08:14:52 PM
"Stand and deliver" he yelled, proudly leaping in front of the car and waiting for it to stop.

"Thwuckrunch" his bones yelled, as the bumper turned his pelvis into powder.

"Third one this week" said the driver, turning the wheel and pulling his car over to call the police.

"Self-driving cars cause rash of suicides" said the anchor. "More at 11."
Title: Re: Self-Driving Cars
Post by: phooey on March 04, 2016, 09:53:46 PM
Oh, this is great news.  It won't happen!  People won't try to stop people on the road to steal, or to get revenge for a breakup or to rape some girls, or just to fuck with the traffic system because of rebellion, or because they are young kids that are bored, or because they are terrorists, or just panhandlers looking for money. Thank goodness this will never happen.  A panhandler won't keep a car prisoner until he is given ten bucks, because, they just won't do that. Phew!

I was worried that we might need a solution to all these issues before we could have automated cars, but it turns out we don't need a solution to prevent chaos. Lordie I am glad of that.

The programmers answer to this problem-It won't happen.  Stop worrying!!  Hallelujah. 

 
Title: Re: Self-Driving Cars
Post by: Johnny Slick on March 04, 2016, 10:01:36 PM
Guys, what if ninjas who are dressed like sailors jump on top of your self driving car while it is stopped in between a red and a green light? I'll tell you what. SYNTAX ERROR DOES NOT COMPUTE BSOD
Title: Re: Self-Driving Cars
Post by: Alex Simmons on March 05, 2016, 02:37:52 AM
But if you read the article from Slate, he talks about exactly how easy it would be to make a car stop, that you could do for all kinds of nefarious reasons.  All a person would likely need to do if start to stick their leg out, and how is the computer going to know what to do other than stop? 

Think how easy it would be to mug someone in a car if this was the case.  Sit, by the side of the road, where you know someone is going to be, force the computer to stop the car, and...voila. 

How do you propose they will resolve this?

Simple. Permit the occupant to resume control of the vehicle.
Well, to be fair, Google wants to get rid of steering wheels altogether.

Well I doubt I'd buy one then. Not because of any particular fear of not having control but simply because driving is at times an enjoyable activity. So for instance when rolling along some nice roads I expect I'd prefer to drive because it's fun (and at that stage the standard driving aids will be pretty sophisticated - some of them already are in modern cars) but if dealing with city traffic or a boring freeway, then let the car do its thing.
Title: Re: Self-Driving Cars
Post by: arthwollipot on March 05, 2016, 09:12:36 PM
Oh, this is great news.  It won't happen!  People won't try to stop people on the road to steal, or to get revenge for a breakup or to rape some girls, or just to fuck with the traffic system because of rebellion, or because they are young kids that are bored, or because they are terrorists, or just panhandlers looking for money. Thank goodness this will never happen.  A panhandler won't keep a car prisoner until he is given ten bucks, because, they just won't do that. Phew!

Wow, do you really live in a place where all of those things happen on a regular basis? What do you do about them now?
Title: Re: Self-Driving Cars
Post by: brilligtove on March 06, 2016, 12:15:06 AM
At least two of our beloved live in South Africa, where flamethrowes and ankle scythes are not unknown AFAIK. A self driving car in Norway, the Sudan, and South Africa will all have to deal with significant challenges in that environment and culture. At the same time, I expect these challenges to be overcome by technology developing over the coming years.
Title: Re: Self-Driving Cars
Post by: phooey on March 06, 2016, 01:27:51 AM
arthwollipot,

You have never been to East LA?  How about the Mission district in San Francisco?  Downtown Denver? 

Do you drive? 
Title: Re: Self-Driving Cars
Post by: phooey on March 06, 2016, 01:40:24 AM
brillingtrove,

I appreciate your confidence in technologies ability to overcome this problem. 

But at this point you can admit that you have no idea what that technology will be right?  A guys wants to stop your car.  You don't want it to stop.  There is no steering wheel.  That's not an easy one to overcome yeah?

Are you in the flame throwing car camp at this point?   
Title: Re: Self-Driving Cars
Post by: arthwollipot on March 06, 2016, 01:43:52 AM
arthwollipot,

You have never been to East LA?  How about the Mission district in San Francisco?  Downtown Denver? 

Do you drive?

Dodging the question by asking questions. Nice.

When was the last time someone stopped you on the road to get revenge for a breakup?
Title: Re: Self-Driving Cars
Post by: phooey on March 06, 2016, 01:55:19 AM
What question am I supposed to answer?  What do I do now when someone wants to stop my car but I don't want to stop?  I don't stop my car.  Because I am not a computer. 

Now what's your solution?  "Ah, this won't happen much, nevermind" ?
Title: Re: Self-Driving Cars
Post by: arthwollipot on March 06, 2016, 02:03:25 AM
What question am I supposed to answer?

How about the one (ones) I asked? When was the last time someone tried to force you to stop your car?
Title: Re: Self-Driving Cars
Post by: phooey on March 06, 2016, 07:19:54 AM
Oh my goodness, I need to figure out what level of person I am dealing with.  people don't have the ability to just stop my car by standing next to it, get it?  So if someone wanted to stop my car, I wouldn't know, because I have the ability to keep on driving anytime I want!  You see??

But I can tell, you I have had lots of lots of people come up to my car at a red light and ask for change, or just start to clean my windshield and then ask for money.  But guess what, I can drive away from them if I choose!  But in the new game, the computer can't just drive away from some right next to their car, because the computer doesn't know why that person is right next to their car! 

Why does it seem like you are struggling with understanding the most basic of concepts here.  If I am a bad guy, and I pretend I am going to charge right in front of your vehicle, how in the world does the computer know not to do the safe thing, and slow down or stop to avoid hitting something? 

Please next time you post, can't you at least think about what is being said? 
Title: Re: Self-Driving Cars
Post by: Captain Video on March 06, 2016, 02:30:25 PM
you need  "a car that just wont stop" Sorry I couldn't resist, I have been thinking of this character ever since you joined.

https://www.youtube.com/watch?v=-n91lSO0wJQ&index=3&list=PLUJxBbj6p4NF7F5I9aDWakk8cMSi8SF2G

Could it be that the car asks you what to do at the point some idiot decides to wash your windows? Anti personal devices could be installed like spraying windshield wiper fluid (or pepper spray) in all directions but i guess you wont need windshield wipers any more so i doubt bums will try to wash your windows nobody will pay. Perhaps the annoying concept of roadside solicitation will finally be put to an end, I yell at those people out all the time especially idiot parents dangerously standing roadside with kids collecting money for whatever.

the next time you post try using some imagination, life as YOU know it will change.
Title: Re: Self-Driving Cars
Post by: PANTS! on March 06, 2016, 02:33:34 PM
Did phooey just confess to mowing down hobos in his car?

Is that why he is so affronted by self-driving cars that won't mow down window washers?

Title: Re: Self-Driving Cars
Post by: DG on March 06, 2016, 04:29:50 PM
Oh my goodness, I need to figure out what level of person I am dealing with.  people don't have the ability to just stop my car by standing next to it, get it?  So if someone wanted to stop my car, I wouldn't know, because I have the ability to keep on driving anytime I want!  You see??

But I can tell, you I have had lots of lots of people come up to my car at a red light and ask for change, or just start to clean my windshield and then ask for money.  But guess what, I can drive away from them if I choose!  But in the new game, the computer can't just drive away from some right next to their car, because the computer doesn't know why that person is right next to their car! 

Why does it seem like you are struggling with understanding the most basic of concepts here.  If I am a bad guy, and I pretend I am going to charge right in front of your vehicle, how in the world does the computer know not to do the safe thing, and slow down or stop to avoid hitting something? 

Please next time you post, can't you at least think about what is being said?

You make an interesting point - what would I do if someone engaged in conduct that was calculated to FORCE me to stop. Arthwolipot makes a good point - there are already ways that a person can force you top stop.

If your issue is that "the car wont know why someone is next to the door" - that seems like an issue of coding - simply teaching the car that if someone is in the drivers seat, the driver's door should be locked. My current car already does something similar. Equally the same can be applied to ensure that the only door with a proximity sensor is the drivers door (as is the case with my car). It seems that the "person along side the car" isn't really an issue. I don't understand why you think that it would be an issue. This seems to deal with someone is next to the car - whether it's a loved one saying goodbye or someone trying to convince you to pay them for smearing dirty water over your windscreen.

If your concern is that "someone might try to hijack my car" (keeping in mind this hypothetical car is readily capable of being tracked, that will avoid accidents so they can't use it for a ram raid or anything of that nature, and can readily be shut down remotely and so on). I'm sure that you'd agree that persons who wish to cause you to stop your car already have means to do so (I'm sure you'd agree that this is the case) if someone steps onto the road in front of you, you are legally required to attempt to avoid the collision now. if someone is standing in the middle of the road you're not allowed to just run them down in case they are trying to steal your car. By the time they have drawn a weapon to take the car, current best practice and advice from law enforcement is to "just give them the car - it's not worth dying over" (this seem to be doubly true with the features referred to above that significantly reduce the benefit of having said stolen car).

Of course, if you're taking the view "I should be able to run someone down if I am worried that they pose a threat to me" - please say so clearly. 

This really leave the situation in which a person desires to injure the driver of the vehicle, again, this is already possible.

Title: Re: Self-Driving Cars
Post by: phooey on March 06, 2016, 06:42:26 PM
First, this is not directed at anyone in particular, but I can not believe the level of thought here.  It seems people aren't even trying to imagine all the possible scenarios that evolve around a computer not knowing what things are safe to stop for, and what is an interruption or interference from someone.

Now DG, you are saying right now its fairly easy for someone to get someone to stop their car, if they want to hassle them? I am not sure why you think that is easy right now, if the person is at all concerned about maybe getting their leg ran over. But you have more importantly completely missed a scenario that I think I explained rather clearly.  Its not only about someone wanting to steal your car.  If I am a beggar on the street (this is just one example of 1000 of people who do not have good intent) I can keep your car from moving until you give me ten bucks.  All I have to do is walk up to your car at night, while it is at a red light.  Now I say, your computer is not programmed to run over my foot, so please give me ten bucks to be on your way.  What's the fricking computer going to do about this?

The latest proposal seems to be to load your car with flame throwers or pepper spray which sprays out from any angle of your car apparently.  I guess that would include going towards the motorcyclist who is in the lane next to your car. 

But the larger point now to me is, what is with skeptics?? Is this the reasoning powers of the average skeptic?  Holy shit. 

Title: Re: Self-Driving Cars
Post by: brilligtove on March 06, 2016, 06:53:15 PM
It seems clear that you don't know what either "skeptic" or "reasoning" mean. What you have been facing on this thread is skeptical and well reasoned opposition to your absurd and irrational pronouncements. Also some joking and derision, because the reasoned approach is not getting through.

The problems you raise with self-driving cars are either identical to problems with cars themselves, or are based in a broad and deep ignorance of engineering, robotics, and machine intelligences.

Ignorance can be cured, of course, and we've tried to help with that. You need to want to be cured for it to work, though.
Title: Re: Self-Driving Cars
Post by: SkeptiQueer on March 06, 2016, 07:00:40 PM
We're not convinced that already extremely rare scenarios are proof that self-driving cars can't ever work. That's pretty much it.

Flame throwers aren't necessary, and there's no reason pepper spray has to spray in all directions stupidly. You can't criticize someone for not thinking of 1000 unlikely scenarios where you might need the ability to drive like an idiot, and then completely lose your ability to think critically when it comes to working around those scenarios.

Defensive measures in the vehicle (130db alarm works pretty well, so does just plain water if all you need is someone to take a step back, pepper spray if kidnapping is a risk, or carry a gun like I do) but there's also the option of a big red "emergency" button. Push button, car ignores some protocols, like running over a foot or whipping around someone instead of moving deliberately. Just like anything else, if you cause an accident while using the emergency button, you're liable instead of the company that insures your car. If someone wants to play the extortion game, call the police. They won't be running speed traps or dealing with reckless accidents, so response times should be decent.

Come on, you can think of all these reasons why it can't work, but won't apply a modicum of brainpower to solving those issues.
Title: Re: Self-Driving Cars
Post by: phooey on March 06, 2016, 07:15:25 PM
Just give everyone a gun.  Hey great idea. 

Let me guess, you are American? 

So far the arm everyone to the teeth experiment in America doesn't seem to be working out so good.  I guess that's why pretty much every other country has decided against that particular strategy. 
Title: Re: Self-Driving Cars
Post by: phooey on March 06, 2016, 07:20:40 PM
Skeptiqueer,

When I am on my motorcycle next to you, when you decide to blast your little air horn, I think I will just go ahead and shot you. 

Just to be safe. 
Title: Re: Self-Driving Cars
Post by: Captain Video on March 06, 2016, 09:21:15 PM
Out of curiosity what country do you live in Phooey? 

I assume programming would be unique in different countries, I have seen video of how they drive in India for instance that gives me a heart attack just watching.

Title: Re: Self-Driving Cars
Post by: Alex Simmons on March 06, 2016, 10:57:09 PM
First, this is not directed at anyone in particular, but I can not believe the level of thought here.  It seems people aren't even trying to imagine all the possible scenarios that evolve around a computer not knowing what things are safe to stop for, and what is an interruption or interference from someone.

Now DG, you are saying right now its fairly easy for someone to get someone to stop their car, if they want to hassle them? I am not sure why you think that is easy right now, if the person is at all concerned about maybe getting their leg ran over. But you have more importantly completely missed a scenario that I think I explained rather clearly.  Its not only about someone wanting to steal your car.  If I am a beggar on the street (this is just one example of 1000 of people who do not have good intent) I can keep your car from moving until you give me ten bucks.  All I have to do is walk up to your car at night, while it is at a red light.  Now I say, your computer is not programmed to run over my foot, so please give me ten bucks to be on your way.  What's the fricking computer going to do about this?

The latest proposal seems to be to load your car with flame throwers or pepper spray which sprays out from any angle of your car apparently.  I guess that would include going towards the motorcyclist who is in the lane next to your car. 

But the larger point now to me is, what is with skeptics?? Is this the reasoning powers of the average skeptic?  Holy shit.
I get the scenario you speak of, and in my view the self drive vehicle should always have a manual operational mode, and the occupant be capable of actually driving as necessary.
e.g. at present when car is in cruise control mode but I tap the brake or tap a button on the controls, the throttle control immediately reverts to me, the driver.

Making such a car immediately revert control to the occupant as soon as they choose to is all that is required.

Even starships in sci-fi all have manual flight control mode, I see no reason why that should not also be the case for self drive cars of the future.
Title: Re: Self-Driving Cars
Post by: DG on March 06, 2016, 11:08:15 PM
First, this is not directed at anyone in particular, but I can not believe the level of thought here.  It seems people aren't even trying to imagine all the possible scenarios that evolve around a computer not knowing what things are safe to stop for, and what is an interruption or interference from someone.

Now DG, you are saying right now its fairly easy for someone to get someone to stop their car, if they want to hassle them? I am not sure why you think that is easy right now, if the person is at all concerned about maybe getting their leg ran over. But you have more importantly completely missed a scenario that I think I explained rather clearly.  Its not only about someone wanting to steal your car.  If I am a beggar on the street (this is just one example of 1000 of people who do not have good intent) I can keep your car from moving until you give me ten bucks.  All I have to do is walk up to your car at night, while it is at a red light.  Now I say, your computer is not programmed to run over my foot, so please give me ten bucks to be on your way.  What's the fricking computer going to do about this?

What are you going to do about this now? Running over their foot could well constitute a serious assault - assault with a vehicle. By your own reckoning they don't pose a threat, just an inconvenience. I hope you're not suggestion that you should have the right to run over such a person? Or, perhaps you are. I am unsure based on your position. It seems to me that either you
(a) agree that you're not allowed to run over someone's foot if it's in your way (in which case you are in the same whether or not the care is automated),
(b) that you believe you should be allowed to use, or have the threat of using, a vehicle to cause harm to a person that seeks to stop you to beg money etc (I'm not sure we are starting in the same position of morality if you think that the threat of the use of force to someone begging for money and does not pose an immediate threat is anything other than assault).


I am disappointed that you don't see the ability to build rules that are comparable to the current laws that are binding on all drivers as a reasonable solution to this issue. I am certainly not convinced that allowing drivers to make decisions based on their own assessment is the best way of determining the best cause of action. I think you'll agree that humans can, and do, make suboptimal decisions in emergencies. Allowing that sometimes people make good decisions in emergencies, gets us no where near saying "it's not viable for a computer to make the decision because they may make a decision that I would not have made".

After all. A significant strength of computerised vehicles is to take the decision making away from humans. Having unpredictable drivers and unpredictable pedestrians sharing a space is a problem that could do with fixing. WE are unlikely to fix pedestrians, but taking some of the unpredictable behaviour away from drivers may be beneficial.

That said, I do not understand why you elect to ignore the more serious suggestion of an override that would allow manual control in those circumstances and focus on what only can be thought of as tongue-in-cheek suggestions of tear gas and the likes. Further, manual control of a vehicle could be instantly reported to the authorities as demonstration of some kind of hazard or threat that requires intervention by the authorities, with a penalty for misusing the manual over ride, akin to misusing the emergency brake on a train.

That said, I really don't think that I would like a computerised car. Hell, I don't like it my wife is driving, nor a taxi driver or basically anyone else. I prefer to have control of the vehicle myself. I generally would rather walk than catch a taxi (but that's mostly because I see my car as a private personal space that I don't like sharing with others and getting in a taxi feels like invading their space in a way that I am not comfortable with. For those reasons, more so than technical difficulties, I think you'll have trouble convincing people to share cars or give up that 'personal' control over their own little space.
Title: Re: Self-Driving Cars
Post by: SkeptiQueer on March 06, 2016, 11:28:19 PM
Skeptiqueer,

When I am on my motorcycle next to you, when you decide to blast your little air horn, I think I will just go ahead and shot you. 

Just to be safe.
Cool. That's assault with a deadly weapon/attempted murder, whereas blowing a 130db horn is just normal traffic. Truck horns are 150dB (permanent hearing damage at 3ft) and the pipes on my Harley, back in the day, were between 110 and 120dB, enough to set off certain car alarms if I opened the butterfly valve. What is it with people who don't own a gun turning into a complete psychopath at the thought of carrying a gun? Do you normally dismount and try to strangle someone who blows their horn at you?

You mention guns and America, but currently we have about 10,000 firearm homicides (that's all non-suicide gun deaths, including justifiable homicide) per year. It's generally a bit lower, 9200 or whatever, but 10k is close enough. We also have 32,000 traffic deaths every year. Again, not suicides, that's people killed by a car driven by a person. Clearly, we're better off switching every country to American gun laws and self-driving cars. Even if we just cut traffic deaths in half, say Google cars just cause half as many fatal accidents (which is easy, considering DUIs are no longer a thing) we've saved 15,000 lives in the US. Even if 5,000 additional people die in attempted carjackings, we have a societal net gain in the cost of insurance, healthcare, and vehicle repair.

Silly math problem aside, you nicely demonstrated my point when you took the gun bait. You're accusing us of being blind and stupid, but you're jumping at any chance to declare automated vehicles impossible. You've ignored anti-carjacking measures already in place or assumed that they function as if designed by a villain in a Pixar movie, spraying flames and OC gas in a 360 when someone tries to panhandle, or suddenly starting and gunfight when someone blows their horn. You've gone to invent ridiculous scenarios (semi trucks barreling out of the fog while husband and wife argue about directions) and generally ignored the realities of the programming and safety.

Nobody is saying the things are perfect now. Tesla is saying 2 years, right? 2 years was the difference between the Osprey V22 falling out of the air and successfully delivering SF teams onto rooftops via a SPIES rig like a Blackhawk. 2 years was the difference between Bush and Obama. 2 years was the difference between Google cars getting trapped in parking lots and cleverly navigating the streets of Monterey during school hours or bar traffic and encountering children and bicycles. A lot can happen in 2 years. It is blind hubris to think a problem you just thought of is proof that the problem cannot be solved.

Still waiting on that evidence that pilots land better than autoland.
Title: Re: Self-Driving Cars
Post by: SkeptiQueer on March 07, 2016, 01:00:55 AM
A couple things that bear noting.

Google will likely aim for cheap commuters, the equivalent to the Android. Functional, lower end. Apple/Tesla will aim for your high end, and I would wager might have some models that retain the capability for human control. Minority Report and iRobot might have been way off in having cars travelling at 200mph in a commute, but higher-end cars retaining the capability to be manually operated were something that I would expect. I would also expect manual operation to cost a whole lot more in insurance. 

I remember having a conversation about two years ago at a bar here in the city with several Boeing engineers, a cop, an accountant, and a lawyer. When I say having a conversation, I was mostly listening as the people who designed UAV programs talked about the problems they'd solved with programming a UAV that could be directed or remotely controlled by a college drop-out without stalling, crashing, or hitting other aircraft. The insurance thing came up. And about a year later I was reading about the lower cost of insurance and accidents that would be expected. Someone (non-UAV code monkey) at the time mentioned the idea of having company-owned vehicles that could be summoned by text message. Google Wallet was starting to really ramp up, and the idea of texting for a car, having it drive you somewjere, and automatically processing payment when you got out was ridiculously mind-blowing. That helps solve the liability issue for the passengers as well.

Honestly the biggest problem I expect to see will be with bad actors trying to take control either remotely or without paying.
Title: Re: Self-Driving Cars
Post by: Alex Simmons on March 07, 2016, 02:23:49 AM
But the larger point now to me is, what is with skeptics?? Is this the reasoning powers of the average skeptic?  Holy shit.
Well it's a fallacy to assume all people on here are skeptics. They are members of a forum. That doesn't make them all skeptics.

Secondly, even if your assertion that the reasoning powers of people you've barely interacted with are lacking is true (which I doubt it is), it's only matched by your apparent lack of ability to adequately communicate your own thoughts in a rational and reasonable manner, to engage others in a constructive discussion or debate, or to listen to what others have been saying and thoughtfully respond to their points.
Title: Re: Self-Driving Cars
Post by: Captain Video on March 07, 2016, 04:42:08 AM
First, this is not directed at anyone in particular, but I can not believe the level of thought here.  It seems people aren't even trying to imagine all the possible scenarios that evolve around a computer not knowing what things are safe to stop for, and what is an interruption or interference from someone.

Now DG, you are saying right now its fairly easy for someone to get someone to stop their car, if they want to hassle them? I am not sure why you think that is easy right now, if the person is at all concerned about maybe getting their leg ran over. But you have more importantly completely missed a scenario that I think I explained rather clearly.  Its not only about someone wanting to steal your car.  If I am a beggar on the street (this is just one example of 1000 of people who do not have good intent) I can keep your car from moving until you give me ten bucks.  All I have to do is walk up to your car at night, while it is at a red light.  Now I say, your computer is not programmed to run over my foot, so please give me ten bucks to be on your way.  What's the fricking computer going to do about this?

The latest proposal seems to be to load your car with flame throwers or pepper spray which sprays out from any angle of your car apparently.  I guess that would include going towards the motorcyclist who is in the lane next to your car. 

But the larger point now to me is, what is with skeptics?? Is this the reasoning powers of the average skeptic?  Holy shit.
I get the scenario you speak of, and in my view the self drive vehicle should always have a manual operational mode, and the occupant be capable of actually driving as necessary.
e.g. at present when car is in cruise control mode but I tap the brake or tap a button on the controls, the throttle control immediately reverts to me, the driver.

Making such a car immediately revert control to the occupant as soon as they choose to is all that is required.

Even starships in sci-fi all have manual flight control mode, I see no reason why that should not also be the case for self drive cars of the future.

I disagree, I don't think people are trained well enough to drive, when people start "piloting" vehicles then I would agree but they don't they drive and that's why we have accidents. BTW some people are anti-gun, i'm anti-car, I don't trust you behind the wheel, you are dangerous and so am I.  As cars become more automated, manual drive licencing should have much stricter learning and testing requirements then eventually eliminated. 

"Most" vehicle drivers do not do a safety inspection of a car before driving. for example, How many of you walk around the car to check the tires before you get in?  How many of you are willing to drive a car with an under pressured tire?  We have few plane crashes because they are inspected, the pilot is well trained and performs a safety checklist before flight, most of the tough decisions are made by a guy thats not even on the plane directing traffic.

I'm not trying to make a comparison between cars and guns being safer, (sorry SkeptiQueer), i'll fight for your right to have a gun but I don't trust you with it. However I do trust you not to accidentally shoot me with that gun a hell of a lot more than I trust you not to accidentally crash into me while behind the wheel.  There is rarely if ever a domino effect when someone shoots a gun unless you are in a spaghetti western. A computer car network would see the whole picture of what was happening and be able to minimize if not eliminate the domino effect in a car crash.

As cars become smarter traffic control can be handled by computers which would not only be safer its eliminating the massive expense of everything it takes to manage a system of roads like traffic lights, street lights, traffic signs, road paint, etc as well as the expense of patrolling and enforcing traffic rules.   If you keep letting people drive manually those things will still be needed.
Title: Re: Self-Driving Cars
Post by: SkeptiQueer on March 07, 2016, 06:45:25 AM
I only brought up the gun thing because it seemed to get a really strong reaction from the anti-safer-vehicle advocate.

There's just this tacit acceptance of the number of deaths because of shitty drivers, and it pisses me off. I just checked NTHSA again. 3100+ deaths last year from distracted driving, 800+ deaths from drowsy driving. About (year depending, more or less) 10,000 impaired driving (drugs or alcohol) deaths and $37 billion dollars each year, between medical care (which we all pay for), the justice system, lost wages, and insurance (which we all pay for) because people drive drunk or high.

How many people will have a beer or two with dinner and hen drive? Take cold medicine? Most people, even if they can afford pain medicine, can't afford to not drive for 6 months while they rehab from a spinal surgery, and that's an impaired driver.

Texting is an easy demon to name, but all you fuckers (pun intended) with all your having kids and needing to take them places are at a high risk for distracted driving. Little poopsmith screaming in the back, radio's on, boom you're distracted even if you're not trying to give it back its binky. Eating and driving, jamming out to the radio, whatever.

Unless we're worried there are going to suddenly be bands of highwaymen holding up cars again and killing 20,000 more people in the US every year, there's a really really strong case for not just adjusting to allow self-driving cars, but pushing for their acceptance and investing government funding in them and their infrastructure.
Title: Re: Self-Driving Cars
Post by: Captain Video on March 07, 2016, 01:56:34 PM
I am 100% behind you on this issue SkeeptiQueer, I only added the little gun note as a reminder that i dont understand but still stand behind your rights on the gun issue, The car issue drives me nuts! (see what I did there)

Everything you said plus the damage they do to the environment and our lungs, Its one big nightmare that most people seem to accept as normal.
Title: Re: Self-Driving Cars
Post by: Fast Eddie B on March 07, 2016, 02:03:57 PM
Late to this thread.

Was this recent accident mentioned?

http://www.wired.com/2016/02/googles-self-driving-car-may-caused-first-crash/ (http://www.wired.com/2016/02/googles-self-driving-car-may-caused-first-crash/)

I only know about it from yesterday's Tech News Today.
Title: Re: Self-Driving Cars
Post by: SkeptiQueer on March 07, 2016, 02:47:20 PM
Late to this thread.

Was this recent accident mentioned?

http://www.wired.com/2016/02/googles-self-driving-car-may-caused-first-crash/ (http://www.wired.com/2016/02/googles-self-driving-car-may-caused-first-crash/)

I only know about it from yesterday's Tech News Today.
Yeah, it came up a few days ago. Consensus was that the car assumed the bus would yield to a lane change in progress. The bus didn't.
Title: Re: Self-Driving Cars
Post by: Morvis13 on March 07, 2016, 03:02:23 PM
Late to this thread.

Was this recent accident mentioned?

http://www.wired.com/2016/02/googles-self-driving-car-may-caused-first-crash/ (http://www.wired.com/2016/02/googles-self-driving-car-may-caused-first-crash/)

I only know about it from yesterday's Tech News Today.
Yeah, it came up a few days ago. Consensus was that the car assumed the bus would yield to a lane change in progress. The bus didn't.

Can you imagine if we had this kind of headlines every time a human caused a crash?
Title: Re: Self-Driving Cars
Post by: amysrevenge on March 07, 2016, 03:11:47 PM
Late to this thread.

Was this recent accident mentioned?

http://www.wired.com/2016/02/googles-self-driving-car-may-caused-first-crash/ (http://www.wired.com/2016/02/googles-self-driving-car-may-caused-first-crash/)

I only know about it from yesterday's Tech News Today.
Yeah, it came up a few days ago. Consensus was that the car assumed the bus would yield to a lane change in progress. The bus didn't.

Can you imagine if we had this kind of headlines every time a human caused a crash?

Bringing back gun analogies (ugh, sorry), it would be kinda like in Canada when we see headlines like "some guy almost got shot 2000km away from you" that I get sometimes in my local news feed.

ETA:  In self-driving car land, national news:  "Panic in the streets of Dallas today, as an attempted mugging almost turned into a 20th century traffic accident, when the would-be mugger was gently nudged by a Googlematic 9000 as it performed its standard dodge and retreat maneuver."
Title: Re: Self-Driving Cars
Post by: Alex Simmons on March 07, 2016, 04:18:28 PM
First, this is not directed at anyone in particular, but I can not believe the level of thought here.  It seems people aren't even trying to imagine all the possible scenarios that evolve around a computer not knowing what things are safe to stop for, and what is an interruption or interference from someone.

Now DG, you are saying right now its fairly easy for someone to get someone to stop their car, if they want to hassle them? I am not sure why you think that is easy right now, if the person is at all concerned about maybe getting their leg ran over. But you have more importantly completely missed a scenario that I think I explained rather clearly.  Its not only about someone wanting to steal your car.  If I am a beggar on the street (this is just one example of 1000 of people who do not have good intent) I can keep your car from moving until you give me ten bucks.  All I have to do is walk up to your car at night, while it is at a red light.  Now I say, your computer is not programmed to run over my foot, so please give me ten bucks to be on your way.  What's the fricking computer going to do about this?

The latest proposal seems to be to load your car with flame throwers or pepper spray which sprays out from any angle of your car apparently.  I guess that would include going towards the motorcyclist who is in the lane next to your car. 

But the larger point now to me is, what is with skeptics?? Is this the reasoning powers of the average skeptic?  Holy shit.
I get the scenario you speak of, and in my view the self drive vehicle should always have a manual operational mode, and the occupant be capable of actually driving as necessary.
e.g. at present when car is in cruise control mode but I tap the brake or tap a button on the controls, the throttle control immediately reverts to me, the driver.

Making such a car immediately revert control to the occupant as soon as they choose to is all that is required.

Even starships in sci-fi all have manual flight control mode, I see no reason why that should not also be the case for self drive cars of the future.

I disagree, I don't think people are trained well enough to drive, when people start "piloting" vehicles then I would agree but they don't they drive and that's why we have accidents. BTW some people are anti-gun, i'm anti-car, I don't trust you behind the wheel, you are dangerous and so am I.  As cars become more automated, manual drive licencing should have much stricter learning and testing requirements then eventually eliminated. 

"Most" vehicle drivers do not do a safety inspection of a car before driving. for example, How many of you walk around the car to check the tires before you get in?  How many of you are willing to drive a car with an under pressured tire?  We have few plane crashes because they are inspected, the pilot is well trained and performs a safety checklist before flight, most of the tough decisions are made by a guy thats not even on the plane directing traffic.

I'm not trying to make a comparison between cars and guns being safer, (sorry SkeptiQueer), i'll fight for your right to have a gun but I don't trust you with it. However I do trust you not to accidentally shoot me with that gun a hell of a lot more than I trust you not to accidentally crash into me while behind the wheel.  There is rarely if ever a domino effect when someone shoots a gun unless you are in a spaghetti western. A computer car network would see the whole picture of what was happening and be able to minimize if not eliminate the domino effect in a car crash.

As cars become smarter traffic control can be handled by computers which would not only be safer its eliminating the massive expense of everything it takes to manage a system of roads like traffic lights, street lights, traffic signs, road paint, etc as well as the expense of patrolling and enforcing traffic rules.   If you keep letting people drive manually those things will still be needed.
I'm not sure the plane or the gun analogies help because there are clear differences (to be fair though I've not examined them closely in this context), but anyway, if cities placed human life at the top of the priority tree over convenience, well there would be a far more bicycle friendly culture in places like the USA, UK and Australia than there actually is. Indeed at times it's actively hostile. Self drive cars can play a huge role in improving this but it still leaves the issues of health/obesity, pollution, air quality and congestion (although better managed by computers, at some stage the number of vehicles will still cause congestion not matter what system is used for driving them).

This is what one city/country did to make the transition from motor vehicle killing machines dominating the streets, to having people dominate the streets:

https://www.youtube.com/watch?v=DDLb6biq39A

Annual deaths of children by motor vehicles dropped from over 400 in 1970 to 14 in 2010. That's approx 3000 to 4000 lives saved, and young lives at that making the payback approximately 250,000 person-years. Not to mention the big savings to environment, air quality, health and well being. Replace the remaining motor vehicle drivers with self drive vehicles and in such a location one may well be able to drive that stat to almost zero child deaths while maintaining all the large societal benefits to health and well being, and the environment.
Title: Re: Self-Driving Cars
Post by: Alex Simmons on March 07, 2016, 05:23:08 PM
(http://45.media.tumblr.com/ea5ad28788d75b33fdca215c4560b4e0/tumblr_o2dp32QZBu1ux3xbyo1_500.gif)
Title: Re: Self-Driving Cars
Post by: Fast Eddie B on March 07, 2016, 05:36:53 PM

Yeah, it came up a few days ago. Consensus was that the car assumed the bus would yield to a lane change in progress. The bus didn't.

Sorry - I didn't have the patience to read all the prior posts.

At least now I'm subscribed to a good discussion on the topic.

As an aside, that same episode of This Week In Tech had a good segment on the latest in drone technology - which is pretty amazing in its own right.
Title: Re: Self-Driving Cars
Post by: phooey on March 17, 2016, 08:41:19 AM
Robotics Expert: Self-Driving Cars Not Ready for Deployment

http://abcnews.go.com/Politics/wireStory/robotics-expert-driving-cars-ready-deployment-37667633

"The cars aren't yet able to handle bad weather, including standing water, drizzling rain, sudden downpours and snow, Missy Cummings, director of Duke University's robotics program, told the Senate commerce committee. And they certainly aren't equipped to follow the directions of a police officer, she said."

"The dangers: Research shows that self-driving cars can be paralyzed using a $60 device similar to a laser pointer in combination with an Arduino computer, Cummings said. Autonomous vehicles are also susceptible to GPS jammers, which drivers sometimes keep in the back of their cars. Using these jammers, hackers can take control of a self-driving vehicle."

Well, I am sure someone will come along and explain how a few hijacked cars is nothing compared to how many people are killed on the roads now. And really, how often does it actually rain, anyway.   

No problem, no problem.  Besides who needs a jammer, when you can do the same thing with a spray bottle and a bucket asking for money?  Still better than real drivers, right? 



Title: Re: Self-Driving Cars
Post by: brilligtove on March 17, 2016, 09:49:41 AM
Nah, I'd respond with facts and arguments related to the statements in the article.
Title: Re: Self-Driving Cars
Post by: starnado on March 17, 2016, 09:50:55 AM
(http://45.media.tumblr.com/ea5ad28788d75b33fdca215c4560b4e0/tumblr_o2dp32QZBu1ux3xbyo1_500.gif)

OMFG how in the world did s/he survive? Terrifying.
Title: Re: Self-Driving Cars
Post by: brilligtove on March 17, 2016, 10:08:57 AM
1. Difficult Conditions
Quote
The cars aren't yet able to handle bad weather, including standing water, drizzling rain, sudden downpours and snow...
Agreed. Self-driving is not ready to handle these conditions yet. The rate of improvement has been impressive so far, however, and powerful machine learning systems will help keep that rate high.

2. Follow Police Instructions
Quote
...they certainly aren't equipped to follow the directions of a police officer...
But they could be equipped to do so, depending on what directions you want them to follow and how you feel about privacy and security. This strikes me as falling into the same category as police access to my phone. If the car is to be secured from #3 there can't be a master key that police can use to take over your vehicle. If this is just a question of obeying a traffic cop, I'd lump this under #1.

3. Hacking for Control and Hacking for Data
Quote
"It is feasible that people could commandeer self-driving vehicles ... to do their bidding, which could be malicious or simply just for the thrill of it," she said...
Quote
...privacy of personal data is another concern.
Identical to many of today's cars, as demonstrated last year. Information and physical security really should be built in to these vehicles from the ground up - conceptually similar to how modern iPhones are designed.

I hadn't considered the parallels between a self-driving car and cell phones before; it seems to me that many of the security and privacy concerns are the same. Perhaps most or all the concerns are the same. Interesting.
Title: Re: Self-Driving Cars
Post by: SkeptiQueer on March 17, 2016, 10:43:29 AM
Manual cars can be disabled with $4 worth of nails, an icepick, and the drivers can also be disabled with a laser pointer.
Title: Re: Self-Driving Cars
Post by: phooey on March 17, 2016, 08:40:23 PM
Just like I said, right on cue someone will say, well so what if someone could hack the car and take control of it. Plus people already drive bad in the rain.   

This is a skeptic website afterall.
Title: Re: Self-Driving Cars
Post by: SkeptiQueer on March 17, 2016, 08:48:03 PM
Not "so what" but that this is a problem with driver-operated cars, and needs to be addressed no matter who's driving.
Title: Re: Self-Driving Cars
Post by: phooey on March 17, 2016, 09:21:05 PM
Yes, I get it, you are arguing that people being able to hack into a self-driving cars system, and control the car remotely rather easily, is not reason enough to stop them from being put into use-since people can flatten car tires right now.

I got it.  That is for others to decide how the skeptic argument succeeds or doesn't.   
Title: Re: Self-Driving Cars
Post by: arthwollipot on March 17, 2016, 09:30:00 PM
Yes, I get it, you are arguing that people being able to hack into a self-driving cars system, and control the car remotely rather easily, is not reason enough to stop them from being put into use-since people can flatten car tires right now.

Wow. You must be getting dizzy with all that spin.

No, he's arguing that this is a known and acknowledged problem that is being addressed in the current phase of development.

At worst, he is arguing that people being able to hack into a self-driving cars system, and control the car remotely rather easily, is not reason enough to stop them from being put into use-since people can hack into a regular cars system, and control the car remotely rather easily right now. [sic]
Title: Re: Self-Driving Cars
Post by: SkeptiQueer on March 18, 2016, 05:26:37 AM
Mostly that the problems are known and appear to be solvable, and most are not unique to self-driving cars. Several regular cars can be hacked through the cellphone or entertainment system and will allow a third party to remotely control the vehicle.

Part of any technology is hiring people to try to break it, hack it, or find issues with it. Then people address those issues. Just pointing at a problem and screaming "IT WILL NEVER WORK SEEEEEE?????" is the technological equivalent of "then why are there still monkeys?!"
Title: Re: Self-Driving Cars
Post by: phooey on March 18, 2016, 05:46:06 AM
In fact asking why there are still monkeys is a good question, regardless of whether or not its a question you like. 

The assumption is that humans evolved from an animal very similar to a monkey.  What was it about that particular type of monkey that was not well enough adapted, that humans had a better chance at survival then them?  The answer is we haven't a clue.  Surely, if humans were as inept physically as they are now, and there were not enough of them originally to form large groups which could co-operate effectively (and communicate??), a monkey would probably be a much more effective survivalist.  Even if one is holding a club, a fight between a human and a chimpanzee is no contest, the chimpanzee would crush the human.  It takes a long time to invent a gun. 

As far as your straw man argument about self driving cars NEVER being able to overcome their technological issues-like I said that is a straw man.  What I said was that it is going to be a very very long time, and that it might be so fucked up as to cause companies like Google to finally admit, this isn't worth it, we will go bankrupt fighting lawsuits. 
Title: Re: Self-Driving Cars
Post by: SkeptiQueer on March 18, 2016, 07:52:12 AM
What is your basis for claiming it will take a very long time? Are you a code expert or an expert-system designer? Do you have experience in building and designing automated systems?

You're incredibly wrong about the monkey question. There's a number of things that allowed early hominids an advantage. Would you like a few book recommendations, or would you prefer to just assume you know better than the experts on the topic of archaeology and anthropology as well as on automation, AI, and computer programming?
Title: Re: Self-Driving Cars
Post by: stonesean on March 18, 2016, 08:33:40 AM
 :munch:
Title: Re: Self-Driving Cars
Post by: Fast Eddie B on March 18, 2016, 08:57:41 AM
The assumption is that humans evolved from an animal very similar to a monkey. 

Not necessarily.

It could have been closer to a human, and then "devolved" into a smaller, less intelligent arboreal animal.

Just as whales and snakes "lost" legs to become better adapted to their surroundings, and cave dwelling fish "lost" eyes, our common ancestor with monkeys may have "lost" some attributes along the way to be better suited to their environment.

But this is a huge derail, and has zero to do with self driving cars and deserves a thread of its own.
Title: Re: Self-Driving Cars
Post by: Morvis13 on March 18, 2016, 09:02:52 AM
The assumption is that humans evolved from an animal very similar to a monkey. 

Not necessarily.

It could have been closer to a human, and then "devolved" into a smaller, less intelligent arboreal animal.

Just as whales and snakes "lost" legs to become better adapted to their surroundings, and cave dwelling fish "lost" eyes, our common ancestor with monkeys may have "lost" some attributes along the way to be better suited to their environment.

But this is a huge derail, and has zero to do with self driving cars and deserves a thread of its own.

^This^

we lost our tail but again not on topic.

The monkeys are building the cars. Just as sticks became better weapons when we sharpen them, The car will change to suit our needs. This is not evolution. It is innovation. It is a matter of time. All we are arguing over here is if it is 5 years or 50. Either way, it doesn't matter. Monkeys should not be driving.
Title: Re: Self-Driving Cars
Post by: phooey on March 18, 2016, 09:58:04 AM
Perhaps.

Or maybe by the time we are close to getting the technology, we will have already evolved into a purple slime, with no thumbs, and won't have a need for transportation-since we won't have the desire to go anywhere. 
Title: Re: Self-Driving Cars
Post by: arthwollipot on March 18, 2016, 09:43:24 PM
Perhaps.

Or maybe by the time we are close to getting the technology, we will have already evolved into a purple slime, with no thumbs, and won't have a need for transportation-since we won't have the desire to go anywhere.

You still haven't said why you think that these problems - which many very smart people are working on solving right now - will take so long to solve.
Title: Re: Self-Driving Cars
Post by: phooey on March 18, 2016, 09:46:46 PM
Well, I don't know.  Maybe for one, because we still haven't figured out how to make safe drinking water yet. 

I guess there are other reasons too. 
Title: Re: Self-Driving Cars
Post by: arthwollipot on March 18, 2016, 10:17:10 PM
Well, I don't know.  Maybe for one, because we still haven't figured out how to make safe drinking water yet. 

I guess there are other reasons too.

We have figured out how to watch movies on a tiny device that fits in my pocket. We have figured out how to prevent millions of children from dying from polio. We have figured how how to send robots to distant planets.

And by the way, we figured out how to make safe drinking water millennia ago. Your argument is invalid.
Title: Re: Self-Driving Cars
Post by: brilligtove on March 18, 2016, 10:23:05 PM
Well, I don't know.  Maybe for one, because we still haven't figured out how to make safe drinking water yet. 

I guess there are other reasons too.

Your continuing failure to provide anything remotely resembling a reason we should trust your opinions leaves me wondering where you sit in the Venn diagram of intelligence, ignorance, and malice. At this point I can see you have some intelligence, and a vast panorama of ignorance. I'm not sure if you're intentionally malicious, though.

:munch:

Agreed.
Title: Re: Self-Driving Cars
Post by: phooey on March 18, 2016, 10:42:09 PM
Ignorance can be cured.  Stupidity, you deal with. 
Title: Re: Self-Driving Cars
Post by: brilligtove on March 18, 2016, 10:50:15 PM
Yes. I just haven't figured out whether you have a manageable condition or a curable one.
Title: Re: Self-Driving Cars
Post by: phooey on March 18, 2016, 10:55:35 PM
Its irrelevant to the problem of how self-driving cars will deal with people pretending they will walk in front of a car to force it to stop, or how it will know who to listen to when two people are drunk and fighting over whose house to go to. 

Taxi drivers can just tell their customers to get out, self-driving cars are on salary, so have no such rights. 
Title: Re: Self-Driving Cars
Post by: Johnny Slick on March 18, 2016, 11:31:12 PM
The reason you're being lampooned is that these are relatively trivial answers that can be fixed by setting heuristics *if* it can be proven that they're an issue in the first place, not some actual, devastating reason why automatic cars won't work. For the person who runs out in front of a car, the heuristic is going to be to swerve when possible and then, once it has assessed that no person was hit, to get out of the area as quickly as possible. As it stands now, this is better than how humans normally handle the situation, which is to stop, put their hands up, and get out of the car, or the way 1980s Tom Wolfe novels handled it, which was to run over the person who stepped out in front of the car and then keep on driving. If there is some other crazy permutation, guess what? It can be coded in. This side of things is not the hard part. There *are* hard parts of automated driving that have to be resolved before this becomes a reality but you're not even coming close to addressing them and if you really want to do so rather than puff up your chest about how this one time you took on a whole forum of skeptics and "won" because you never acknowledged how foolish your arguments were, I would recommend actually reading about the subject.
Title: Re: Self-Driving Cars
Post by: phooey on March 19, 2016, 12:04:16 AM
Actually Johnny its you who is being lampooned, which is made all the more humorous by you not realizing it.

For instance, you believe that you are going to be able to program the car into knowing that it was someone intentionally pretending to run in front of the car, rather than a kid on a skateboard who didn't see the car coming.  Its going to swerve for one, brake for the other, and drive off out of the area as quickly as possible for a grandma on a walker (and the car is going to know what "the area" is). 

Please continue.  I also happen to love when skeptics say "read a book on the subject", like they think they just came up with a great retort ( every skeptic says this when they are at a loss for words, just to let you know why I am laughing at you next time). 

I want to here more about the coding for street thugs/grandmas with canes detection ability.  The street thugs will have to wear grey wigs? 

Great stuff Johnny. 
Title: Re: Self-Driving Cars
Post by: SkeptiQueer on March 19, 2016, 01:58:51 AM
Once again, you're making some as-yet unsubstantiated claims about the difficulty of solving these problems. What makes you know better than the people who are designing the things and setting timetables for them?
Title: Re: Self-Driving Cars
Post by: phooey on March 19, 2016, 06:55:13 AM
Well if the programmers think anything like you skeptics, for instance believing that they are going to be able to distinguish those who have good intentions with those who have bad, and program the car to just swerve out of the way and then drive out of the area as fast as possible, then I feel pretty safe in saying that they haven't got it very well thought out. 
Title: Re: Self-Driving Cars
Post by: Johnny Slick on March 19, 2016, 08:19:14 AM
no u see taht mens i win
Title: Re: Self-Driving Cars
Post by: daniel1948 on March 19, 2016, 09:23:12 AM
Robotics Expert: Self-Driving Cars Not Ready for Deployment <...snip...>

It is perfectly true that self-driving cars are not yet ready for deployment. Note the word "YET." And that's why no company is deploying them yet. There was a time, for every gadget you own, when it was "not ready for deployment" yet.

So what?

Everybody reading this knows that self-driving cars are not ready for deployment. Yet. There are lots of issues that remain to be solved, and lots of really smart people working on solutions. I personally don't believe that we'll ever have true A.I. But I do think there will come a time not too far from now, when self-driving cars will be safer overall than human-driven cars.

And that's really the goal and the metric: When self-driving cars reach the point of being safer overall than human drivers. When you are statistically less likely to be hurt or killed by a computer-operated car than by a human-operated car. The types of accidents caused by computer-driven cars will be different than the types of accidents that people cause. But that's a distraction and not relevant except to the people working to solve the problems. What matters to the public policy decision of allowing computer-driven cars is the overall safety of people on the roads. Which is more likely to hurt or kill you: a human driver or a computer driver? We have not YET reached the tipping point. And that's why we don't YET have those cars on the road.

End of story. The time is coming. Get used to it.
Title: Re: Self-Driving Cars
Post by: phooey on March 19, 2016, 09:34:43 AM
No I don't think you are right Daniel. 

Its not a simple calculation of which would cause more overall deaths, self-driven cars or people driven.

Because in the case of the self-driven cars, the people who are harmed, ALL will be innocent victims who had the bad luck of being in one of these cars.  Now trying to tell those people, that well, sorry its unfortunate for you that your baby was killed because the program made an error, but look, you have to expect these things to happen sometimes.  Its the price we all pay. get used to it.

No, I don't think most people will accept that as just the cold calculations of computers.  As it is now, people feel they have some control over their fate (and they mostly do).  Telling them they will have zero control over that fate is going to be a pretty hard sell I reckon.   
Title: Re: Self-Driving Cars
Post by: daniel1948 on March 19, 2016, 10:03:27 AM
No I don't think you are right Daniel. 

Its not a simple calculation of which would cause more overall deaths, self-driven cars or people driven.

Because in the case of the self-driven cars, the people who are harmed, ALL will be innocent victims who had the bad luck of being in one of these cars.  Now trying to tell those people, that well, sorry its unfortunate for you that your baby was killed because the program made an error, but look, you have to expect these things to happen sometimes.  Its the price we all pay. get used to it.

No, I don't think most people will accept that as just the cold calculations of computers.  As it is now, people feel they have some control over their fate (and they mostly do).  Telling them they will have zero control over that fate is going to be a pretty hard sell I reckon.   

That's just silly! Do you really think a grieving parent will care whether it was a computer or a drunk or incompetent driver that killed their baby? Perhaps when the technology reaches the tipping point and you are only 3% safer in a computer-driven car, a lot of people, being ignorant of statistics, and having an inflated opinion of their own driving skills, will prefer to drive themselves. But when you are 95% less likely to be killed or injured by a computer-driven car than a human-driven car, only a moron will reject the obvious choice.

What you are ignoring, phooey, is that humans are outrageously incompetent drivers. They have poor reaction times, they get easily distracted, and there seems to be no way to dissuade them from such pure idiocy as eating, drinking, applying make-up, talking on their phone, or allowing pets loose in the car while driving. 32,675 people were killed in traffic accidents in the U.S. in 2014. That's many times more than the number of people killed by terrorists even in 2001. I didn't bother looking up injuries or accidents that only resulted in property damage. When the time comes (and it's when, not if) that computer-driven cars can reduce that thirty-two thousand to one thousand, nobody in their right mind is going to argue that it's "better" to be killed by a human than by a computer.
Title: Re: Self-Driving Cars
Post by: phooey on March 19, 2016, 11:33:26 AM
No its not a question of better Daniel. 

I am pretty sure the mother is going to feel that she would prefer to take her chances protecting her baby herself, rather than letting a computer program spin a roulette wheel. 
Title: Re: Self-Driving Cars
Post by: SkeptiQueer on March 19, 2016, 01:25:26 PM
Some people trust prayer to protect their children better than vaccines. Those people are still wrong, and that's still child abuse.

Title: Re: Self-Driving Cars
Post by: daniel1948 on March 19, 2016, 01:42:12 PM
No its not a question of better Daniel. 

I am pretty sure the mother is going to feel that she would prefer to take her chances protecting her baby herself, rather than letting a computer program spin a roulette wheel. 

If she'd rather subject her baby to a much greater likelihood of death, because she wrongly believes she can drive better than a computer, then perhaps the baby would be better off with a foster parent who is not an ignorant neo-Luddite.

You do have a point, though: A segment of the American public is so misinformed, and so irrational, that their policy preferences lead them to extremely self-destructive practices. One of the reasons we have government is to protect us from idiots. And once computer-driven cars become available, I want my government to protect me from idiots like your hypothetical mom who will kill me by driving while texting on her phone and distracted by her baby's crying, because she's too stupid to recognize that she, her baby, and I are all safer if she lets a computer operate her car.

(And yes, that day is not here YET, but will be soon.)

And P.S.: Not all mothers are idiots. Many will prefer the safer option!
Title: Re: Self-Driving Cars
Post by: Alex Simmons on March 19, 2016, 05:51:32 PM
Satire, but makes a point:
http://eveningharold.com/2016/03/15/driverless-car-chases-cyclist-for-three-miles-in-road-rage-incident/
Title: Re: Self-Driving Cars
Post by: phooey on March 19, 2016, 10:16:10 PM
Skeptiqueer,

Yep, I know you check all the skeptic boxes.  Big AG good, big pharma good, alternative medicines bad, evolution must be true, conspiracies don't happen, there is life on other planets, climate change science has been settled, yadda, yadda...

No need to always repeat it skeptiqueer, just say you are a Skeptic!  Then we will already know exactly everything you believe.  Sign on the bottom line. 
Title: Re: Self-Driving Cars
Post by: Johnny Slick on March 19, 2016, 10:24:07 PM
Guys I have nothing more to say and haven't for 10 pages so instead I will say you are not a True Skeptic because you won't accept my idiotic ideas about cars! What if a flock of penguins gain sentience and ask a car to make 3 left turns while simultaneously a dog human hybrid asks it to make a right?? Until we solve this problem which we never will, automated driving is a JOKE. And you would know this if you were a True Skeptic.
Title: Re: Self-Driving Cars
Post by: SkeptiQueer on March 19, 2016, 11:00:58 PM
Skeptiqueer,

Yep, I know you check all the skeptic boxes.  Big AG good, big pharma good, alternative medicines bad, evolution must be true, conspiracies don't happen, there is life on other planets, climate change science has been settled, yadda, yadda...

No need to always repeat it skeptiqueer, just say you are a Skeptic!  Then we will already know exactly everything you believe.  Sign on the bottom line.
So once more, what expert knowledge do you have that makes you more qualified to make judgements about production of expert systems or AI than the people who get paid lots of money to build expert systems and AI?
Title: Re: Self-Driving Cars
Post by: teethering on March 20, 2016, 01:07:59 AM
Actually Johnny its you who is being lampooned, which is made all the more humorous by you not realizing it.

For instance, you believe that you are going to be able to program the car into knowing that it was someone intentionally pretending to run in front of the car, rather than a kid on a skateboard who didn't see the car coming.  Its going to swerve for one, brake for the other, and drive off out of the area as quickly as possible for a grandma on a walker (and the car is going to know what "the area" is). 

Please continue.  I also happen to love when skeptics say "read a book on the subject", like they think they just came up with a great retort ( every skeptic says this when they are at a loss for words, just to let you know why I am laughing at you next time). 

I want to here more about the coding for street thugs/grandmas with canes detection ability.  The street thugs will have to wear grey wigs? 

Great stuff Johnny.

Wait, what?

Humans aren't perfect at instantly detecting those differences either.  And humans are slower at reacting to anything that's happening, whether it's someone pretending to run in front of a car or a kid on a skateboard.

Yes, pattern recognition is still hard for computers, but that's why self-driving cars are actually very promising, because the patterns to recognize and care about are in a very limited space, as opposed to the set of problems we need human brains for.

If you were trying to make a point that humans are inherently better here, you have failed.  If you that's not the point you were making, I don't know what it is.  We already have fully autonomous transportation systems deployed for decades.  Like with self-driving cars the problem space is reduced so that computers can deal with the problems.  If we're talking about trains then being on rails, for example, makes things a lot simpler.  If we're talking about planes then requiring human pilots to takeoff and land leaves computers with what they can safely deal with.  There are lots of contingencies on the road that need to be accounted for a car, but at the same time we're dealing with much lower speeds than other modes of transport (which means there are a lot more computing cycles available in which to solve more complex problems, because distances close slower and you have more time) and there are more options to avoid crashes, like stop, swerve, accelerate etc.

And finally the only consideration is whether computerized systems would be safer than humans at driving and there's just no doubt that the answer will be yes, because as was mentioned before computers don't drink, they don't text and they don't road-rage.  It doesn't mean it's a simple engineering problem to solve, but there's no good reason to be skeptical that it is very close to being solved.  The idea isn't that self-driving cars will be perfect at driving, but they'll definitely reduce problems associated with driving.

I find this thread very strange.
Title: Re: Self-Driving Cars
Post by: arthwollipot on March 20, 2016, 01:30:49 AM
Guys I have nothing more to say and haven't for 10 pages so instead I will say you are not a True Skeptic because you won't accept my idiotic ideas about cars! What if a flock of penguins gain sentience and ask a car to make 3 left turns while simultaneously a dog human hybrid asks it to make a right?? Until we solve this problem which we never will, automated driving is a JOKE. And you would know this if you were a True Skeptic.

I think it's pretty clear that phooey has no idea what a True Skeptic even is.
Title: Re: Self-Driving Cars
Post by: Alex Simmons on March 20, 2016, 02:28:22 AM
I expect we'll get to self drive via an evolutionary process. We already have an amazing range of tech in cars now.

At the present I am considering a new(ish) vehicle. I can already buy a used car at a relatively modest price level that:
- has active cruise control, which automatically adjusts the car's speed (within the set desired speed) to maintain a desired distance from vehicles ahead of you, or that are moving into the lane ahead of you.
- continually monitors the line you are taking on the road and provides warnings about maintaining a good line, and if appropriate/necessary makes steering adjustments for you
- self parks into parallel space or rear facing spaces, and can also drive out of such spaces
- has collision avoidance / severity reduction systems to engage an range of tech, from simple warnings through to actively undertaking emergency braking/steering for you
- if a collision has occurred that may have incapacitated the driver, will engage systems such as braking and hazard warnings to help prevent or at least significantly reduce potential for subsequent collisions
- engages an emergency routine to safely stop the vehicle when it detects that you are no longer providing any steering input for a given amount of time (e.g. you've fallen asleep at the wheel)
- that automatically operates the throttle and brake for you when driving in heavy traffic / traffic jam type conditions
- has increasingly sophisticated mapping, traffic monitoring and guidance systems
- uses voice control
- hill descent control
- blind spot monitoring / lane change control

In more expensive cars there are a range of other driver aids.

This evolution will continue such that self drive cars will sort of creep up on us.
Title: Re: Self-Driving Cars
Post by: SkeptiQueer on March 20, 2016, 03:49:09 AM
I just want a highway autopilot so I can fuck off until we take the designated exit, and enjoy the massively improved flow of traffic.
Title: Re: Self-Driving Cars
Post by: Fast Eddie B on March 20, 2016, 09:18:28 AM
What you are ignoring, phooey, is that humans are outrageously incompetent drivers. They have poor reaction times, they get easily distracted, and there seems to be no way to dissuade them from such pure idiocy as eating, drinking, applying make-up, talking on their phone, or allowing pets loose in the car while driving. 32,675 people were killed in traffic accidents in the U.S. in 2014.

I think you're guilty of a hasty generalization here.

For the most part, human drivers do manage to get from point A to point B with quite amazing regularity. Accidents, while fairly common, have to be viewed against the background of total miles traveled accident free.

Not to say 32,000+ is not a large number, but to give it meaning one has to view it against the huge numbers of drivers driving safely, day and night, sometimes over decades or even a lifetime without a single accident.

Also not saying self-driving cars don't have the potential to do better - just credit where credit is due that most drivers do a pretty good job most of the time.
Title: Re: Self-Driving Cars
Post by: phooey on March 20, 2016, 10:24:20 AM
Indeed I concur. 

I happen to drive in a country that is far more difficult to maneuver through than American roads, and after many a year, I have never had a real accident (a bicyclist did run into me once though). 

I listened to a guy who owns an expensive car that has certain self -driving help features in it.  It will apply the brakes for you if it feels there is a dangerous situation.  He absolutely hates it. Scares the hell out of him.     
Title: Re: Self-Driving Cars
Post by: SkeptiQueer on March 20, 2016, 10:41:07 AM
Indeed I concur. 

I happen to drive in a country that is far more difficult to maneuver through than American roads, and after many a year, I have never had a real accident (a bicyclist did run into me once though). 

I listened to a guy who owns an expensive car that has certain self -driving help features in it.  It will apply the brakes for you if it feels there is a dangerous situation.  He absolutely hates it. Scares the hell out of him.   

Then maybe they aren't right for your country, but declaring that Tesla can't have a car capable of traversing roads with a sleeping driver in the near future is misguided at best.

You forgot to mention what your credentials were regarding programming, btw.
Title: Re: Self-Driving Cars
Post by: phooey on March 20, 2016, 10:50:07 AM
Is the car going to use voice commands or manual input into a screen?  Can it be done by one, or many people? 

Having a car that can follow a highway line is not the same challenge as having a car the can navigate city streets, where other forms of transportation and pedestrians are allowed to roam.  Highways are basically a restricted airspace, like a plane navigates, a much easier proposition.

You forgot to tell me how a self-driving car is going to get away from a beggar holding a car ransom at a traffic light until it gets paid.  Oh wait, that's right, you said a loud horn.  Pepper spray.  A loaded gun. 

Do you think Google will team up with Smith and Wesson? 

 
Title: Re: Self-Driving Cars
Post by: Captain Video on March 20, 2016, 11:13:47 AM
Can we be done with this kid already?
Title: Re: Self-Driving Cars
Post by: Johnny Slick on March 20, 2016, 11:16:54 AM
You forgot to tell me how a self-driving car is going to get away from a beggar holding a car ransom at a traffic light until it gets paid.  Oh wait, that's right, you said a loud horn.  Pepper spray.  A loaded gun.
Here's a crazy thought: driving around the beggar. OH SNAP I JUST SOLVED SCIENCE
Title: Re: Self-Driving Cars
Post by: Morvis13 on March 20, 2016, 11:19:05 AM
Is the car going to use voice commands or manual input into a screen?  Can it be done by one, or many people? 

Having a car that can follow a highway line is not the same challenge as having a car the can navigate city streets, where other forms of transportation and pedestrians are allowed to roam.  Highways are basically a restricted airspace, like a plane navigates, a much easier proposition.

You forgot to tell me how a self-driving car is going to get away from a beggar holding a car ransom at a traffic light until it gets paid.  Oh wait, that's right, you said a loud horn.  Pepper spray.  A loaded gun. 

Do you think Google will team up with Smith and Wesson?

Ridiculous. I would bet that each car company would have their own way to switch manual and robot driver. Voice, button or whatever. That really doesn't matter. The point is if a beggar stands in front of a manual driver there isn't much you can do. If the car was on robot driver then the manual over ride can take care of these pests. Eventually the robot can be programmed to avoid these. I don't think this is a huge problem.


Title: Re: Self-Driving Cars
Post by: phooey on March 20, 2016, 11:24:56 AM
It seems a lot of people just jump into the conversation without reading many of the posts which have discussed the very issue.

To wit, many of the car companies have explicitly said that they want to remove the steering wheel from the cars completely, so that there can be no conflicts between the computers decision and a humans.  So I guess its just another of the many stumbling blocks the car companies are going to have to deal with.

Title: Re: Self-Driving Cars
Post by: daniel1948 on March 20, 2016, 12:49:19 PM
What you are ignoring, phooey, is that humans are outrageously incompetent drivers. They have poor reaction times, they get easily distracted, and there seems to be no way to dissuade them from such pure idiocy as eating, drinking, applying make-up, talking on their phone, or allowing pets loose in the car while driving. 32,675 people were killed in traffic accidents in the U.S. in 2014.

I think you're guilty of a hasty generalization here.

For the most part, human drivers do manage to get from point A to point B with quite amazing regularity. Accidents, while fairly common, have to be viewed against the background of total miles traveled accident free.

Not to say 32,000+ is not a large number, but to give it meaning one has to view it against the huge numbers of drivers driving safely, day and night, sometimes over decades or even a lifetime without a single accident.

Also not saying self-driving cars don't have the potential to do better - just credit where credit is due that most drivers do a pretty good job most of the time.

It is amazing that so few people die every year in car crashes. Note that not so long ago the number was around 40,000. Advances in the safety of cars have greatly reduced the number of deaths. It's safe to say that the number of accidents that do not involve deaths is many times greater. And I'm sure you will agree, even though phooey seems not to, that when computer-driven cars can lower that thirty-two thousand deaths to one thousand deaths, the obvious rational choice will be to switch to computer-driven cars.

Is the car going to use voice commands or manual input into a screen?

Different auto makers will use different methods.

Can it be done by one, or many people?

You'll probably have the choice whether to allow others to program your car or only allow yourself to do it.

Having a car that can follow a highway line is not the same challenge as having a car the can navigate city streets, where other forms of transportation and pedestrians are allowed to roam.  Highways are basically a restricted airspace, like a plane navigates, a much easier proposition.

That's why we're not quite there. Yet.

You forgot to tell me how a self-driving car is going to get away from a beggar holding a car ransom at a traffic light until it gets paid.

How does a human driver get away from a beggar "holding a car ransom"? This has never happened to me, but if it did, I really don't know what I could do except back up if it's possible, or else just wait there until he gets tired and leaves. (Though maybe the reason I've never been held "ransom" in my car is that I'm not a stingy a-hole and I always give money to beggars if it's safe to stop. I don't stop if the traffic is moving and there is a car behind me.)

Really, phooey, the only legitimate objection you've raised that makes any sense at all is that people are afraid of computer-driven cars. Many people are. Car makers will have to educate the public concerning the true risks and benefits.

I bought a Toyota Prius in 2004. The Prius was still pretty new technology, having been on American roads for only 3 years (IIRC) in an earlier model. A lot of people were scared of it. They didn't trust the relatively new technology. A few people had bought the earlier model so it was actually well-tested, but it hadn't really entered the public awareness. So I got all sorts of silly questions about the car based on ignorance, and for the first year or two it was mostly tree-huggers like me who bought it. But eventually the public came to understand the technology and the car is now totally mainstream and extremely popular.

Computer-driven cars will go through a similar process of public ignorance and fear, to gradual acceptance, and eventually they'll become mainstream.
Title: Re: Self-Driving Cars
Post by: SkeptiQueer on March 20, 2016, 04:32:21 PM
It seems a lot of people just jump into the conversation without reading many of the posts which have discussed the very issue.

To wit, many of the car companies have explicitly said that they want to remove the steering wheel from the cars completely, so that there can be no conflicts between the computers decision and a humans.  So I guess its just another of the many stumbling blocks the car companies are going to have to deal with.
Why do you think these stumbling blocks will take a significant amount of time?

Why are solutions used by human drivers in countries where traffic disruptions are common suddenly  I good when the driver is a computer and the passenger is the one activating the pepper spray or loud alarm?

So far every solution has been ignored or waved off, but without a skeptical examination.
Title: Re: Self-Driving Cars
Post by: PANTS! on March 20, 2016, 05:33:03 PM
What do they do in Russia when some jackhole gets infront of their car?  How is the solution Russians use inherently safer than an automated one?  Seems to me that backing up or going around would be safer when all the surrounding cars know what maneuver the car will be performing.
Title: Re: Self-Driving Cars
Post by: phooey on March 20, 2016, 10:09:37 PM
Lots of people's minds get stuck into one way of thinking, and they simply can not open their minds.  I see that with plenty of skeptics, and you can see it plainly here.

I mean look what you have people saying.  "Well, the car will just have a special program to swerve away from people who try to fuck with it, and drive as fast as it can out of that area" (How does it know who is real and who is fucking with it, well, we'll figure that out later. Will it break the speed limit?).

"It will just run over beggars."(It will run over Syrian refugees asking for a quarter?)"

"Well, it will have pepper spray shooters built in, plus you can always carry a gun."  This is what skeptics think is a solution?  Soccer moms armed with glocks is their answer? 

"Some will have manual input and some will use voice commands."  Yea, that's gonna really work well with voice commands.  Anyone's voice?  When someone's says stop the car does it do so immediately?  Does it look for a safe place to stop?  Then someone else says, no, don't stop, keep driving.  Does it listen to the first or the second command?  The first voice repeats, stop the car.  A third voice says, turnaround.  Who is it taking orders from now?  Its in the middle of Manhattan, cars are piling up behind it as it tries to figure out a solution.  People behind are getting pissed and beeping their horns.  Or its on a country road.  It stopped because you told it to stop.  But there is no safe shoulder to stop on, so it doesn't know if it should really stop here.  Should it drive onto the grass?  Oh right just go then.  No don't just go, I want to stop and take a leak.  I said go.  I said stop! 

Others have suggested we can have a fully self driving car that you can sleep in, but also allows manual takeover at anytime. The manufacturers already know this idea is preposterous, because what do you do about drunks, or about people who just woke up and see the car doing something that is weird, does it allow them to take over immediately, when actually the car was making a safe maneuver that you just fucked up?  Plus eventually people's own driving skills will get worse, since they don't use them, do you still let them takeover?  This is why they want to remove the steering wheel completely.  So now you have a mishmash of half thought-out ideas that won't work together. 

The skeptics all believe the great minds of the world's elite programmers.  Like those programmers who thought Y2K might shut down the entire world, because they got the decimal points wrong on the dates. The same programmers who have a system now that is built on top of so many layers of bad programming that it can't keep kids in their basement from stealing people's credit cards, or from Russians taking over computers from 10,000 miles away and demanding ransom so hospitals can get their patients files back.  Rest assured, Elon Musk has already thought of all of this, don't worry! 

I got news for you.  All Elon Musk has done is make an electric car.  Its a nice electric car, sure.  But we have been making electric cars since 1834.  Now you can finally buy one.  Great.

That's probably about how long it is going to take to perfect self-driving cars. 

People calling themselves "skeptics" is so humorously poignant.   

Title: Re: Self-Driving Cars
Post by: daniel1948 on March 20, 2016, 10:44:31 PM
You seem to be awfully angry over a technology you don't believe will be available for a hundred years.
Title: Re: Self-Driving Cars
Post by: phooey on March 20, 2016, 11:13:04 PM
Well, if it is like the electric car, it was more like 200 years, not 100.   
Title: Re: Self-Driving Cars
Post by: arthwollipot on March 21, 2016, 04:15:33 AM
You seem remarkably convinced that you are coming up with new obstacles, that the manufacturers haven't already thought of. Why do you think you're so original, when the manufacturers are employing the best people that they can find to think of and deal with these things?

I guarantee that the people working for the manufacturers have already thought of and solved more problems that you can think of. I absolutely guarantee it.
Title: Re: Self-Driving Cars
Post by: bimble on March 21, 2016, 05:36:17 AM
Well, if it is like the electric car, it was more like 200 years, not 100.

you know that there were electric cars back in the 1880's, right?? And if it hadn't been for the ease and cost of the internal combustion engine we might all be driving electric cars now.

Lots of people's minds get stuck into one way of thinking, and they simply can not open their minds.  I see that with plenty of skeptics, and you can see it plainly here.


Yes, yes you can. Very plainly, one person absolutely stuck into one way of thinking...  ;)
Title: Re: Self-Driving Cars
Post by: SkeptiQueer on March 21, 2016, 06:03:46 AM


Lots of people's minds get stuck into one way of thinking, and they simply can not open their minds.  I see that with plenty of skeptics, and you can see it plainly here.

I mean look what you have people saying.  "Well, the car will just have a special program to swerve away from people who try to fuck with it, and drive as fast as it can out of that area" (How does it know who is real and who is fucking with it, well, we'll figure that out later. Will it break the speed limit?).

"It will just run over beggars."(It will run over Syrian refugees asking for a quarter?)"

"Well, it will have pepper spray shooters built in, plus you can always carry a gun."  This is what skeptics think is a solution?  Soccer moms armed with glocks is their answer? 

"Some will have manual input and some will use voice commands."  Yea, that's gonna really work well with voice commands.  Anyone's voice?  When someone's says stop the car does it do so immediately?  Does it look for a safe place to stop?  Then someone else says, no, don't stop, keep driving.  Does it listen to the first or the second command?  The first voice repeats, stop the car.  A third voice says, turnaround.  Who is it taking orders from now?  Its in the middle of Manhattan, cars are piling up behind it as it tries to figure out a solution.  People behind are getting pissed and beeping their horns.  Or its on a country road.  It stopped because you told it to stop.  But there is no safe shoulder to stop on, so it doesn't know if it should really stop here.  Should it drive onto the grass?  Oh right just go then.  No don't just go, I want to stop and take a leak.  I said go.  I said stop! 

Others have suggested we can have a fully self driving car that you can sleep in, but also allows manual takeover at anytime. The manufacturers already know this idea is preposterous, because what do you do about drunks, or about people who just woke up and see the car doing something that is weird, does it allow them to take over immediately, when actually the car was making a safe maneuver that you just fucked up?  Plus eventually people's own driving skills will get worse, since they don't use them, do you still let them takeover?  This is why they want to remove the steering wheel completely.  So now you have a mishmash of half thought-out ideas that won't work together. 

You made up a bunch of scenarios and said they were proof that self-driving cars couldn't work. Several people responded to the problem in different ways, much in the way that and design team would. We can't test them all (Google won't lend us one, I asked) but since we didn't all immediately come up with the same perfect tested solution, that means the designers can't either? Or *gasp* maybe different companies will have different approaches, the way Boeing and Airbus approach autopilot programming differently.

I like how solutions to carjacking that are used successfully in other countries aren't solutions just because you said so. That's a really convenient way to avoid actually having a discussions.

Quote
  Like those programmers who thought Y2K might shut down the entire world, because they got the decimal points wrong on the dates.

No, they needed to add 2 digits to add bunch of databases. The problem was anticipated several years before, in fact being part of the plot of the movie Office Space. It wasn't a mad panic in the minds of anyone except people who didn't understand what was happening.

Quote
The same programmers who have a system now that is built on top of so many layers of bad programming that it can't keep kids in their basement from stealing people's credit cards, or from Russians taking over computers from 10,000 miles away and demanding ransom so hospitals can get their patients files back.

Hah, one of my friends works pretty high up in infosec. Stopping credit card theft usually involves things like watching for skimmers or following the protocols they set (which Target didn't do, btw) and generally viruses can be anyone voided with very simple security measures. The stuff they worry about is like people using the hard drive or BUS speaker to transmit data across airgaps. As others noted, many modern cars can be remotely shut down or controlled, and that's''s a problem infosec is working on already and it's not unique to self-driving cars.

What does this have to do with cars, aside from providing several examples of how little you understand the technology?

Quote
I got news for you.  All Elon Musk has done is make an electric car.  Its a nice electric car, sure.  But we have been making electric cars since 1834.  Now you can finally buy one.  Great.

No, that's not all. Improved battery technology, fast charging, infrastructure, interface, and it's an attractive car. Saying all Tesla did is make an electric car is like saying that all Tesla did was make a rocket ship, and we've had rockets since 1500BCE.

Quote
That's probably about how long it is going to take to perfect self-driving cars. 
Why? Modern electric commuter cars became viable because battery tech finally caught up. What specific technologies are lacking for self-driving cars?

Also, golf carts running on batteries have been around quite a bit longer, as have fully street-legal EEVs. If you're going to claim that electric cars were invented in 1834, then electric golf carts would be consumer electric automobiles, as they're as functional as a mule wagon circa 1834 was.

Title: Re: Self-Driving Cars
Post by: daniel1948 on March 21, 2016, 09:57:01 AM
Phooey: I ask again (or rather I turn my statement into a question): Why are you so angry about a technology you think will never exist? Why can't you just say "I think the problems won't be surmountable and we'll never get them."?

Lots of folks here think we'll colonize Mars. I disagree with them. But I don't get all angry and claim they are bad skeptics. I just say I think there are insurmountable obstacles and we'll never do it.

But you are practically steaming out the ears in anger against the idea of computer-driven cars, all the while saying they'll never work. I can't figure out where all the anger is coming from. Or are you afraid that they really will work and by the time you are old enough to get a driver's license there won't be any more human-driven cars and you'll never get a chance to drive a car? If that's the case, don't worry: There will always be tracks where people can drive cars.
Title: Re: Self-Driving Cars
Post by: PANTS! on March 21, 2016, 10:47:23 AM
https://www.youtube.com/watch?v=an2moViurLk

As I watch this, it came to mind that automated cars will probably handle this better anyway.  It could be set up to automatically file a police report with the dash cam attached.  Maybe even get permission from the police to go around.  As things stand now, meat drivers will typically have to wait until the police get there before they can proceed.
Title: Re: Self-Driving Cars
Post by: phooey on March 21, 2016, 11:55:48 AM
Daniel calm down.  Why are you getting all worked up about a technology that isn't ready for you yet?  take it easy, you are going to give yourself an aneurysm. 

Seriously, why are you raging?  Your mom won't let you drive the car, so you think if you get a babysitter car, she might let you out? 

I know its a bummer for you, but just accept it for now.  You also aren't going to get to fly on the Starship Enterprise probably, but you can still get a Playstation 2. 

I am a little worried about you, are you always so emotional?  Try that counting to ten trick they taught you.   
Title: Re: Self-Driving Cars
Post by: brilligtove on March 21, 2016, 01:19:33 PM
Well I learned two things about Phooey in all this.

First, it appears to have a simple heuristic for responding to us.
A. Ignore questions of experience, expertise, or competence.
B. Reflect questions of tone in the "I'm a bitchy bitch" mirror. Also known as the "I'm rubber, you're glue defence" or the "Pee wee Herman defence." (http://sguforums.com/index.php/topic,44070.msg9398016.html#msg9398016)
C. Wield profound ignorance as a club.

Second, Phooey has finally given some testimony that indicates its qualifications to speak authoritatively on the question at hand. I enjoyed its display of deep and profound ignorance of programming and engineering in the recent spate of posts.

This leads me to wonder if Phooey is an Eliza-bot, a troll, an ignoramous, or an idiot.

Well, it could be a combination of some of these options, of course. It could be quite a clever ignorant AI troll, for example.
Title: Re: Self-Driving Cars
Post by: PANTS! on March 21, 2016, 01:26:40 PM
B. Reflect questions of tone in the "I'm a bitchy bitch" mirror. Also known as the "I'm rubber, you're glue" defense.

Sorry, but this is aptly called the "Pee Wee Herman defense"
Title: Re: Self-Driving Cars
Post by: brilligtove on March 21, 2016, 01:32:34 PM
B. Reflect questions of tone in the "I'm a bitchy bitch" mirror. Also known as the "I'm rubber, you're glue" defense.

Sorry, but this is aptly called the "Pee Wee Herman defense"

Corrected. I presume others may be as ignorant of Pee Wee Herman as I am, so I clarified rather than replaced.

Also, I predict Phooeybot will respond with a type B comment, and then make a hash of it by salting it with type C invective. And that it will not understand that particular phrasing and it's relationship to crypto and secure storage of private information.
Title: Re: Self-Driving Cars
Post by: PANTS! on March 21, 2016, 02:00:05 PM
Also, I predict Phooeybot will respond with a type B comment, and then make a hash of it by salting it with type C invective. And that it will not understand that particular phrasing and it's relationship to crypto and secure storage of private information.
I know you are, but what am I?
Title: Re: Self-Driving Cars
Post by: amysrevenge on March 21, 2016, 02:16:24 PM
You know why we shouldn't have smartphones?  Try shouting "Siri, text Mom, 'Eat shit'. Send"* in a crowded room.  No way to stop it, everyone's moms die of broken hearts.  Unsolvable problem.

*Or whatever the exact command would be. I don't use voice recognition (or Apple).
Title: Re: Self-Driving Cars
Post by: Friendly Angel on March 21, 2016, 03:42:26 PM
I presume others may be as ignorant of Pee Wee Herman as I am,



mecalecahi mecahinyho

Title: Re: Self-Driving Cars
Post by: PANTS! on March 21, 2016, 06:04:44 PM
The secret phrase is "true skeptic".
Title: Re: Self-Driving Cars
Post by: Friendly Angel on March 21, 2016, 07:49:08 PM
What if I told you...

(http://33.media.tumblr.com/2e4e68237a69d2bcc19c3a0932252d98/tumblr_inline_naixve8eIO1stmifk.jpg)

Cowboy Curtis went on to achieve a certain notoriety.
Title: Re: Self-Driving Cars
Post by: PANTS! on March 21, 2016, 08:14:39 PM
He took the red pill.
Title: Re: Self-Driving Cars
Post by: Johnny Slick on March 21, 2016, 09:05:36 PM
Also, Chairy went on to star as the psychiatrist's couch in The Sopranos.
Title: Re: Self-Driving Cars
Post by: Captain Video on March 21, 2016, 10:41:26 PM
I presume others may be as ignorant of Pee Wee Herman as I am,



mecalecahi mecahinyho

I had to look that up

Really?
(click to show/hide)

What I dislike the most about this thread is its location, we are in the Tech Talk section and while there has been some argument over which OS is better (cough... windows.. cough cough) as well as other topics in general this section is not suposed to be a big argument like politics, skepticism, or religion, This is where we get to talk about upcoming new tech or even old tech,  to help each other with tech support, geek out, and generally get excited about geeky things. Its my favorite section on the forum and this thread has left a big brown stain all over it. So lets stop/

Phooey I'm sure most of us can accept that you are against self-driving cars and obviously take the position that they will not be safe, I get it, they get it, we all get it, I wont argue it any more. Can you accept that regardless of the future of  self driving safety in either direction at the very least we will learn something by trying?

Can we get back to geeking about new tech and how cool it is regardless of its practicality and safety?

Title: Re: Self-Driving Cars
Post by: Hanes on March 21, 2016, 10:51:53 PM

Can we get back to geeking about new tech and how cool it is regardless of its practicality and safety?



I can't wait to have two hours of my day back (1/3rd of my non-sleeping, non-working time) when I can do whatever I want in the car instead of focusing on the road and all the assholes I'm surrounded by.
Title: Re: Self-Driving Cars
Post by: phooey on March 21, 2016, 11:19:49 PM
Simon,

I think your comment towards me is a bit misplaced.  I am discussing the conceptual flaws with the technology.  Most of the people responding here are not.  Daniel is not alone in being downright angry that I brought up the flaws the technology is going to have problems overcoming (His rage is apparently infectious). 

The point of discussing things online is to see the other side of an argument, to have further insight into your own position. For instance I have learned that the ideas that people have for these problems right now are absolutely nuts.  Now possibly someone else somewhere has some real solutions to these issues, but I remain "skeptical" until someone can show me otherwise. 

I guess I am the only real skeptic here. 
Title: Re: Self-Driving Cars
Post by: Friendly Angel on March 21, 2016, 11:24:10 PM
I presume others may be as ignorant of Pee Wee Herman as I am,



mecalecahi mecahinyho

I had to look that up

Really?


Google FU has failed  YU.

http://youtu.be/XN5kVvR8VQI
Title: Re: Self-Driving Cars
Post by: Captain Video on March 22, 2016, 12:52:20 AM
I presume others may be as ignorant of Pee Wee Herman as I am,



mecalecahi mecahinyho

I had to look that up

Really?

Google FU has failed  YU.

I know where it came from but that silly answer came up in a few places when I googled what it meant, I assume originating from a satire site, it was a joke.
Title: Re: Self-Driving Cars
Post by: kvuo75 on March 22, 2016, 08:55:02 AM

You forgot to tell me how a self-driving car is going to get away from a beggar holding a car ransom at a traffic light until it gets paid

you make intentionally fucking with traffic a serious offense...
Title: Re: Self-Driving Cars
Post by: daniel1948 on March 22, 2016, 10:34:04 AM
I guess I am the only real skeptic here. 

 :roflolmao:
Title: Re: Self-Driving Cars
Post by: phooey on March 22, 2016, 10:40:04 AM

You forgot to tell me how a self-driving car is going to get away from a beggar holding a car ransom at a traffic light until it gets paid

Equivalent to like what, drunk driving?  Selling crack? 

That works, huh? 

you make intentionally fucking with traffic a serious offense...
Title: Re: Self-Driving Cars
Post by: SkeptiQueer on March 22, 2016, 10:41:52 AM
If someone's obstructing your path but not putting hands on you, that is unlawful detention already. Problem solved.
Title: Re: Self-Driving Cars
Post by: Fast Eddie B on March 22, 2016, 10:43:14 AM

I guess I am the only real skeptic here.

I doubt that!  ;)
Title: Re: Self-Driving Cars
Post by: daniel1948 on March 22, 2016, 02:20:58 PM
Phooey confuses "skeptic" with "doubter." He does not understand the definition.
Title: Re: Self-Driving Cars
Post by: Jeremy's Sea on March 22, 2016, 05:10:41 PM
Phooey confuses "skeptic" with "doubter." He does not understand the definition.
I'll be curious to read his/her threads on global warming too.
Title: Re: Self-Driving Cars
Post by: teethering on March 22, 2016, 05:22:02 PM
Simon,

I think your comment towards me is a bit misplaced.  I am discussing the conceptual flaws with the technology.

Conceptual flaws with the technology?  As in, the concept is so flawed it is not a worthwhile pursuit?

Because that's demonstrably false.  As I've said earlier we have a ton of automated transportation already fully functional in different settings, from fully autonomous trains, to computers flying planes and so on and so on.  Cars are not a quantum leap from that spectrum.  Some things are harder with cars and some things are simpler with cars and there's no reason to believe any of the problems brought up in this thread are so radically difficult that we can't solve them in the next few years.

Quote
  Most of the people responding here are not.  Daniel is not alone in being downright angry that I brought up the flaws the technology is going to have problems overcoming (His rage is apparently infectious). 

The point of discussing things online is to see the other side of an argument, to have further insight into your own position. For instance I have learned that the ideas that people have for these problems right now are absolutely nuts.  Now possibly someone else somewhere has some real solutions to these issues, but I remain "skeptical" until someone can show me otherwise. 

I guess I am the only real skeptic here.

Yep, that's it.  You're also the smartest and most knowledgeable person here, so there's that.   :worship:
Title: Re: Self-Driving Cars
Post by: Jeremy's Sea on March 22, 2016, 05:23:10 PM
Yep, that's it.  You're also the smartest and most knowledgeable person here, so there's that.   :worship:
y r u so irationaly angry? u must be wrong.
Title: Re: Self-Driving Cars
Post by: brilligtove on March 22, 2016, 07:44:01 PM
Thinking about the (largely) self-teaching AI behind the latest AI triump - winning at Go. If driving can be thought of as a game with rules this kind of AI could do a lot of good at several levels of complexity. There's the driving competence of each individual car, the collective competence of a herd, and the efficiency of the overall system.
Title: Re: Self-Driving Cars
Post by: phooey on March 22, 2016, 11:41:04 PM
noun
noun: skeptic

    1.
    a person inclined to question or doubt accepted opinions.


I don't know what the word means??

I reiterate, it appears I am the only skeptic here. 

Its amazing that so many of you have used the word to mean the exact opposite-to be an skeptic in your world means to go along with accepted opinions.   
Title: Re: Self-Driving Cars
Post by: brilligtove on March 22, 2016, 11:51:03 PM
Also, I predict Phooeybot will respond with a type B comment, and then make a hash of it by salting it with type C invective. And that it will not understand that particular phrasing and it's relationship to crypto and secure storage of private information.

While I did not predict the dick shun airy aspect to its response, I still think I nailed the prediction.

Phoobot, you may need to follow up on "inclined" and "question" and "inclined to question" before claiming understanding. Hint: these words do not refer to tilting or to rejection of evidence.
Title: Re: Self-Driving Cars
Post by: SkeptiQueer on March 23, 2016, 12:17:18 AM
noun
noun: skeptic

    1.
    a person inclined to question or doubt accepted opinions.


I don't know what the word means??

I reiterate, it appears I am the only skeptic here. 

Its amazing that so many of you have used the word to mean the exact opposite-to be an skeptic in your world means to go along with accepted opinions.   
I've been skeptical of your claims of the difficulty of solving problems, and have questioned your credentials and expertise from the get-go.

Boom. Skeptified.
Title: Re: Self-Driving Cars
Post by: phooey on March 23, 2016, 12:22:09 AM
Skeptiqueer,

Thank you for agreeing that mine is the accepted opinion. 

I bet you are one of those guerilla skeptics, aren't you? 
Title: Re: Self-Driving Cars
Post by: SkeptiQueer on March 23, 2016, 01:50:17 AM
I'm skeptical of your claim of skepticism, given that you've continually refused to participate in the evidence-based examination or critical discussion that is the core of skeptical discourse.

Boom. 2 for 2.
Title: Re: Self-Driving Cars
Post by: arthwollipot on March 23, 2016, 03:00:10 AM
You know why we shouldn't have smartphones?  Try shouting "Siri, text Mom, 'Eat shit'. Send"* in a crowded room.  No way to stop it, everyone's moms die of broken hearts.  Unsolvable problem.

*Or whatever the exact command would be. I don't use voice recognition (or Apple).

You'd have to hold down the home button first, or if it's plugged in, say "hey Siri" in the owner's voice before Siri would recognise the command.

In other words, that's a prank that technology already prevents, and it's therefore an absurd argument.
Title: Re: Self-Driving Cars
Post by: daniel1948 on March 23, 2016, 08:26:01 AM
You know why we shouldn't have smartphones?  Try shouting "Siri, text Mom, 'Eat shit'. Send"* in a crowded room.  No way to stop it, everyone's moms die of broken hearts.  Unsolvable problem.

*Or whatever the exact command would be. I don't use voice recognition (or Apple).

You'd have to hold down the home button first, or if it's plugged in, say "hey Siri" in the owner's voice before Siri would recognise the command.

In other words, that's a prank that technology already prevents, and it's therefore an absurd argument.

Does this mean that "Hey, Siri" won't work if you have a cold and your voice sounds different? I've never tried to use "Hey, Siri" without pressing the home button, but I'm pretty sure I've never seen a procedure for teaching Siri to recognize my voice.

But I do know that on the latest hardware "Hey, Siri" does not require that the device be plugged into power.
Title: Re: Self-Driving Cars
Post by: SkeptiQueer on March 23, 2016, 08:55:53 AM
Anybody see the vids where people were making their display name "XboxOneOff" or whatever so that people would yell the command to turn off their xbox mid-match?

Not a new problem, and one fairly quickly fixable.

For cars, the key fob+voice print seems to work, and I have yet to hear a realistic reason why multiple inputs would cause havoc instead of having the car do what a human would do and just say "please make up your mind."

I'm wondering if we'll see different styles with more autonomous and less autonomous programming (Google removing steering wheels, but maybe Sonyssan retains it but locks it forward out of the way when in autopilot, while perhaps TeslApple allows you to program in custom movement paths on a touchscreen in you want to park a certain way or pilot around an obstacle.

The idea has occurred to me that with that much integration, dash cams should be simple enough, and a small "last 5 minutes" hard drive cheap as hell. Other countries already have hire cars with anticarjacking measures, should that be an issue.

Anybody remember the video from the Ferguson protest where Dudebro drove through a crowd of people? How would a self-driving car handle that? I guess that has to start by asking how everyone would handle that, beyond just locking the doors and live-tweeting it.
Title: Re: Self-Driving Cars
Post by: arthwollipot on March 23, 2016, 07:43:35 PM
You know why we shouldn't have smartphones?  Try shouting "Siri, text Mom, 'Eat shit'. Send"* in a crowded room.  No way to stop it, everyone's moms die of broken hearts.  Unsolvable problem.

*Or whatever the exact command would be. I don't use voice recognition (or Apple).

You'd have to hold down the home button first, or if it's plugged in, say "hey Siri" in the owner's voice before Siri would recognise the command.

In other words, that's a prank that technology already prevents, and it's therefore an absurd argument.

Does this mean that "Hey, Siri" won't work if you have a cold and your voice sounds different? I've never tried to use "Hey, Siri" without pressing the home button, but I'm pretty sure I've never seen a procedure for teaching Siri to recognize my voice.

But I do know that on the latest hardware "Hey, Siri" does not require that the device be plugged into power.

On my iPhone 5S, it definitely has to be plugged in before it will recognise "Hey Siri" and I also definitely went through a process of training it to recognise my voice (it's in Settings - General - Siri). It may have changed in the 6.
Title: Re: Self-Driving Cars
Post by: DG on March 23, 2016, 09:36:48 PM
Anybody see the vids where people were making their display name "XboxOneOff" or whatever so that people would yell the command to turn off their xbox mid-match?


on the PS4, in FIFA15, the in game commentary could, if loud enough, affect the voice activation and change the in game instructions (i.e forcing the offside trap, or man marking).
Title: Re: Self-Driving Cars
Post by: arthwollipot on March 24, 2016, 04:51:13 AM
I remember when we first got our Kinect, my son would troll me while I was playing Mass Effect 3 by shouting in the middle of a battle "Garrus, go forward!"
Title: Re: Self-Driving Cars
Post by: daniel1948 on March 24, 2016, 09:26:30 AM
You know why we shouldn't have smartphones?  Try shouting "Siri, text Mom, 'Eat shit'. Send"* in a crowded room.  No way to stop it, everyone's moms die of broken hearts.  Unsolvable problem.

*Or whatever the exact command would be. I don't use voice recognition (or Apple).

You'd have to hold down the home button first, or if it's plugged in, say "hey Siri" in the owner's voice before Siri would recognise the command.

In other words, that's a prank that technology already prevents, and it's therefore an absurd argument.

Does this mean that "Hey, Siri" won't work if you have a cold and your voice sounds different? I've never tried to use "Hey, Siri" without pressing the home button, but I'm pretty sure I've never seen a procedure for teaching Siri to recognize my voice.

But I do know that on the latest hardware "Hey, Siri" does not require that the device be plugged into power.

On my iPhone 5S, it definitely has to be plugged in before it will recognise "Hey Siri" and I also definitely went through a process of training it to recognise my voice (it's in Settings - General - Siri). It may have changed in the 6.

I guess I never set up "Hey Siri" on any of my tablets. I always use the button on the very rare occasions I use Siri at all. I usually start my command with "Hey, Siri..." which I guess is unnecessary after activating it with the button. I wonder if it's as bad at verifying that the voice is yours as it is at recognizing your fingerprint. On my iPad Pro, fingerprint recognition is so bad that I quite using it and just use my passcode.

Tim Cook made a point of stressing that "Hey, Siri" now works even without power plugged in. Something to do with a new bit of hardware. He was speaking about the iPhone SE, but I think he said the iPhone 6 has the same capability. I think the iPad Pro has it also, but I've never tried it because the only times I've used Siri were just to see if it could do something, or to see what funny reply it would have to a silly question.
Title: Re: Self-Driving Cars
Post by: brilligtove on March 24, 2016, 10:25:04 AM
I use voice commands on Google all the time, and they are excellent. The voice dictation is very good too. It is one of the reasons I like Android more than iOS.

I'm surprised that your fingerprint reader is so bad. Mine is generally pretty good.
Title: Re: Self-Driving Cars
Post by: amysrevenge on March 24, 2016, 10:59:26 AM
You know why we shouldn't have smartphones?  Try shouting "Siri, text Mom, 'Eat shit'. Send"* in a crowded room.  No way to stop it, everyone's moms die of broken hearts.  Unsolvable problem.

*Or whatever the exact command would be. I don't use voice recognition (or Apple).

You'd have to hold down the home button first, or if it's plugged in, say "hey Siri" in the owner's voice before Siri would recognise the command.

In other words, that's a prank that technology already prevents, and it's therefore an absurd argument.

So the whole "what about conflicting commands within the car" argument is garbage, then?  Well.  How about that.
Title: Re: Self-Driving Cars
Post by: daniel1948 on March 24, 2016, 11:18:33 AM
But wait a minute! What if your passenger imitates your voice and shouts "Hey, Siri! Crash the car!"  :mmmhmm:  :raise:
Title: Re: Self-Driving Cars
Post by: brilligtove on March 24, 2016, 11:19:29 AM
Just stick a banana in the tailpipe and be done with it already.
Title: Re: Self-Driving Cars
Post by: SkeptiQueer on March 24, 2016, 12:13:17 PM
But wait a minute! What if your passenger imitates your voice and shouts "Hey, Siri! Crash the car!"  :mmmhmm:  :raise:
This just in: League of Ventriloquists claim responsibility for terror in the streets!
Title: Re: Self-Driving Cars
Post by: Johnny Slick on March 24, 2016, 12:35:28 PM
Just stick a banana in the tailpipe and be done with it already.
What if it's a talking banana and it tells Siri to take you to the banana split store??
Title: Re: Self-Driving Cars
Post by: daniel1948 on March 24, 2016, 01:57:37 PM
Just stick a banana in the tailpipe and be done with it already.
What if it's a talking banana and it tells Siri to take you to the banana split store??

A masochist/suicidal banana?
Title: Re: Self-Driving Cars
Post by: Johnny Slick on March 24, 2016, 02:47:16 PM
Look. I don't go around asking bananas about what gets their banana rocks off, and I suggest you follow this practice as well. Let's just leave it at this: in a world where bananas can talk, this is within the realm of possibility.
Title: Re: Self-Driving Cars
Post by: daniel1948 on March 24, 2016, 03:03:46 PM
Would there even be banana split stores in a world where bananas could talk? I mean, wouldn't they protest them, perhaps by scattering banana peels around the entrance to the store to keep people out, the way the Brits used to spread flypaper around their safari camps in India to keep tigers out?
Title: Re: Self-Driving Cars
Post by: amysrevenge on March 24, 2016, 03:11:43 PM
Would there even be banana split stores in a world where bananas could talk? I mean, wouldn't they protest them, perhaps by scattering banana peels around the entrance to the store to keep people out, the way the Brits used to spread flypaper around their safari camps in India to keep tigers out?

If you've seen Bojack Horseman, consider the chicken farmer played by Ron Funches...
Title: Re: Self-Driving Cars
Post by: Jeremy's Sea on March 24, 2016, 03:39:54 PM
Would there even be banana split stores in a world where bananas could talk? I mean, wouldn't they protest them, perhaps by scattering banana peels around the entrance to the store to keep people out, the way the Brits used to spread flypaper around their safari camps in India to keep tigers out?
*thread drift extreme*
We're still butchering whales and dolphins, and we're still using rare metals from Africa dug up by children under the duress of warlords, I'm sure we'd still eat banana splits...
Title: Re: Self-Driving Cars
Post by: phooey on March 24, 2016, 05:50:08 PM
You know why we shouldn't have smartphones?  Try shouting "Siri, text Mom, 'Eat shit'. Send"* in a crowded room.  No way to stop it, everyone's moms die of broken hearts.  Unsolvable problem.

*Or whatever the exact command would be. I don't use voice recognition (or Apple).

You'd have to hold down the home button first, or if it's plugged in, say "hey Siri" in the owner's voice before Siri would recognise the command.

In other words, that's a prank that technology already prevents, and it's therefore an absurd argument.

So the whole "what about conflicting commands within the car" argument is garbage, then?  Well.  How about that.

Uh yea, actually no Amy.  The skeptics have missed the point once again, so they are off on this idea that the problem is the car knowing who the owner is.  For any number of reasons the car can't be designed so that only ONE person can ever control it.  If I need to explain all the reason why this is so, there is no point, because that would mean I am dealing with people who have no clue at all. 

Oh, and you have crushed skeptiqueers dream of not owning a car, and calling up Uber for a lift.  You people are so cruel. 
Title: Re: Self-Driving Cars
Post by: arthwollipot on March 24, 2016, 08:21:51 PM
You know why we shouldn't have smartphones?  Try shouting "Siri, text Mom, 'Eat shit'. Send"* in a crowded room.  No way to stop it, everyone's moms die of broken hearts.  Unsolvable problem.

*Or whatever the exact command would be. I don't use voice recognition (or Apple).

You'd have to hold down the home button first, or if it's plugged in, say "hey Siri" in the owner's voice before Siri would recognise the command.

In other words, that's a prank that technology already prevents, and it's therefore an absurd argument.

So the whole "what about conflicting commands within the car" argument is garbage, then?  Well.  How about that.

Uh yea, actually no Amy.  The skeptics have missed the point once again, so they are off on this idea that the problem is the car knowing who the owner is.  For any number of reasons the car can't be designed so that only ONE person can ever control it.  If I need to explain all the reason why this is so, there is no point, because that would mean I am dealing with people who have no clue at all.

Nobody is in any way disputing that you can't have a single-person car. We're just disputing that this is even remotely a problem.
Title: Re: Self-Driving Cars
Post by: brilligtove on March 24, 2016, 09:21:46 PM
Aww. Phoobot is soooooooo cute when it tries to pretend to be an intelligent agent.
Title: Re: Self-Driving Cars
Post by: Morvis13 on March 24, 2016, 10:04:21 PM
You know why we shouldn't have smartphones?  Try shouting "Siri, text Mom, 'Eat shit'. Send"* in a crowded room.  No way to stop it, everyone's moms die of broken hearts.  Unsolvable problem.

*Or whatever the exact command would be. I don't use voice recognition (or Apple).

You'd have to hold down the home button first, or if it's plugged in, say "hey Siri" in the owner's voice before Siri would recognise the command.

In other words, that's a prank that technology already prevents, and it's therefore an absurd argument.

So the whole "what about conflicting commands within the car" argument is garbage, then?  Well.  How about that.

Uh yea, actually no Amy.  The skeptics have missed the point once again, so they are off on this idea that the problem is the car knowing who the owner is.  For any number of reasons the car can't be designed so that only ONE person can ever control it.  If I need to explain all the reason why this is so, there is no point, because that would mean I am dealing with people who have no clue at all. 

Oh, and you have crushed skeptiqueers dream of not owning a car, and calling up Uber for a lift.  You people are so cruel.

Yeah I think you do kind of have to explain why you think the design is limited in this way. If I am the only owner and have the only key then I'm the only one that can drive the car. That is the way current cars work. Sure I can give my key to my mechanic to change my snowtires. I'm pretty sure that mean they can drive my car now. Am I missing something?
Title: Re: Self-Driving Cars
Post by: brilligtove on March 24, 2016, 10:18:03 PM
SERIOUSLY, Morvis?! I'm disappointed in you. Clearly you have missed the obvious reality that all mechanics are evile.

...Oh wait, I may be drunk.
Title: Re: Self-Driving Cars
Post by: Jeremy's Sea on March 25, 2016, 10:46:41 AM
SERIOUSLY, Morvis?! I'm disappointed in you. Clearly you have missed the obvious reality that all mechanics are evile.

...Oh wait, I may be drunk.
No, no. Mechanics are evil. Also you may still be drunk.
Title: Re: Self-Driving Cars
Post by: brilligtove on March 25, 2016, 12:52:07 PM
In boost to self-driving cars, computers can qualify as drivers (http://www.reuters.com/article/us-alphabet-autos-selfdriving-exclusive-idUSKCN0VJ00H)

Quote
The National Highway Traffic Safety Administration told Google ... of its decision in a previously unreported Feb. 4 letter to the company posted on the agency's website this week.

...

"NHTSA will interpret 'driver' in the context of Google's described motor vehicle design as referring to the (self-driving system), and not to any of the vehicle occupants," NHTSA's letter said.

"We agree with Google its (self-driving car) will not have a 'driver' in the traditional sense that vehicles have had drivers during the last more than one hundred years."

Pretty cool!
Title: Re: Self-Driving Cars
Post by: daniel1948 on March 25, 2016, 01:49:02 PM
Nice!
Title: Re: Self-Driving Cars
Post by: Jeremy's Sea on March 25, 2016, 02:42:50 PM
Think of it like horses, they kind of drove too and mass chaos didn't ensue. This is like an intelligent horse.
Title: Re: Self-Driving Cars
Post by: brilligtove on March 25, 2016, 04:43:17 PM
Think of it like horses, they kind of drove too and mass chaos didn't ensue. This is like an intelligent horse.

My grandfather, born in 1901, told me that the great thing about horses I'd you could drive the buggy out for an hour, get in the back, and let the horse take you home for an hour. Hard to get caught with your pants down when you're on the move. :)
Title: Re: Self-Driving Cars
Post by: daniel1948 on March 25, 2016, 05:37:10 PM
And of course, if you lived on a farm, you grew your own fuel and the exhaust helped to fertilize your fields.
Title: Re: Self-Driving Cars
Post by: Jeremy's Sea on March 25, 2016, 06:58:08 PM
Google needs to invent this horse thing.
Title: Re: Self-Driving Cars
Post by: phooey on March 25, 2016, 07:56:56 PM
"Google "expresses concern that providing human occupants of the vehicle with mechanisms to control things like steering, acceleration, braking... could be detrimental to safety because the human occupants could attempt to override the (self-driving system's) decisions," the NHTSA letter stated."


So no steering wheel, no brake pedals, and the car can only ever take commands from one person. 

Gee, this sounds like Shangri-la.  I can't imagine any problems with that at all. 
Title: Re: Self-Driving Cars
Post by: daniel1948 on March 25, 2016, 08:02:03 PM
No more idiot drivers, and the car takes commands from the owner and whoever the owner authorizes. I don't know where you came up with the squirrely notion that a computer-driven car could not be authorized to have more than one driver.
Title: Re: Self-Driving Cars
Post by: Morvis13 on March 25, 2016, 08:14:31 PM
"Google "expresses concern that providing human occupants of the vehicle with mechanisms to control things like steering, acceleration, braking... could be detrimental to safety because the human occupants could attempt to override the (self-driving system's) decisions," the NHTSA letter stated."


So no steering wheel, no brake pedals, and the car can only ever take commands from one person. 

Gee, this sounds like Shangri-la.  I can't imagine any problems with that at all.

(http://forums.windowscentral.com/attachments/ask-question/92711d1420167753t-wikipedian_protester.png)
Title: Re: Self-Driving Cars
Post by: phooey on March 25, 2016, 08:32:16 PM
Citation for what??  That was a quote from the article that brillingtrove just posted. 
Title: Re: Self-Driving Cars
Post by: phooey on March 25, 2016, 08:34:48 PM
I didn't come up with the squirrely notion that the car could only have one person to take commands from, you all did ( I already knew that was silly) .  I said from the beginning that that would be impossible.  It was you guys talking about Siri not getting confused by other voices, not me. 
Title: Re: Self-Driving Cars
Post by: arthwollipot on March 25, 2016, 08:41:30 PM
"Google "expresses concern that providing human occupants of the vehicle with mechanisms to control things like steering, acceleration, braking... could be detrimental to safety because the human occupants could attempt to override the (self-driving system's) decisions," the NHTSA letter stated."


So no steering wheel, no brake pedals, and the car can only ever take commands from one person. 

Gee, this sounds like Shangri-la.  I can't imagine any problems with that at all.

For a start, the car only taking commands from one person is in no way suggested in the excerpt you quoted. Second, in what way is more than one person allowed to be in control of a car now?

You know that today you can get cars that have individually coded keyfobs, which will automatically adjust the seat, mirror and steering wheel position depending on which person is getting into the car? That technology has been around for a decade now. Does the fact that the keyfob is individually coded to my preferred driving position mean that no-one else can ever drive it? No, because the car recognises more than one individual. Why would that existing technology not be applied to a self-driving car?
Title: Re: Self-Driving Cars
Post by: daniel1948 on March 25, 2016, 10:00:18 PM
I didn't come up with the squirrely notion that the car could only have one person to take commands from, you all did ( I already knew that was silly) .  I said from the beginning that that would be impossible.  It was you guys talking about Siri not getting confused by other voices, not me. 

Nobody ever said the car could only be controlled by one person. What we said was that the car would know who is authorized to control it at any given time (the person whose key or fob turned it on, or the person who activated the controls upon starting it). That's an entirely different thing than the car only responding to one person ever. This should be obvious to anyone paying attention.
Title: Re: Self-Driving Cars
Post by: brilligtove on March 25, 2016, 10:27:22 PM
I have started thinking about the short and medium term implications of gradual introduction of self driving vehicles in the next decade. I don't have any particular insights, but some questions come to mind.

Title: Re: Self-Driving Cars
Post by: arthwollipot on March 25, 2016, 11:02:37 PM
Well, Uber is reportedly entering the market for self-driving cars (http://www.reuters.com/article/us-daimler-uber-idUSKCN0WK1C8), so I assume that they're positioning themselves in that space.

I think long-haul goods and mass public transport will be the two markets that self-driving cars will take over first.
Title: Re: Self-Driving Cars
Post by: brilligtove on March 25, 2016, 11:10:13 PM
I just wonder how that takeover will progress. Will truckers be allowed to sleep on the highways to start with, and then take over when conditions are more difficult? Then be required to have hands-on-wheel oversight/override for difficult conditions? (That's a situation that Google wants to never see happen, if I understand their position correctly.) And then be out of work?

Autonomous cars will be a long time coming in places that are not Silicon Valley, of course, for regulatory and engineering reasons. Uber works in a lot of those markets... Kudos for long term strategy. This is cool:

Quote
Earlier this week Mercedes rival BMW (BMWG.DE) said it was considering launching its own ride hailing service in what would amount to a rival business to Uber.

I didn't know about that. There's a lot of money to be made in the horseless carriage game, it seems.
Title: Re: Self-Driving Cars
Post by: PANTS! on March 25, 2016, 11:25:57 PM
Unless there is union influence, I bet that truckers will go straight in the dustbin - no transition for them. If that were my job, and I were under 55, I'd be shittin my Dickies.

I do not think there will be any dual driver options in any car released to market.  Either that or I expect it will be like a 4WD ootion.  You will have to stop and makre a switch, and once you do it is no longer street legal.  My bet is that insurance companies will not want to deal with it, nor the manufacturers.  I bet that the auto-automobiles will simply slow down to super slow speeds when faced with nasty conditions.

And yes, there will be times when that restriction sucks.  I love driving on snow.  Or in a back field filled with mud.  But, I would love to automate my morning commute more. And so probably will everyone else.  Or if they don't their insurance bill is going to go through the roof when their risk pool plummets.

All just me spitballing, based off how the transition to elevators went, and the transition to autopilot is going.

My fucking God!  I build commercial real estate, and once of the biggest pains in my ass is parking.  So much space in a modern city is devoted to car storage.  But now you will be able to tell you car to go park in a central storage place, and retreive it when you need it again. 
Title: Re: Self-Driving Cars
Post by: brilligtove on March 25, 2016, 11:43:41 PM
Unless there is union influence, I bet that truckers will go straight in the dustbin - no transition for them. If that were my job, and I were under 55, I'd be shittin my Dickies.

I do not think there will be any dual driver options in any car released to market.  Either that or I expect it will be like a 4WD ootion.  You will have to stop and makre a switch, and once you do it is no longer street legal.  My bet is that insurance companies will not want to deal with it, nor the manufacturers.  I bet that the auto-automobiles will simply slow down to super slow speeds when faced with nasty conditions.

And yes, there will be times when that restriction sucks.  I love driving on snow.  Or in a back field filled with mud.  But, I would love to automate my morning commute more. And so probably will everyone else.  Or if they don't their insurance bill is going to go through the roof when their risk pool plummets.

All just me spitballing, based off how the transition to elevators went, and the transition to autopilot is going.

My fucking God!  I build commercial real estate, and once of the biggest pains in my ass is parking.  So much space in a modern city is devoted to car storage.  But now you will be able to tell you car to go park in a central storage place, and retreive it when you need it again.

Well I did ask for balls slathered in spit, so thank you for that. :)

I see what you're saying about truckers. I think couriers will have a place for a while longer - at least until autonomodrones (autodrones? intelidrones?) are common. As the recipient of medical supplies, I can see a need for human deliveries of several categories of products.

As I consider it, a drone could deliver supplies (like pharmaceuticals) that are GPS tracked and thumbprint-locked to a registered nurse (for example). That way the supplies would be on hand when the person who can use/administer them arrives.

Interesting...
Title: Re: Self-Driving Cars
Post by: Captain Video on March 26, 2016, 12:05:52 AM
Unless there is union influence, I bet that truckers will go straight in the dustbin - no transition for them. If that were my job, and I were under 55, I'd be shittin my Dickies.

I do not think there will be any dual driver options in any car released to market.  Either that or I expect it will be like a 4WD ootion.  You will have to stop and makre a switch, and once you do it is no longer street legal.  My bet is that insurance companies will not want to deal with it, nor the manufacturers.  I bet that the auto-automobiles will simply slow down to super slow speeds when faced with nasty conditions.

And yes, there will be times when that restriction sucks.  I love driving on snow.  Or in a back field filled with mud.  But, I would love to automate my morning commute more. And so probably will everyone else.  Or if they don't their insurance bill is going to go through the roof when their risk pool plummets.

All just me spitballing, based off how the transition to elevators went, and the transition to autopilot is going.

My fucking God!  I build commercial real estate, and once of the biggest pains in my ass is parking.  So much space in a modern city is devoted to car storage.  But now you will be able to tell you car to go park in a central storage place, and retreive it when you need it again.

Well I did ask for balls slathered in spit, so thank you for that. :)

I see what you're saying about truckers. I think couriers will have a place for a while longer - at least until autonomodrones (autodrones? intelidrones?) are common. As the recipient of medical supplies, I can see a need for human deliveries of several categories of products.

As I consider it, a drone could deliver supplies (like pharmaceuticals) that are GPS tracked and thumbprint-locked to a registered nurse (for example). That way the supplies would be on hand when the person who can use/administer them arrives.

Interesting...

I was once forced to listen to a Lyft driver tell me how much Uber sucks because it is trying to eliminate driver jobs all together with automaton, I kept thinking "man I cant wait till that happens so I don't have to listen to guys like this any more"  Still 1000 times better than a cab and I prefer a Lyft front seat over an Uber back seat.

As a Union man I have to say I would side with any teamster strike over automation of the trucker industry wile at the same time trying to convince them to learn new tech skills so they can find new jobs controlling said trucks, Its a tough call, Computers will run the industry so much safer but that's a ton of people out of work.  I suppose computers have put millions of other jobs in the grave but created just as many. Its a tough thing to be on either side of the argument.
Title: Re: Self-Driving Cars
Post by: SkeptiQueer on March 26, 2016, 12:10:18 AM
So, trains are self-piloting for the most part. They can't entirely self-pilot because of the time it takes to stop, but moreso because of the unions, which lobbied for safety rules to require the engineer and conductor to remain awake...except for some of the trains that run a small enough load on a short enough run, some of them get away with a single engineer and the train doing the rest.

I imagine the teamsters doing something similar, requiring an awake and alert driver to monitor the truck.

I'm sure a market will remain for vehicles with a driver-operated option, because 'Merika, but I imagine that will be limited by the cost of insurance.
Title: Re: Self-Driving Cars
Post by: bimble on March 26, 2016, 06:37:11 AM
You will always want someone present in commercial vehicles, if only to check the load, load/unload, refuel the vehicle. An autonomous van can't take a parcel out of the back and take it to reception/front door for delivery.
Title: Re: Self-Driving Cars
Post by: PANTS! on March 26, 2016, 07:34:18 AM
Parcels, maybe.  Like UPS drivers I see having a greatly improved job.  Refueling will assuredly be automated.  Truckers don't load and empty their trucks now.  I am pretty sure that will not change.
Title: Re: Self-Driving Cars
Post by: phooey on March 26, 2016, 08:09:13 AM
In fact, studies have shown that not only would congestion likely be worse with automated cars, but also computer simulations showed that self driving cars on the road would be much slower, because they are not able to make the kind of traffic flowing maneuvers humans do as quickly, because of the need to keep certain distances and allow for more margin of error. 

So your commute will likely take longer.
Title: Re: Self-Driving Cars
Post by: phooey on March 26, 2016, 08:18:52 AM
Brillingtrove,

"The first stage is driver-assist technology, as we have today. I'd expect the next step to be autopilot on highways under good road conditions. What's the step after that?"

That is exactly the question.  I think it is not that hard to imagine automated vehicles where there are no intersections, or pedestrians or mixed speed traffic.  I don't think that is the challenge.  Once you start dealing with the entire spectrum of flow of people and vehicles and animals, and traffic lights, and emergency vehicles, and kids playing in the road, its gets quite a bit more tricky.  Anyone who has driven in any foreign country, or even in places like Manhattan knows that its a whole different ball game out there.  Trying to mix humans and computers driving at the same time is going to be a cluster fuck for a long time. 

As far as the truckers go, I wouldn't worry too much.  The trucks eventually have to come off the highway, then who is going to drive them.  If there are no more truck drivers with skills, well, there will be a lot of abandoned loads by the side of the road, trying to download the latest Google maps, and getting a "no signal available right now" message.   

Title: Re: Self-Driving Cars
Post by: Johnny Slick on March 26, 2016, 08:38:00 AM
Google needs to invent this horse thing.
Really, so called skeptic? What's a horse going to do when a panhandling banana jumps on top of it and demands oats?
Title: Re: Self-Driving Cars
Post by: daniel1948 on March 26, 2016, 09:31:45 AM
Google needs to invent this horse thing.
Really, so called skeptic? What's a horse going to do when a panhandling banana jumps on top of it and demands oats?

A banana that eats oats? What is this, some kind of cross-kingdom GMO banana?
Title: Re: Self-Driving Cars
Post by: daniel1948 on March 26, 2016, 09:47:04 AM
The transition to computer-driven cars will certainly include hybrid human/computer-driven cars. Effectively we already have that now with adaptive cruise control and self-parking cars. I see the goal as pure computer, but to get there we may need intermediate stages.

I don't think mass transit will shift to individual cars, but there will certainly be cab companies and Uber or Uber-like companies that will use computer-driven cars. There will also be subscription services where you pay a fixed price and a computer-driven car owned and maintained by a company picks you up and takes you to and from work. The efficiency of not leaving the car parked most of the day will reduce the total cost of transportation. And you no longer need to hassle with maintenance and repairs. This will also mean fewer total cars on the road and less congestion, though if population continues to grow, traffic congestion will increase no matter what kind of cars are used.

It's good to hear that Uber is planning for computer-driven cars, but I think it's far too early. They're still too far out for anybody to be selecting a manufacturer, much less making arrangements to actually buy them. I personally like Tesla best, but in a decade, when the cars are actually available, I'll buy the best car from whoever is making it. I'm guessing Tesla, but not placing my order until the car is built.

And of course, once the technology is mature and all the cars on the road have it, and they can communicate with each other, traffic flow will be much more efficient. No more sitting at a red light when there are no cars crossing your path. The computer can allow cars through whenever it's clear.
Title: Re: Self-Driving Cars
Post by: PANTS! on March 26, 2016, 11:14:51 AM
The transition to computer-driven cars will certainly include hybrid human/computer-driven cars. Effectively we already have that now with adaptive cruise control and self-parking cars. I see the goal as pure computer, but to get there we may need intermediate stages.


I guess thats fair.  As you say, we already have the humcomp hybrids.  When I  said there would likely not be such hybrids, think I was envisioning a mainly computer driven vehicle with a button to switch to human steering whenever the human wants to.  Which I still contend will not work.  We have tons of evidence that it does not work. 
Title: Re: Self-Driving Cars
Post by: SkeptiQueer on March 26, 2016, 11:20:53 AM
In fact, studies have shown that not only would congestion likely be worse with automated cars, but also computer simulations showed that self driving cars on the road would be much slower, because they are not able to make the kind of traffic flowing maneuvers humans do as quickly, because of the need to keep certain distances and allow for more margin of error. 

So your commute will likely take longer.
[CITATION NEEDED]
Title: Re: Self-Driving Cars
Post by: PANTS! on March 26, 2016, 12:42:25 PM
Oh, I fully expect many commutes to take longer - including, mine because I don't drive the speed limit.
Title: Re: Self-Driving Cars
Post by: wastrel on March 26, 2016, 01:23:58 PM
In fact, studies have shown that not only would congestion likely be worse with automated cars, but also computer simulations showed that self driving cars on the road would be much slower, because they are not able to make the kind of traffic flowing maneuvers humans do as quickly, because of the need to keep certain distances and allow for more margin of error. 

So your commute will likely take longer.

Oh perfect, we are finally talking about actual studies.  So....links?
Title: Re: Self-Driving Cars
Post by: daniel1948 on March 26, 2016, 06:07:10 PM
The transition to computer-driven cars will certainly include hybrid human/computer-driven cars. Effectively we already have that now with adaptive cruise control and self-parking cars. I see the goal as pure computer, but to get there we may need intermediate stages.


I guess thats fair.  As you say, we already have the humcomp hybrids.  When I  said there would likely not be such hybrids, think I was envisioning a mainly computer driven vehicle with a button to switch to human steering whenever the human wants to.  Which I still contend will not work.  We have tons of evidence that it does not work. 

I actually think the button-to-switch model could work EXCEPT that once cars are able to drive without human intervention nobody will want humans to be allowed to drive. Computers will be so much safer, that we'll outlaw human drivers on public roads. There will definitely be tracks where people can drive cars.

Of course phooey has it backwards: Computer-controlled cars will be able to drive closer together and eliminate wasted time at intersections, reducing commuter time (unless you habitually drive over the limit.)

Imagine this: Networked cars that at low-traffic times can adjust their speed slightly to zip through intersections without the need for traffic lights. Car trains that on the highway can reduce distance between cars because they're all networked and if an obstacle is detected by one, all the cars behind it can brake simultaneously. Seamless merging. A fully-networked highway is probably a decade behind fully computer driven cars, but if we don't nuke ourselves off the face of the planet or squander all the energy before building renewable generation, it will come. I probably won't be here to see it if my time line is accurate.
Title: Re: Self-Driving Cars
Post by: SkeptiQueer on March 26, 2016, 08:04:35 PM
Commutes that depend on people driving 75 in a 55 will be slower, naturally. OTOH, my commute is 13 miles and 20 minutes, and I drive like a stoned grandmother with a cop behind her. My commute can only get faster if more cars are following speed limits, merge orders, and not fighting to get there 30 seconds faster.

There's also a decent possibility to expect robot-only bypasses running at speeds that are still road-safe, but not safe for a driver that takes a half second to respond at best.
Title: Re: Self-Driving Cars
Post by: arthwollipot on March 26, 2016, 08:23:34 PM
I was once forced to listen to a Lyft driver tell me how much Uber sucks because it is trying to eliminate driver jobs all together with automaton, I kept thinking "man I cant wait till that happens so I don't have to listen to guys like this any more"  Still 1000 times better than a cab and I prefer a Lyft front seat over an Uber back seat.
You're more than welcome to use the front seat in my Uber. Complementary bottle of water in the door pocket.
Title: Re: Self-Driving Cars
Post by: PANTS! on March 26, 2016, 08:24:24 PM
In fact, studies have shown that not only would congestion likely be worse with automated cars, but also computer simulations showed that self driving cars on the road would be much slower, because they are not able to make the kind of traffic flowing maneuvers humans do as quickly, because of the need to keep certain distances and allow for more margin of error. 

So your commute will likely take longer.

Oh perfect, we are finally talking about actual studies.  So....links?

Oh for God's sakes Wastrel.  Quit talking to the shouty vagrant with the wack ass ideas. 
Title: Re: Self-Driving Cars
Post by: arthwollipot on March 26, 2016, 08:29:53 PM
Yes, I believe that the truck and bus and taxi, and yes, Uber drivers like me, will be put out of a job. Just like the gas lamp lighters and the buggy whip manufacturers. It's a well-known economic concept known as creative destruction (https://en.wikipedia.org/wiki/Creative_destruction).

Those people will lose their jobs, but they'll get other jobs. Better jobs. That's what happens.
Title: Re: Self-Driving Cars
Post by: daniel1948 on March 26, 2016, 10:43:14 PM
I believe it's already been mentioned, way back in the thread, but bears repeating: Even a longer commute (due to driving at the speed limit when you were accustomed to driving way over the limit) would be a gain since the time could be used to work or nap in the car. Some cars will be sold with seats that lay down flat into a bed. Some will have a workspace: desk & internet-connected computer. Some will have both. An SUV might have an exercise bike. (Which might seem ridiculous, because why not bicycle to work, but if your commute is 50 miles each way and there's snow on the ground, cycling would not be practical, but you could get your work-out while your car takes you to work, and take a nap while it takes you home.)
Title: Re: Self-Driving Cars
Post by: teethering on March 26, 2016, 11:00:16 PM
In fact, studies have shown that not only would congestion likely be worse with automated cars, but also computer simulations showed that self driving cars on the road would be much slower, because they are not able to make the kind of traffic flowing maneuvers humans do as quickly, because of the need to keep certain distances and allow for more margin of error. 

So your commute will likely take longer.

Oh perfect, we are finally talking about actual studies.  So....links?

Oh for God's sakes Wastrel.  Quit talking to the shouty vagrant with the wack ass ideas.

I'm suspecting our friend here is trying to pretend that the study he's talking about is making a general prediction about self-driving cars in general, rather than simulating the specific conditions of "making self-driving cars as comfortable as light rail".

http://www.citylab.com/tech/2015/01/how-driverless-cars-could-make-traffic-dramatically-worse/384821/
Title: Re: Self-Driving Cars
Post by: arthwollipot on March 26, 2016, 11:34:08 PM
I believe it's already been mentioned, way back in the thread, but bears repeating: Even a longer commute (due to driving at the speed limit when you were accustomed to driving way over the limit) would be a gain since the time could be used to work or nap in the car. Some cars will be sold with seats that lay down flat into a bed. Some will have a workspace: desk & internet-connected computer. Some will have both. An SUV might have an exercise bike. (Which might seem ridiculous, because why not bicycle to work, but if your commute is 50 miles each way and there's snow on the ground, cycling would not be practical, but you could get your work-out while your car takes you to work, and take a nap while it takes you home.)

You could even harness the work you do on the bike to recharge the car's batteries.

What, does anyone seriously think these self-drivers aren't going to be electric?
Title: Re: Self-Driving Cars
Post by: SkeptiQueer on March 26, 2016, 11:42:43 PM
I agree, electric makes the most sense for a self-driving commuter. Probably not for freight.

Should the finances allow, SO will probably get an EV of some sort to replace the aging cruiseliner. I likely won't until someone makes a decent light SUV that can handle 200mi or so on a charge.

The idea of an exercise bike is certainly intriguing.
Title: Re: Self-Driving Cars
Post by: Alex Simmons on March 27, 2016, 12:26:19 AM
(http://autoguide.com.vsassets.com/blog/wp-content/uploads/2012/04/original.jpg)

http://www.autoguide.com/auto-news/2012/04/stationary-bike-makes-suv-more-like-mobile-gym.html
Title: Re: Self-Driving Cars
Post by: arthwollipot on March 27, 2016, 12:39:21 AM
I'd need a shower when I got to work though, and I hate showering at work.
Title: Re: Self-Driving Cars
Post by: teethering on March 27, 2016, 01:38:09 AM
I agree, electric makes the most sense for a self-driving commuter. Probably not for freight.

Should the finances allow, SO will probably get an EV of some sort to replace the aging cruiseliner. I likely won't until someone makes a decent light SUV that can handle 200mi or so on a charge.

The idea of an exercise bike is certainly intriguing.

Why not freight?  I know very little about car things, but my understanding is that electric engines have a lot of torque and are very efficient, which sounds like that would be good for hauling freight.
Title: Re: Self-Driving Cars
Post by: SkeptiQueer on March 27, 2016, 01:47:35 AM
I agree, electric makes the most sense for a self-driving commuter. Probably not for freight.

Should the finances allow, SO will probably get an EV of some sort to replace the aging cruiseliner. I likely won't until someone makes a decent light SUV that can handle 200mi or so on a charge.

The idea of an exercise bike is certainly intriguing.

Why not freight?  I know very little about car things, but my understanding is that electric engines have a lot of torque and are very efficient, which sounds like that would be good for hauling freight.
Fuel density. A Peterbuilt 379 with a 550hp diesel runs a 110gal tank and gets 4-6mpg. I don't think there's a comparable electric that can put out 410kw for 500mi and hot-swap a battery in 15 minutes.
Title: Re: Self-Driving Cars
Post by: phooey on March 27, 2016, 06:34:28 AM
There also aren't self-driving cars you can sleep in now.  And yet you are skeptical about the potential for battery power??

Ok

Title: Re: Self-Driving Cars
Post by: arthwollipot on March 27, 2016, 07:03:03 AM
Fuel density. A Peterbuilt 379 with a 550hp diesel runs a 110gal tank and gets 4-6mpg. I don't think there's a comparable electric that can put out 410kw for 500mi and hot-swap a battery in 15 minutes.

The Tesla Model S (https://www.teslamotors.com/en_AU/models?redirect=no) has a maximum range of 550km on a single charge with a top speed of 250kph and reaches 100kph in 4.4 seconds. Is that what you're referring to?
Title: Re: Self-Driving Cars
Post by: SkeptiQueer on March 27, 2016, 07:39:38 AM
Fuel density. A Peterbuilt 379 with a 550hp diesel runs a 110gal tank and gets 4-6mpg. I don't think there's a comparable electric that can put out 410kw for 500mi and hot-swap a battery in 15 minutes.

The Tesla Model S (https://www.teslamotors.com/en_AU/models?redirect=no) has a maximum range of 550km on a single charge with a top speed of 250kph and reaches 100kph in 4.4 seconds. Is that what you're referring to?

Not quite. The biggest and baddest tesla S battery is the 310 megajoule model, right? 500km range, curb weight tops out around 2200kg.

Feel free to double check my math, but the 110 gallon tank on a standard tractor (I know international nomenclature varies, I mean the rig pulling a standard van/trailer) is holding upwards of 15 000 megajoules in the form of diesel, hauling a total curb weight up to 38,000kg (I rounded, 80,000lb is the max gross for most US highways) travelling up and down hills at 65mph, usually 12hr per day and 800+ km in that time. Team drivers will keep the truck running as many hours of possible by swapping drivers.

Now, the last piece of the puzzle. A Tesla 85kw/h 310 mega joule battery weighs 1200lb/544kg. 110 US gallons of diesel weighs 825lb/380kg(ish). Fuel tank weighs another 150-ish lb, the engine weighs another 3000lb. Without factoring in the weight of the motors for the tesla or the transmission for the truck, you would need about 50 of 85kw/h battery packs, weighing a total of 60000lb, or about the entire load capacity of a 54' trailer, to have the same raw power available.

Now I realize this is bad engineering at it's finest. I haven't calculated whether the tesla manages to turn battery power into torque 10x more efficiently than a Cummins ISX15 turns diesel into torque, but no matter which way you slice it diesel packs about 4kw/h per pound, and the best commercial battery packs on the market get .07kw/h per pound.

The fuel density for electric doesn't scale well for the loads. Same reason you can reasonably well run UAVs on electric, but not a Cessna, much less a C130.
Title: Re: Self-Driving Cars
Post by: phooey on March 27, 2016, 07:47:32 AM
In fact, studies have shown that not only would congestion likely be worse with automated cars, but also computer simulations showed that self driving cars on the road would be much slower, because they are not able to make the kind of traffic flowing maneuvers humans do as quickly, because of the need to keep certain distances and allow for more margin of error. 

So your commute will likely take longer.

Oh perfect, we are finally talking about actual studies.  So....links?

Oh for God's sakes Wastrel.  Quit talking to the shouty vagrant with the wack ass ideas.

I'm suspecting our friend here is trying to pretend that the study he's talking about is making a general prediction about self-driving cars in general, rather than simulating the specific conditions of "making self-driving cars as comfortable as light rail".

http://www.citylab.com/tech/2015/01/how-driverless-cars-could-make-traffic-dramatically-worse/384821/

That is only one of the many studies that suggest this problem.  There is a computer model which shows that self-driving cars are slower not just because they drive the speed limit, but also because they brake more conservatively and because they don't drive as aggressively to fill in gaps in traffic or to switch lanes. 

That along with cars being parked less, and driving to more destinations is predicted to increase, not decrease traffic. 

I don't save every article, and magazine I read, so I guess you all are just going to have to do some of your own research for a change. 

 
Title: Re: Self-Driving Cars
Post by: Johnny Slick on March 27, 2016, 08:31:41 AM
If you have more studies, cite the damn studies already. Stop trying to get others to do your work for you.
Title: Re: Self-Driving Cars
Post by: brilligtove on March 27, 2016, 09:16:49 AM
In fact, studies have shown that not only would congestion likely be worse with automated cars, but also computer simulations showed that self driving cars on the road would be much slower, because they are not able to make the kind of traffic flowing maneuvers humans do as quickly, because of the need to keep certain distances and allow for more margin of error. 

So your commute will likely take longer.

Oh perfect, we are finally talking about actual studies.  So....links?

Oh for God's sakes Wastrel.  Quit talking to the shouty vagrant with the wack ass ideas.

I'm suspecting our friend here is trying to pretend that the study he's talking about is making a general prediction about self-driving cars in general, rather than simulating the specific conditions of "making self-driving cars as comfortable as light rail".

http://www.citylab.com/tech/2015/01/how-driverless-cars-could-make-traffic-dramatically-worse/384821/

It never even occurred to me that anyone would want minimal acceleration from a SDV. For as long as we have human drivers in the mix, I would expect SDVs to behave a lot like humans do - well, if humans had millisecond response times and could network behaviours with other drivers. Based on the abstract I don't think they looked into network effects where SDVs can communicate and coordinate reliably.
Title: Re: Self-Driving Cars
Post by: daniel1948 on March 27, 2016, 09:56:13 AM
I agree, electric makes the most sense for a self-driving commuter. Probably not for freight.

Should the finances allow, SO will probably get an EV of some sort to replace the aging cruiseliner. I likely won't until someone makes a decent light SUV that can handle 200mi or so on a charge.

I think the Tesla Model X may be what you're looking for. Strictly speaking it's a crossover, rather than an SUV, but should have all the benefits of an SUV, and Tesla quality. I believe it's just coming to market now, as I write this. Probably around 300 miles on a charge.

I'd need a shower when I got to work though, and I hate showering at work.

Workplaces will have to (and probably will) install better showering facilities.

I agree, electric makes the most sense for a self-driving commuter. Probably not for freight.

Should the finances allow, SO will probably get an EV of some sort to replace the aging cruiseliner. I likely won't until someone makes a decent light SUV that can handle 200mi or so on a charge.

The idea of an exercise bike is certainly intriguing.

Why not freight?  I know very little about car things, but my understanding is that electric engines have a lot of torque and are very efficient, which sounds like that would be good for hauling freight.
Fuel density. A Peterbuilt 379 with a 550hp diesel runs a 110gal tank and gets 4-6mpg. I don't think there's a comparable electric that can put out 410kw for 500mi and hot-swap a battery in 15 minutes.

Batteries are the biggest problem with EVs right now. Progress is incremental. I agree that at present, an electric long-haul truck tractor is not practical.
Title: Re: Self-Driving Cars
Post by: Fast Eddie B on March 27, 2016, 10:04:26 AM
Our main driver is a 2005 Honda Element with 5-speed, bought new and coming up on 180,000 miles.

Its been quite reliable, and we plan on driving it for as long as possible - until the wheels fall off, as we put it.

We're really hoping by then plug-in hybrids continue to improve until we can get decent electric range at a reasonable price. We live in the country, and most of our driving is trips into town or to towns maybe 20 to 40 miles away. So 200 mile range would be in a sweet spot, and might work even for a pure electric. Where a pure electric would not work at that range would be our semi-regular trips to Knoxville and back - about 100 miles each way.

We also bought 60 shares of Tesla stock at various prices, now up just a bit overall - so we're especially interested in any Tesla under $40k to be our next car - if the little Element can keep its wheels on that long!
Title: Re: Self-Driving Cars
Post by: daniel1948 on March 27, 2016, 12:01:28 PM
Our main driver is a 2005 Honda Element with 5-speed, bought new and coming up on 180,000 miles.

Its been quite reliable, and we plan on driving it for as long as possible - until the wheels fall off, as we put it.

We're really hoping by then plug-in hybrids continue to improve until we can get decent electric range at a reasonable price. We live in the country, and most of our driving is trips into town or to towns maybe 20 to 40 miles away. So 200 mile range would be in a sweet spot, and might work even for a pure electric. Where a pure electric would not work at that range would be our semi-regular trips to Knoxville and back - about 100 miles each way.

We also bought 60 shares of Tesla stock at various prices, now up just a bit overall - so we're especially interested in any Tesla under $40k to be our next car - if the little Element can keep its wheels on that long!

Tesla Model 3, $35,000, will be unveiled in four days, March 31 at 8:30 p.m. Pacific time and is scheduled to go into production late 2017.

The problem with plug-in hybrids is that you have two completely separate drive systems adding weight and complexity. That's why they typically have fairly limited electric range.
Title: Re: Self-Driving Cars
Post by: teethering on March 28, 2016, 02:18:42 AM
In fact, studies have shown that not only would congestion likely be worse with automated cars, but also computer simulations showed that self driving cars on the road would be much slower, because they are not able to make the kind of traffic flowing maneuvers humans do as quickly, because of the need to keep certain distances and allow for more margin of error. 

So your commute will likely take longer.

Oh perfect, we are finally talking about actual studies.  So....links?

Oh for God's sakes Wastrel.  Quit talking to the shouty vagrant with the wack ass ideas.

I'm suspecting our friend here is trying to pretend that the study he's talking about is making a general prediction about self-driving cars in general, rather than simulating the specific conditions of "making self-driving cars as comfortable as light rail".

http://www.citylab.com/tech/2015/01/how-driverless-cars-could-make-traffic-dramatically-worse/384821/

That is only one of the many studies that suggest this problem.  There is a computer model which shows that self-driving cars are slower not just because they drive the speed limit, but also because they brake more conservatively and because they don't drive as aggressively to fill in gaps in traffic or to switch lanes. 

That along with cars being parked less, and driving to more destinations is predicted to increase, not decrease traffic. 

I don't save every article, and magazine I read, so I guess you all are just going to have to do some of your own research for a change. 

Cars driving less aggressively doesn't necessarily translate into longer commutes, even the article I linked to showed a very wide range of impact depending on the parameters of car behaviour they were simulating.  And what they were simulating was a very narrow slice of the overall traffic system, a single four-way intersection, which necessarily omits all the potential gains from computer driven cars not causing congestion by reducing numbers of accidents and so on.

The problem with you making claims without citing sources is that you are deliberately distorting what the studies are actually saying and we can't examine these caveats.  If I do your work for you and guess which studies you are referring to and go through their caveats and how they don't actually match your claims you have a chicken-shit out in the form of "but these other studies..."  which is one of many reasons why skeptics typically come armed with links to studies they're citing, to avoid this sort of endless switcharoo.
Title: Re: Self-Driving Cars
Post by: phooey on March 28, 2016, 02:33:15 AM
No teethering, actually there is no problem at all with my assertions, you are the one who has problems with it, not me. 

I don't need to convince you of anything.  You are a skeptic, so you already have your believes set in stone.  That is the paradox of the way the word is used by you folks. So I have no desire to convince you of anything.  It is for others, who actually have an open mind and wish to learn things. 

Everyone here believed that as soon as you have self-driving cars, traffic will be faster, with less congestion.  This study shows just the opposite.  I don't care about the spin you want to try to throw at it.  That is all skeptics do, try to throw spin.  Its why the SGU never allows opposing viewpoints on their show-they like to control their spin propaganda messages. 

So now people can read it and see for themselves.  I am well versed in your sides tactics.  They don't interest me in the slightest.  Its not a problem for me at all. 

 
Title: Re: Self-Driving Cars
Post by: SkeptiQueer on March 28, 2016, 03:51:38 AM
Swing and a miss, SGU has had opposing views on several times, but you'd have to have listened to the show to know. Strike two, several people have said that commutes may not be faster, particularly if they typically exceed the speed limit.

Wanna try for three?
Title: Re: Self-Driving Cars
Post by: arthwollipot on March 28, 2016, 05:07:36 AM
The problem with plug-in hybrids is that you have two completely separate drive systems adding weight and complexity. That's why they typically have fairly limited electric range.

My problem with plug-in hybrids is that you're getting power from the grid - in most places in Australia, that still means coal-fired power plants. With a non-plug-in hybrid, all you're doing is getting more distance out of the fuel that you are using. You're not drawing power from anything other than the fuel and your own momentum (regenerative braking).
Title: Re: Self-Driving Cars
Post by: phooey on March 28, 2016, 05:45:50 AM
Swing and a miss, SGU has had opposing views on several times, but you'd have to have listened to the show to know. Strike two, several people have said that commutes may not be faster, particularly if they typically exceed the speed limit.

Wanna try for three?

Citations needed!!
Title: Re: Self-Driving Cars
Post by: phooey on March 28, 2016, 06:06:43 AM
Swing and a miss, SGU has had opposing views on several times, but you'd have to have listened to the show to know. Strike two, several people have said that commutes may not be faster, particularly if they typically exceed the speed limit.

Wanna try for three?

Also no one admitted their commute would be slower UNTIL I SAID that it would be. 

Strike Two!
Title: Re: Self-Driving Cars
Post by: SkeptiQueer on March 28, 2016, 06:07:17 AM
This thread.

SGU episodes featuring guests who do not agree with the panel include, but are not limited to Neal Adams (Episode 51), B. Alan Wallace (Ep 73), Jimmy Carter (Ep 105), Michio Kaku (Ep 182), Billy West (Ep 373), Don McElroy (Ep 408).

Further, as SGU is a panel discussion show about science, it's likely not the best place to gear someone talking about a view that doesn't stand up to scientific scrutiny. If you want to hear that, I suggest "Be Reasonable" hosted by Michael Marshal of the Merseyside Sceptics, apologies for misspellings if any, where a prominent sceptic invites on promoters of what would normally be considered woo for a patient discussion. Marsh, if I can call him that, is likely the most patient man in the world given the number of ridiculous people he has calmly interviewed while being screamed at.

So you see, an actual skeptical discussion where you made a claim, and I disputed it with a counterclaim and provided evidence has just occurred, and I've given plenty of evidence to contradict your claims that skeptics just blindly accept whatever.
 
Will you accept that you were wrong in your broad character assassination and incorrect assessment of the SGU?
Title: Re: Self-Driving Cars
Post by: SkeptiQueer on March 28, 2016, 06:13:43 AM
Swing and a miss, SGU has had opposing views on several times, but you'd have to have listened to the show to know. Strike two, several people have said that commutes may not be faster, particularly if they typically exceed the speed limit.

Wanna try for three?

Also no one admitted their commute would be slower UNTIL I SAID that it would be. 

Strike Two!
Several of us changed our view when new evidence (not speeding) was introduced to the equation. That's the opposite of the blind belief you accused skeptics of.

What has been questioned repeatedly are your assertions that all traffic will be slower, that the programming will take 200 years, and that robots are sitter drivers than humans.

[Citation still needed]

Also I think a Jay mentioned cars not speeding in ep 509, like a year before you posted about slowdowns.
Title: Re: Self-Driving Cars
Post by: phooey on March 28, 2016, 06:18:49 AM
What does Jay mentioning about cars being slow have to do with this discussion.  I brought it up here, so why are you trying to take credit for it? 

You are strange.

Now about those so called opposing viewpoints on SGU broadcasts....????
Title: Re: Self-Driving Cars
Post by: SkeptiQueer on March 28, 2016, 06:19:50 AM
What does Jay mentioning about cars being slow have to do with this discussion.  I brought it up here, so why are you trying to take credit for it? 

You are strange.

Now about those so called opposing viewpoints on SGU broadcasts....????
Scroll up three posts.
Title: Re: Self-Driving Cars
Post by: Guillermo on March 28, 2016, 09:21:32 AM
In fact, studies have shown that not only would congestion likely be worse with automated cars, but also computer simulations showed that self driving cars on the road would be much slower, because they are not able to make the kind of traffic flowing maneuvers humans do as quickly, because of the need to keep certain distances and allow for more margin of error. 

So your commute will likely take longer.
Source please. I think the above is unlikely. I doubt the congestion would be worse because congestion happens mostly due to humans reacting slowly to moving flow of cars and incorrectly blocking passages things that should not happen with Autonomous cars. 
Besides, I got no problem with slower commute if I Spend it sleeping or reading.
Title: Re: Self-Driving Cars
Post by: Guillermo on March 28, 2016, 09:40:24 AM
In fact, studies have shown that not only would congestion likely be worse with automated cars, but also computer simulations showed that self driving cars on the road would be much slower, because they are not able to make the kind of traffic flowing maneuvers humans do as quickly, because of the need to keep certain distances and allow for more margin of error. 

So your commute will likely take longer.

Oh perfect, we are finally talking about actual studies.  So....links?

Oh for God's sakes Wastrel.  Quit talking to the shouty vagrant with the wack ass ideas.

I'm suspecting our friend here is trying to pretend that the study he's talking about is making a general prediction about self-driving cars in general, rather than simulating the specific conditions of "making self-driving cars as comfortable as light rail".

http://www.citylab.com/tech/2015/01/how-driverless-cars-could-make-traffic-dramatically-worse/384821/

It never even occurred to me that anyone would want minimal acceleration from a SDV. For as long as we have human drivers in the mix, I would expect SDVs to behave a lot like humans do - well, if humans had millisecond response times and could network behaviours with other drivers. Based on the abstract I don't think they looked into network effects where SDVs can communicate and coordinate reliably.
My criticism with this study is that it assumes that Autos will drive like light rails and that they will be smooth. Obviously, if you want to face backwards and be confortable, you could set it that way. But I feel that people are used to jenky driving esperiences, Anybody who has gotten on a bus  knows that the ride wont be as smooth as depicted on this paper.
Title: Re: Self-Driving Cars
Post by: daniel1948 on March 28, 2016, 09:45:49 AM
The problem with plug-in hybrids is that you have two completely separate drive systems adding weight and complexity. That's why they typically have fairly limited electric range.

My problem with plug-in hybrids is that you're getting power from the grid - in most places in Australia, that still means coal-fired power plants. With a non-plug-in hybrid, all you're doing is getting more distance out of the fuel that you are using. You're not drawing power from anything other than the fuel and your own momentum (regenerative braking).

Charge your car from the sun! A growing number of people are doing that. Depending on your location, photovoltaics on your roof can supply your car and your house with all the electricity they need. Here in the U.S., if you qualify, Solar City will install those panels for you and all you pay for is the electricity that comes off of them. And Solar City pays very competitive interest to the bondholders who supply the capital for the installations. 5% on ten-year bonds. It's a sweet deal for the investors, for the homeowners, and for the stockholders. The only loser is the electric utility that can no longer sell coal-fired electricity at high prices to the homeowners.

But plug-in hybrids, while they have their uses, are more complex and heavier, with more to go wrong (the Chevy Volt has TWO clutches). I think a better solution is pure EV. Right now you can drive a Tesla all the way across the U.S., just stopping for a half an hour about every four hours.
Title: Re: Self-Driving Cars
Post by: Guillermo on March 28, 2016, 10:00:27 AM
Found this interesting article. I like the last position talking about the ethical programming and how really there should be no problem with it.

Top misconceptions of autonomous cars and self-driving vehicles (http://www.driverless-future.com/?page_id=774)

Quote
Self-driving cars are a rapidly evolving technology which only a few years ago was still considered science fiction. In such a dynamic context, quick intuitions can be very misleading and misconceptions about the technology, its impact, and the nature of the innovation process abound. In the following we address some of the most widely held misconceptions about autonomous vehicles:
  • Driver assistance systems will evolve gradually into fully autonomous cars
  • The first models of fully autonomous cars will be targeted to the consumer and will be available for purchase
  • It will take decades until most of the vehicles on the road are capable of autonomous driving
  • Self-driving cars are controlled by classical computer algorithms (if-then rules)
  • Public demonstrations of self-driving cars provide an indication of their capabilities
  • Self-driving cars need to make the right ethical judgements
Title: Re: Self-Driving Cars
Post by: PANTS! on March 28, 2016, 10:19:23 AM
Fantastic article.  Thanks, G
Title: Re: Self-Driving Cars
Post by: brilligtove on March 28, 2016, 10:36:02 AM
Found this interesting article. I like the last position talking about the ethical programming and how really there should be no problem with it.

Top misconceptions of autonomous cars and self-driving vehicles (http://www.driverless-future.com/?page_id=774)

Quote
Self-driving cars are a rapidly evolving technology which only a few years ago was still considered science fiction. In such a dynamic context, quick intuitions can be very misleading and misconceptions about the technology, its impact, and the nature of the innovation process abound. In the following we address some of the most widely held misconceptions about autonomous vehicles:
  • Driver assistance systems will evolve gradually into fully autonomous cars
  • The first models of fully autonomous cars will be targeted to the consumer and will be available for purchase
  • It will take decades until most of the vehicles on the road are capable of autonomous driving
  • Self-driving cars are controlled by classical computer algorithms (if-then rules)
  • Public demonstrations of self-driving cars provide an indication of their capabilities
  • Self-driving cars need to make the right ethical judgements

Interesting article.

On the first point, I think the author has misconstrued what current technology can do and what it should be upgradable to do.

Quote
To summarize: Driver assistance systems can not evolve continuous driving capability gradually! At the moment we entrust them to drive continuously they require a huge, discontinuous jump in capability which will place their capabilities very close to the capabilities of fully autonomous vehicles.

The 'large discontinuous jump in capability' may not be so large. For example, consider a fleet of vehicles with driver assist capabilities but which are otherwise fully under the control of a human. Assume that the sensor technology and computer hardware are fully capable of driving autonomously, but we don't have the software yet. Over the course of a tens of millions Km under the control of real human drivers, the machine learning systems will have a vast wealth of information about when the SDV systems disagreed with the human driver (as per point #4). In cases where humans screw up, the SDV intelligence will learn what not to do; in cases where humans outperform the SDV it will learn what to do. As different aspects of the SDV performance improve, the whole system can be upgraded to take advantage of the learning. It might take a year for SDVs to be competent enough to let you sleep on the highway in good weather. It might be several more before SDVs are good enough to let you sleep while in city streets. These do represent gradual improvements, and not discontinuous improvements.

Points 4 and 6 seem particularly well thought out to me, though the writing! is a bit! breathless! in places. ;)
Title: Re: Self-Driving Cars
Post by: Jeremy's Sea on March 28, 2016, 03:00:20 PM
Our main driver is a 2005 Honda Element with 5-speed, bought new and coming up on 180,000 miles.

Its been quite reliable, and we plan on driving it for as long as possible - until the wheels fall off, as we put it.

We're really hoping by then plug-in hybrids continue to improve until we can get decent electric range at a reasonable price. We live in the country, and most of our driving is trips into town or to towns maybe 20 to 40 miles away. So 200 mile range would be in a sweet spot, and might work even for a pure electric. Where a pure electric would not work at that range would be our semi-regular trips to Knoxville and back - about 100 miles each way.

We also bought 60 shares of Tesla stock at various prices, now up just a bit overall - so we're especially interested in any Tesla under $40k to be our next car - if the little Element can keep its wheels on that long!
I'm so glad to see this! I have a 2004 Element with 132K on it and have been debating on how long I should continue to drive it. Most people assure me they're solid mechanically over 200K and so far I have had very little go wrong, and nothing engine related. So I have trying to hold out for an all electric. I went solar this year and am waiting to see if I generate a surplus, since LADWP doesn't buy back power and only credits it against usage, I would have a strong incentive to power a car from my solar production. A friend just bought a Volt because he was over producing and basically giving free energy to the power company. Exciting times!
Title: Re: Self-Driving Cars
Post by: teethering on March 28, 2016, 04:45:40 PM
No teethering, actually there is no problem at all with my assertions, you are the one who has problems with it, not me.
 

This confuses me... do you not know how arguments in general work?  When disagreements between two reasonable people happen, it's typical for one person to have problems with the other person's position.  That's actually the nutshell of the whole thing.  Yes, I have a problem with your assertions.

Now in the case of reasonable people when the disagreement is about factual claims the only way to resolve it is to bring forth objective evidence to support these factual claims.  You made factual claims in the form of "there are studies that show self-driven cars will slow down traffic".  You haven't provided evidence for it.  We therefore have no way to resolve the disagreement.

Now in my experience when someone so obviously completely full of shit is talking out of their ass and when called on it refuse to provide evidence it's because they were lying, there is no evidence.  Occasionally they're just being stubborn and/or lazy.  But either way, it's bad form and it's not anyone's problem but the person who is talking out of their ass.

Quote
I don't need to convince you of anything.  You are a skeptic, so you already have your believes set in stone.

In this case my opinion of the topic has nothing to do with skepticism, it has a lot to do with understanding how computer systems work, in particular how real-time systems work (which self-driven cars happens to be a subset of).  Believe me when I tell you that I had dealt with a lot of problems that arise from programming computers to react to contingencies in linear or constant time.  It is hard to make them work well when you're dealing with controlling mechanical systems, but there's no reason to believe there is some insurmountable impediment.  The key to understanding this problem is in comparing the scope of the computational tasks and the power of the hardware that you can throw at it, because you don't have any problems in this path that can't be dealt with by O(n) or O(1) heuristics in a reasonable way.  The other key is this word "reasonable" and to me the bar is clearly set at "as well or better than human behaviour".

The other piece of evidence that does have to do with skepticism is that we appear to have fairly promising data from real world trials like Google's.  This further confirms my opinion on this matter.

Nothing you've presented that I've read raises real concerns.

Quote
  That is the paradox of the way the word is used by you folks. So I have no desire to convince you of anything.  It is for others, who actually have an open mind and wish to learn things.
 

Are you under the impression that someone could possibly learn something from you?  You're even more confused than I thought.

Quote
Everyone here believed that as soon as you have self-driving cars, traffic will be faster, with less congestion.  This study shows just the opposite.

What study?  The study I found for you to try and give a good faith hearing to your claim?

If that's the case, you're wrong.  The study didn't look at the full picture, it looked at a particular, narrow slice of the traffic and their claims aren't about the whole picture.  They are pointing out possible problems with a specific way of implementing self-driving cars in specific portions of the grid.  You are bait-and-switching it to apply to the whole thing, whatever the implementation.

Quote
  I don't care about the spin you want to try to throw at it.  That is all skeptics do, try to throw spin.  Its why the SGU never allows opposing viewpoints on their show-they like to control their spin propaganda messages. 

So now people can read it and see for themselves.  I am well versed in your sides tactics.  They don't interest me in the slightest.  Its not a problem for me at all.

This is particularly rich coming from someone who refuses to cite their sources.  Citing sources isn't spin or a tactic, lying about what's in the studies is.  Citing sources is intellectual honesty and transparency, something you apparently are completely lacking.
Title: Re: Self-Driving Cars
Post by: daniel1948 on March 28, 2016, 04:58:23 PM
No teethering, actually there is no problem at all with my assertions, you are the one who has problems with it, not me.

This confuses me... do you not know how arguments in general work?

Phooey does not know how arguments work, as he demonstrates over and over again in his posts. He does not care how arguments work, as demonstrated by his style of argument. And he does not know what the word "skeptic" means, as demonstrated by his incorrect definition of the word numerous times in this thread. At some point, one must recognize that one is dealing with a troll. He's having fun throwing nonsense at us.
Title: Re: Self-Driving Cars
Post by: teethering on March 28, 2016, 06:05:37 PM
No teethering, actually there is no problem at all with my assertions, you are the one who has problems with it, not me.

This confuses me... do you not know how arguments in general work?

Phooey does not know how arguments work, as he demonstrates over and over again in his posts. He does not care how arguments work, as demonstrated by his style of argument. And he does not know what the word "skeptic" means, as demonstrated by his incorrect definition of the word numerous times in this thread. At some point, one must recognize that one is dealing with a troll. He's having fun throwing nonsense at us.

In my defense, it's my first time encountering him.  Looking back at his previous posts I do see a familiar pattern.
Title: Re: Self-Driving Cars
Post by: brilligtove on March 28, 2016, 07:35:33 PM
No teethering, actually there is no problem at all with my assertions, you are the one who has problems with it, not me.

This confuses me... do you not know how arguments in general work?

Phooey does not know how arguments work, as he demonstrates over and over again in his posts. He does not care how arguments work, as demonstrated by his style of argument. And he does not know what the word "skeptic" means, as demonstrated by his incorrect definition of the word numerous times in this thread. At some point, one must recognize that one is dealing with a troll. He's having fun throwing nonsense at us.

In my defense, it's my first time encountering him.  Looking back at his previous posts I do see a familiar pattern.

On the positive side, we have determined with this thread that it is possible to us shame and ridicule to provoke a shift from Troll to Sealion. Also, I expect this post will cause an immediate regression to Troll. The Phoobot doesn't really understand any ot the stuff that is happening on these forums though, so 'm not worried about it.
Title: Re: Self-Driving Cars
Post by: daniel1948 on March 28, 2016, 07:43:52 PM
Sealion?
Title: Re: Self-Driving Cars
Post by: arthwollipot on March 28, 2016, 07:56:49 PM
Sea-Lioning (http://knowyourmeme.com/memes/sea-lioning)
Title: Re: Self-Driving Cars
Post by: brilligtove on March 28, 2016, 08:30:31 PM
Sea-Lioning (http://knowyourmeme.com/memes/sea-lioning)

I have been reading that comic for YEARS and was thrilled when one of his very insightful ideas became a meme.
Title: Re: Self-Driving Cars
Post by: daniel1948 on March 28, 2016, 09:24:34 PM
Sea-Lioning (http://knowyourmeme.com/memes/sea-lioning)

Thanks. I had found this, and it didn't seem to fit the context:

http://www.urbandictionary.com/define.php?term=sea+lion

Quote
sea lion

A lesbian cougar. An older woman that preys on younger women at bars and other prime locations in much the same way that cougars prey on younger men. Sea lion is derived from cougar as an aquatic homage to the large cat, namely because their diet consists of fish.

Did you see that sea lion that hit on Alexa? She was old enough to be her mom!
Title: Re: Self-Driving Cars
Post by: brilligtove on March 28, 2016, 09:37:20 PM
Sea-Lioning (http://knowyourmeme.com/memes/sea-lioning)

Thanks. I had found this, and it didn't seem to fit the context:

http://www.urbandictionary.com/define.php?term=sea+lion

Quote
sea lion

A lesbian cougar. An older woman that preys on younger women at bars and other prime locations in much the same way that cougars prey on younger men. Sea lion is derived from cougar as an aquatic homage to the large cat, namely because their diet consists of fish.

Did you see that sea lion that hit on Alexa? She was old enough to be her mom!

That is a mew one for me.
Title: Re: Self-Driving Cars
Post by: brilligtove on March 28, 2016, 09:38:44 PM
Sea-Lioning (http://knowyourmeme.com/memes/sea-lioning)

Thanks. I had found this, and it didn't seem to fit the context:

http://www.urbandictionary.com/define.php?term=sea+lion

Quote
sea lion

A lesbian cougar. An older woman that preys on younger women at bars and other prime locations in much the same way that cougars prey on younger men. Sea lion is derived from cougar as an aquatic homage to the large cat, namely because their diet consists of fish.

Did you see that sea lion that hit on Alexa? She was old enough to be her mom!

That is a mew one for me.


I swear that was an honest typo. I meant "new"!
Title: Re: Self-Driving Cars
Post by: DG on March 28, 2016, 10:36:50 PM
Sea-Lioning (http://knowyourmeme.com/memes/sea-lioning)

Thanks. I had found this, and it didn't seem to fit the context:

http://www.urbandictionary.com/define.php?term=sea+lion

Quote
sea lion

A lesbian cougar. An older woman that preys on younger women at bars and other prime locations in much the same way that cougars prey on younger men. Sea lion is derived from cougar as an aquatic homage to the large cat, namely because their diet consists of fish.

Did you see that sea lion that hit on Alexa? She was old enough to be her mom!

That is a mew one for me.

I'm fairly sure that I am not allowed to mod box for bad puns. If I was....
Title: Re: Self-Driving Cars
Post by: SkeptiQueer on March 28, 2016, 11:37:10 PM
Hey, just wanted to make sure you got that list of episodes featuring dissenting views that completely contradicts your narrative of SGU as an idealogical echochamber.

Thanks!
Title: Re: Self-Driving Cars
Post by: phooey on March 28, 2016, 11:52:13 PM
Teethering,

Let me make sure I understand your whining, because I am having a hard time seeing the screen clearly through my tears of laughter.  You are questioning my contention that there are studies that show self-driving cars could make traffic worse, yes?  And you are questioning my contention that there are such studies, whilst you have an article right in front of you which is titled, wait for it..... "HOW DRIVERLESS CARS COULD MAKE TRAFFIC DRAMATICALLY WORSE"!!

And I am supposed to take your criticism seriously?  Are you trying to spoof Skeptiqueer?

So in the future, when you say stupid shit like  "The other piece of evidence that does have to do with skepticism is that we appear to have fairly promising data from real world trials like Google's.", do I really need to point out to you that we have ZERO data from Google about how this will affect real world traffic, or won't it just be much easier for me to say, why are you such a troll?  Do you understand how arguments work?  Do you understand what evidence means?  Do you understand the simplest fucking point?     
Title: Re: Self-Driving Cars
Post by: phooey on March 28, 2016, 11:53:49 PM
Hey, just wanted to make sure you got that list of episodes featuring dissenting views that completely contradicts your narrative of SGU as an idealogical echochamber.

Thanks!

I must have missed that list.  Please show me. 
Title: Re: Self-Driving Cars
Post by: SkeptiQueer on March 29, 2016, 12:06:17 AM


Hey, just wanted to make sure you got that list of episodes featuring dissenting views that completely contradicts your narrative of SGU as an idealogical echochamber.

Thanks!

I must have missed that list.  Please show me.



SGU episodes featuring guests who do not agree with the panel include, but are not limited to Neal Adams (Episode 51), B. Alan Wallace (Ep 73), Jimmy Carter (Ep 105), Michio Kaku (Ep 182), Billy West (Ep 373), Don McElroy (Ep 408).

Further, as SGU is a panel discussion show about science, it's likely not the best place to gear someone talking about a view that doesn't stand up to scientific scrutiny. If you want to hear that, I suggest "Be Reasonable" hosted by Michael Marshal of the Merseyside Sceptics, apologies for misspellings if any, where a prominent sceptic invites on promoters of what would normally be considered woo for a patient discussion. Marsh, if I can call him that, is likely the most patient man in the world given the number of ridiculous people he has calmly interviewed while being screamed at.

So you see, an actual skeptical discussion where you made a claim, and I disputed it with a counterclaim and provided evidence has just occurred, and I've given plenty of evidence to contradict your claims that skeptics just blindly accept whatever.
 
Will you accept that you were wrong in your broad character assassination and incorrect assessment of the SGU?

Three posts above the last time you asked.
Title: Re: Self-Driving Cars
Post by: teethering on March 29, 2016, 01:38:08 AM
So it won't be a problem in the future for me to call people lying out of their asshole trolls, or to say, "Oh, teethering is sooooo cute when he tries to pretend to be an intelligence agent!"

Because previously, I just politely ignored the vacuous non sequiturs or ad hominem bullshit posts.  I see now, that I can reply to them in kind though, with your blessing.   Thanks for that.

You are doing one hell of a great job as a moderator by the way.  Do I have permission to share your private messages to me to show just how great a moderator you are?   Because you are sooooo cute when you try to pretend to be an intelligence agent.

FTFY

 :ninja:
Title: Re: Self-Driving Cars
Post by: teethering on March 29, 2016, 01:44:32 AM
Teethering,

Let me make sure I understand your whining, because I am having a hard time seeing the screen clearly through my tears of laughter.  You are questioning my contention that there are studies that show self-driving cars could make traffic worse, yes?  And you are questioning my contention that there are such studies, whilst you have an article right in front of you which is titled, wait for it..... "HOW DRIVERLESS CARS COULD MAKE TRAFFIC DRAMATICALLY WORSE"!!

And I am supposed to take your criticism seriously?  Are you trying to spoof Skeptiqueer?

So in the future, when you say stupid shit like  "The other piece of evidence that does have to do with skepticism is that we appear to have fairly promising data from real world trials like Google's.", do I really need to point out to you that we have ZERO data from Google about how this will affect real world traffic, or won't it just be much easier for me to say, why are you such a troll?  Do you understand how arguments work?  Do you understand what evidence means?  Do you understand the simplest fucking point?   

To be clear, I explained how your general claim that there are studies that show that in general self-driving cars would make traffic problems worse are not substantiated by the article I linked to, because the study it describes is far narrower than your general claim.  I explained this twice already, this is the third time, but apparently you haven't yet finished your hookt on fonix, so reading is difficult for you.  That's ok, I'll try to use smaller words.

But basically this is what I suspected about your bullshit in the first place, you are deliberately distorting claims made by actual researchers, which is why you can't actually link to any of the studies.  That's how I know you're lying.
Title: Re: Self-Driving Cars
Post by: Fast Eddie B on March 29, 2016, 07:56:22 AM
I remember a friend's girlfriend worked as a grocery checker when bar code scanners first came out.

She thought there was no way they could be as fast and accurate as a human.

Similarly, think of other forms of "traffic"...

Would the packets that take info to and from your computer move more rapidly if routed by humans?

And do we all remember how much faster phone calls connected back in the day?

(https://upload.wikimedia.org/wikipedia/commons/8/8e/Photograph_of_Women_Working_at_a_Bell_System_Telephone_Switchboard_(3660047829).jpg)

No, we don't, because automation virtually always makes thing faster in the end.

I find zero reason to believe the same thing will not happen with automated cars.
Title: Re: Self-Driving Cars
Post by: brilligtove on March 29, 2016, 07:59:52 AM
My mom was an operator once, way back in the day.
Title: Re: Self-Driving Cars
Post by: Morvis13 on March 29, 2016, 08:14:24 AM
My mom was an operator once, way back in the day.

That's what she said.
Title: Re: Self-Driving Cars
Post by: brilligtove on March 29, 2016, 08:15:36 AM
My mom was an operator once, way back in the day.

That's what she said.

Yes. She told me she was good at it too.
Title: Re: Self-Driving Cars
Post by: phooey on March 29, 2016, 08:28:13 AM
Teethering,

The title of the article is "HOW DRIVERLESS CARS COULD MAKE TRAFFIC DRAMATICALLY WORSE"!!...for fuck's sake!

But no, no, you are right, there are no studies which say driverless cars could make traffic worse, I just made that up. 

Holy shit, are you insane?
Title: Re: Self-Driving Cars
Post by: daniel1948 on March 29, 2016, 08:50:28 AM
Teethering,

The title of the article is "HOW DRIVERLESS CARS COULD MAKE TRAFFIC DRAMATICALLY WORSE"!!...for fuck's sake!

But no, no, you are right, there are no studies which say driverless cars could make traffic worse, I just made that up. 

Holy shit, are you insane?

Something you would know if you ever actually listened to the SGU podcast is that headlines are written by editors whose job is to write headlines that will get people to read or click on the article. They go for the dramatic and as a rule they mischaracterize or even completely reverse what the article says. In this case, the study did not address the overall effect on traffic, but only looked at a single hypothetical intersection, and only looked at a situation in which computer-driven cars would be mixed in with human-driven cars, and would accelerate slowly enough that the occupants would feel as though they were on a train. The word "could" in the title should have clued you in, because "could" means the headline writer is pulling crap out of his hat.

The article assumed that the computer-driven cars would not accelerate with the flow of traffic, and did not consider the condition of computer-driven cars all accelerating in synchronization, which would actually speed up the flow of traffic.

But of course all this has been mentioned above and you choose to ignore facts that do not suit your narrative and cherry-pick those that can be shoe-horned into your "self-driving cars are bad" dogma.
Title: Re: Self-Driving Cars
Post by: Morvis13 on March 29, 2016, 08:55:18 AM
Yep, traffic is dramatically worse by 10 whole seconds at a 4 way stop IF the car is designed to slow down like a train. Fact. You got us there. Better strap the whole project.

OR we could do the whole experiment again with 100% robot drivers and remove the stop sign. Or even let the robots flow with traffic as Daniel suggests.
Title: Re: Self-Driving Cars
Post by: phooey on March 29, 2016, 09:23:00 AM
Whrrrr, whrrr, whrrr, watch the skeptics spin. 

The article is about a study which shows that driverless cars could make traffic dramatically worse. 

Spin all you want, that is what the study is about.  Let someone else follow you down the rabbit hole of denial.  I don't need to accept your warped interpretation that the study doesn't mean what it says it means.  I can read just fine.
Title: Re: Self-Driving Cars
Post by: brilligtove on March 29, 2016, 09:49:02 AM
Folks, the Phoobot might be able to read, but it shows no comprehension. Can we agree to ignore it and continue the interesting conversation instead?

I remember a friend's girlfriend worked as a grocery checker when bar code scanners first came out.

She thought there was no way they could be as fast and accurate as a human.

Similarly, think of other forms of "traffic"...

Would the packets that take info to and from your computer move more rapidly if routed by humans?

And do we all remember how much faster phone calls connected back in the day?

(https://upload.wikimedia.org/wikipedia/commons/8/8e/Photograph_of_Women_Working_at_a_Bell_System_Telephone_Switchboard_(3660047829).jpg)

No, we don't, because automation virtually always makes thing faster in the end.

I find zero reason to believe the same thing will not happen with automated cars.

There could be new kinds of congestion with automated car routing, though. Internet packets collide all the time and are dropped and resent. We can't do the same with cars - at least not outside a rally race. :) I suspect that SDVs will still need to stop at intersections to allow pedestrian traffic and human powered vehicles (like bikes) to operate. Maybe we could get to more dedicated non-car lanes if cohesion is lessened. More one way streets might make grid-layout cities easier to route as well.
Title: Re: Self-Driving Cars
Post by: phooey on March 29, 2016, 10:00:52 AM
Right, right, the study doesn't say that driverless cars could make traffic dramatically worse.  They are lying. 

The skeptics have got it all sussed out.  Just more lies from the media. I am just repeating the media lies....

Its a conspiracy!   

Title: Re: Self-Driving Cars
Post by: SkeptiQueer on March 29, 2016, 10:15:27 AM
Right, right, the study doesn't say that driverless cars could make traffic dramatically worse.  They are lying. 

The skeptics have got it all sussed out.  Just more lies from the media. I am just repeating the media lies....

Its a conspiracy!
Still not going to acknowledge that your narrative about SGU and the skeptics is incorrect?
Title: Re: Self-Driving Cars
Post by: PANTS! on March 29, 2016, 10:15:32 AM
Folks, the Phoobot might be able to read, but it shows no comprehension. Can we agree to ignore it and continue the interesting conversation instead?

I remember a friend's girlfriend worked as a grocery checker when bar code scanners first came out.

She thought there was no way they could be as fast and accurate as a human.

Similarly, think of other forms of "traffic"...

Would the packets that take info to and from your computer move more rapidly if routed by humans?

And do we all remember how much faster phone calls connected back in the day?

(https://upload.wikimedia.org/wikipedia/commons/8/8e/Photograph_of_Women_Working_at_a_Bell_System_Telephone_Switchboard_(3660047829).jpg)

No, we don't, because automation virtually always makes thing faster in the end.

I find zero reason to believe the same thing will not happen with automated cars.

There could be new kinds of congestion with automated car routing, though. Internet packets collide all the time and are dropped and resent. We can't do the same with cars - at least not outside a rally race. :) I suspect that SDVs will still need to stop at intersections to allow pedestrian traffic and human powered vehicles (like bikes) to operate. Maybe we could get to more dedicated non-car lanes if cohesion is lessened. More one way streets might make grid-layout cities easier to route as well.

Why not connect bikes and pedestrians to the network?
Title: Re: Self-Driving Cars
Post by: Guillermo on March 29, 2016, 10:28:52 AM
Whrrrr, whrrr, whrrr, watch the skeptics spin. 

The article is about a study which shows that driverless cars could make traffic dramatically worse. 

Spin all you want, that is what the study is about.  Let someone else follow you down the rabbit hole of denial.  I don't need to accept your warped interpretation that the study doesn't mean what it says it means.  I can read just fine.
Let me see if I read the article correctly. Please confirm.

Does the study say that driverless cars could make traffic worse in a 4 way intersection if 25% of the vehicles are driverless and they optimize comfort by stopping like Light rail trains or high speed rail trains do?
Title: Re: Self-Driving Cars
Post by: SkeptiQueer on March 29, 2016, 10:31:55 AM
Folks, the Phoobot might be able to read, but it shows no comprehension. Can we agree to ignore it and continue the interesting conversation instead?

I remember a friend's girlfriend worked as a grocery checker when bar code scanners first came out.

She thought there was no way they could be as fast and accurate as a human.

Similarly, think of other forms of "traffic"...

Would the packets that take info to and from your computer move more rapidly if routed by humans?

And do we all remember how much faster phone calls connected back in the day?

(https://upload.wikimedia.org/wikipedia/commons/8/8e/Photograph_of_Women_Working_at_a_Bell_System_Telephone_Switchboard_(3660047829).jpg)

No, we don't, because automation virtually always makes thing faster in the end.

I find zero reason to believe the same thing will not happen with automated cars.

There could be new kinds of congestion with automated car routing, though. Internet packets collide all the time and are dropped and resent. We can't do the same with cars - at least not outside a rally race. :) I suspect that SDVs will still need to stop at intersections to allow pedestrian traffic and human powered vehicles (like bikes) to operate. Maybe we could get to more dedicated non-car lanes if cohesion is lessened. More one way streets might make grid-layout cities easier to route as well.

Why not connect bikes and pedestrians to the network?

I feel like there's a big disincentive there on the part of the user. Privacy concerns. I don't believe enough eople would use it, especially cyclists and lower-income pedestrians, to make it reliable data for the cars. Sensors at crosswalks, sure.

For those wondering about the viability of connecting a grid and vehicles with good ID, remember that we've done this with airplanes the hard way for most of a century, and now IFF systems like the BlueForce Tracker allow for real-time HUD identification of friendly forces.
Title: Re: Self-Driving Cars
Post by: brilligtove on March 29, 2016, 10:37:06 AM
Folks, the Phoobot might be able to read, but it shows no comprehension. Can we agree to ignore it and continue the interesting conversation instead?

I remember a friend's girlfriend worked as a grocery checker when bar code scanners first came out.

She thought there was no way they could be as fast and accurate as a human.

Similarly, think of other forms of "traffic"...

Would the packets that take info to and from your computer move more rapidly if routed by humans?

And do we all remember how much faster phone calls connected back in the day?

(https://upload.wikimedia.org/wikipedia/commons/8/8e/Photograph_of_Women_Working_at_a_Bell_System_Telephone_Switchboard_(3660047829).jpg)

No, we don't, because automation virtually always makes thing faster in the end.

I find zero reason to believe the same thing will not happen with automated cars.

There could be new kinds of congestion with automated car routing, though. Internet packets collide all the time and are dropped and resent. We can't do the same with cars - at least not outside a rally race. :) I suspect that SDVs will still need to stop at intersections to allow pedestrian traffic and human powered vehicles (like bikes) to operate. Maybe we could get to more dedicated non-car lanes if cohesion is lessened. More one way streets might make grid-layout cities easier to route as well.

Why not connect bikes and pedestrians to the network?

I feel like there's a big disincentive there on the part of the user. Privacy concerns. I don't believe enough eople would use it, especially cyclists and lower-income pedestrians, to make it reliable data for the cars. Sensors at crosswalks, sure.

For those wondering about the viability of connecting a grid and vehicles with good ID, remember that we've done this with airplanes the hard way for most of a century, and now IFF systems like the BlueForce Tracker allow for real-time HUD identification of friendly forces.
LI had the same thought. Most of us already do walk with phones, but there is a presumption that the location information that the phone company gets is not used (by the government) to track our movements. And volunteering to share my location data is not the same as being required to provide it.
Title: Re: Self-Driving Cars
Post by: SkeptiQueer on March 29, 2016, 10:39:30 AM
Folks, the Phoobot might be able to read, but it shows no comprehension. Can we agree to ignore it and continue the interesting conversation instead?

I remember a friend's girlfriend worked as a grocery checker when bar code scanners first came out.

She thought there was no way they could be as fast and accurate as a human.

Similarly, think of other forms of "traffic"...

Would the packets that take info to and from your computer move more rapidly if routed by humans?

And do we all remember how much faster phone calls connected back in the day?

(https://upload.wikimedia.org/wikipedia/commons/8/8e/Photograph_of_Women_Working_at_a_Bell_System_Telephone_Switchboard_(3660047829).jpg)

No, we don't, because automation virtually always makes thing faster in the end.

I find zero reason to believe the same thing will not happen with automated cars.

There could be new kinds of congestion with automated car routing, though. Internet packets collide all the time and are dropped and resent. We can't do the same with cars - at least not outside a rally race. :) I suspect that SDVs will still need to stop at intersections to allow pedestrian traffic and human powered vehicles (like bikes) to operate. Maybe we could get to more dedicated non-car lanes if cohesion is lessened. More one way streets might make grid-layout cities easier to route as well.

Why not connect bikes and pedestrians to the network?

I feel like there's a big disincentive there on the part of the user. Privacy concerns. I don't believe enough people would use it, especially cyclists and lower-income pedestrians, to make it reliable data for the cars. Sensors at crosswalks, sure.

For those wondering about the viability of connecting a grid and vehicles with good ID, remember that we've done this with airplanes the hard way for most of a century, and now IFF systems like the BlueForce Tracker allow for real-time HUD identification of friendly forces.
LI had the same thought. Most of us already do walk with phones, but there is a presumption that the location information that the phone company gets is not used (by the government) to track our movements. And volunteering to share my location data is not the same as being required to provide it.

PFFFT. My GPS goes on when I need directions and goes off as soon as I no longer do. I'm not even a fan of the E911, but it's a necessary evil. It seems like the networking could just as easily happen in the grid level instead of tracking the individuals.
Title: Re: Self-Driving Cars
Post by: phooey on March 29, 2016, 10:45:20 AM
Skeptiqueer,

Your post is an acknowledgement that the SGu has been an echo-chamber for at least the past two years (the last time they supposedly had an opposing viewpoint expressed). 

Furthermore you claim Michio Kaku and Jimmy Carter are opposing viewpoints.  That's laughable. 

They had a kook on one top who believes the earth is hollow.  Wow. 

How about a real opposing viewpoint.  How about a biologist who doubts Darwinian evolution, like Simon Conway Morris? How about Dr. William Thompson, the man who exposed the fraud at the CDC over vaccine study cover-ups?  How about Lennart Bengtsson or Freeman Dyson to talk about climate change?  How about David Suzuki to talk about GMO's?  How about Rupert Sheldrake to talk about telepathy?  Or Jeffery Long or Sam Parnia to talk about near death experience?   Or Dr. Julia Mossbridge to talk about psychic abilities? 

Give me a break.  Would you be serious.   
Title: Re: Self-Driving Cars
Post by: Johnny Slick on March 29, 2016, 10:46:19 AM
Right, right, the study doesn't say that driverless cars could make traffic dramatically worse.  They are lying. 

The skeptics have got it all sussed out.  Just more lies from the media. I am just repeating the media lies....

Its a conspiracy!
CAMON LIBS YOU CANT EXPECT ME TO READ ARTICLES I DIDN'T CITE BUT AM TAKING CREDIT FOR CITING WHAT DO YOU THINK I AM SOME KIND OF SKEPTIC??
Title: Re: Self-Driving Cars
Post by: Johnny Slick on March 29, 2016, 10:49:09 AM
Skeptiqueer,

Your post is an acknowledgement that the SGu has been an echo-chamber for at least the past two years (the last time they supposedly had an opposing viewpoint expressed). 

Furthermore you claim Michio Kaku and Jimmy Carter are opposing viewpoints.  That's laughable. 

They had a kook on one top who believes the earth is hollow.  Wow. 

How about a real opposing viewpoint.  How about a biologist who doubts Darwinian evolution, like Simon Conway Morris? How about Dr. William Thompson, the man who exposed the fraud at the CDC over vaccine study cover-ups?  How about Lennart Bengtsson or Freeman Dyson to talk about climate change?  How about David Suzuki to talk about GMO's?  How about Rupert Sheldrake to talk about telepathy?  Or Jeffery Long or Sam Parnia to talk about near death experience?   Or Dr. Julia Mossbridge to talk about psychic abilities? 

Give me a break.  Would you be serious.
Have you listened to any of these episodes? The flat earth guy was especially excruciating. The problem with these people is that, like a certain person posting in this thread who shall remain nameless except to say that it is you, they don't argue honestly. They lie outright and when they're not lying outright, they refuse to acknowledge claims. And then when they cite evidence they refuse to actually delve into what the evidence says and instead cherry pick out the bits and pieces that support their case the best.

(and by the way, they did have Bill Nye on to talk about GMOs. Dr. Novella disagrees with much of what Nye says about them, and they spoke about those differences on the show)
Title: Re: Self-Driving Cars
Post by: phooey on March 29, 2016, 10:52:26 AM
You mean like this:

"The final traffic tolls ranged from annoying to frightening. In the baseline situation, without any driverless cars, each vehicle experienced a delay of 20 seconds at the intersection. When driverless cars accelerated and decelerated in the style of light rail, the congestion worsened from 4 percent (21 seconds) to 50 percent (30 seconds). The number of cars traveling through the intersection—at 1,793 in the baseline scenario—also fell between 4 percent (1,724 cars) and 21 percent (1,415 cars).  Consider, for instance, that these simulations didn't include pedestrians. Doing so no doubt would have led to even more starting and stopping, and thus more delay."


Yea right pal. 

Great, two years ago the SGU had on a hollow Earth cartoonist. 

Yea right pal. 

The HSR-smoothness scenario was even scarier. Against the same baseline, autonomous cars that started and stopped like high-speed rail increased delay anywhere from 36 percent (27 seconds) to nearly 2,000 percent (6 minutes and 44 seconds!). Meanwhile, intersection capacity fell between 18 percent (1,469 cars) and 53 percent (850 cars)."
Title: Re: Self-Driving Cars
Post by: SkeptiQueer on March 29, 2016, 11:00:54 AM
Skeptiqueer,

Your post is an acknowledgement that the SGu has been an echo-chamber for at least the past two years (the last time they supposedly had an opposing viewpoint expressed). 

Furthermore you claim Michio Kaku and Jimmy Carter are opposing viewpoints.  That's laughable. 

They had a kook on one top who believes the earth is hollow.  Wow. 

How about a real opposing viewpoint.  How about a biologist who doubts Darwinian evolution, like Simon Conway Morris? How about Dr. William Thompson, the man who exposed the fraud at the CDC over vaccine study cover-ups?  How about Lennart Bengtsson or Freeman Dyson to talk about climate change?  How about David Suzuki to talk about GMO's?  How about Rupert Sheldrake to talk about telepathy?  Or Jeffery Long or Sam Parnia to talk about near death experience?   Or Dr. Julia Mossbridge to talk about psychic abilities? 

Give me a break.  Would you be serious.

You didn't listen to the episodes, so you don't know that Michio Kaku dissents from the SGU on matters of metaphysics, and that Carter is a UFO believer. Further, as I said the list I gave was not a complete list. Most recently, Joe Rogan [conspiracy theorist, UFO believer, etc] was invited on about two weeks ago but refused, having already been soundly spanked by Phil Plait, another frequent flyer. Others have been invited but they frequently refuse the invite.

As I noted, the SGU is a panel discussion show, and interviews have been less common as the show progressed. It's purpose isn't to interview every crackpot who thinks they've discovered dog telepathy. As I noted, there are other podcasts for that. You shouldn't expect Anderson Cooper to be embedding with narcoterrorists in Colombia either, as that's not the format his show takes.

I am being serious, but it's difficult to be serious when you claim we're all an echochamber despite dissent being expressed in this thread, among others, and making claims about a show you haven't listened to.

I still don't understand why you're here if you don't intend to prove anything and don't care about changing minds. Do you just like berating people who think differently from you? 'Cuz that's a shitty way to behave, if you're not going to engage in honest discussion.

You mean like this:

"The final traffic tolls ranged from annoying to frightening. In the baseline situation, without any driverless cars, each vehicle experienced a delay of 20 seconds at the intersection. When driverless cars accelerated and decelerated in the style of light rail, the congestion worsened from 4 percent (21 seconds) to 50 percent (30 seconds). The number of cars traveling through the intersection—at 1,793 in the baseline scenario—also fell between 4 percent (1,724 cars) and 21 percent (1,415 cars).  Consider, for instance, that these simulations didn't include pedestrians. Doing so no doubt would have led to even more starting and stopping, and thus more delay."


Yea right pal. 

Great, two years ago the SGU had on a hollow Earth cartoonist. 

Yea right pal. 

The HSR-smoothness scenario was even scarier. Against the same baseline, autonomous cars that started and stopped like high-speed rail increased delay anywhere from 36 percent (27 seconds) to nearly 2,000 percent (6 minutes and 44 seconds!). Meanwhile, intersection capacity fell between 18 percent (1,469 cars) and 53 percent (850 cars)."

Again, cars don't normally start and stop like trains, so it makes no sense that self-driving cars would do so. That's what's called a faulty premise. An argument built on a faulty premise, like a house with a bad foundation, doesn't stand very well.
Title: Re: Self-Driving Cars
Post by: brilligtove on March 29, 2016, 11:23:06 AM
SQ, Phoobot doesn't understand the context of the research, and what it means in that context. We can ignore it, really we can. :)

Slick, IIRC, Nye has changed his position on GMOs since then, after being presented with new (to him) evidence. He still has ecosystem-level concerns, I think, but the actual genetic engineering isn't his concern anymore. Has Steve mentioned that? I feel like he has, but don't have any clear recollection.

Back to SDVs, I think the study that shows delays at intersections is useful initial work, but too limited to have any real predictive power. And, as SQ notes, it has a premise that is questionable at best. It isn't clear that it accounts for human driver rage when caught behind a super slow acceleration SDV either. One would expect that to have an effect on how humans that SDVS - for example, rushing past them and cutting them off aggressively to get into pole position at the next intersection. Realistically, SDVs are going to have to accelerate like humans for quite some time.

Another scenario occurs to me. We have diamond lanes in Toronto, where cabs, busses, and cars with 3 or more people in them can drive. Everyone else can only use the diamond lane in the run up to turning at an intersection (they are usually the right lane on city streets).

What if that lane became the SDV lane? If that was the case, a train of SDVs could network together and accelerate and decelerate as single incompressible unit, rather than having the springy expansion and contraction of HDVs. Under those circumstances a slower acceleration could still get more cars through a green light. This could be especially effective if the SDVs could arrange to have SDVs that need to turn at the end of the block. That would mean that most pedestrians would be out of the way, and the SDVs would be able to take the corner relatively quickly.

Another thought: if the structure of vehicles no longer requires a driver (in the farther future) could the interior of the vehicle be set up to tilt into accelerations? I'm thinking about a platform with batteries and wheels and such, with an armoured passenger compartment that can tilt left and right to minimize lateral acceleration to passengers. Cars can generally handle much higher lateral forces around corners than drivers are comfortable with. A tilting cabin would allow a lot more cars to get around corners without having passengers tossed around in their seats.

...this would probably require a shift in the programming for crosswalks. Have pedestrian crossing signals that end a bit before the green light ends, so that turning cars can turn quickly.

If traffic is flowing better, we could largely eliminate left turns in urban areas as well - or maybe make them more efficient. For example, allowing left turns and perpendicular (non-intefering) right turns to happen at the same time.

Interesting routing problems.
Title: Re: Self-Driving Cars
Post by: SkeptiQueer on March 29, 2016, 11:29:18 AM
Not making left turns would likely slow some commutes, but again you're likely to see an overall decrease. IIRC, UPS routes to avoid left turns both for safety reasons and to avoid long idles at difficult intersections.
Title: Re: Self-Driving Cars
Post by: amysrevenge on March 29, 2016, 11:33:53 AM
Not making left turns would likely slow some commutes, but again you're likely to see an overall decrease. IIRC, UPS routes to avoid left turns both for safety reasons and to avoid long idles at difficult intersections.

Yeah, if you re-organize your streets/intersections such that "three rights make a left" is do-able, you could change a lot of things about pedestrians and intersection timing.
Title: Re: Self-Driving Cars
Post by: daniel1948 on March 29, 2016, 12:20:10 PM
There's a fundamental flaw in the study: It begins by pointing out that it would be nice to be able to work and sleep in your computer-driven car, like you can on a train. It then makes a leap to "...but trains accelerate and decelerate slowly, which is comfortable..." and from there to "...but cars could do the same, making them as comfortable as trains."

The study then knocks down its own straw man by pointing to the horrid, unacceptable, ten- or twenty-second delay at a 4-way intersection.

In fact, such behavior of computer-driven cars will not be considered acceptable as long as they are sharing the road with human-driven cars, so they will not be programmed to drive that way. The first ones will be programmed to drive just like humans drive. They will not have the level of comfort of trains, though you'll still be able to nap or work. Just not as comfortably as on a train.

Once computers have replaced human drivers, then non-compressible car groups will be much more efficient at getting through intersections. We might never have train-level comfort in computer-driven cars in stop-and-go city traffic. But that's the real problem with the study, and why it is fatally flawed: It assumes we will implement train-level comfort at the cost of the annoyance of a ten-second delay at intersections, or while some cars are still being driven by humans.

One individual in this thread makes the common layman's mistake of citing one single study as "proof" of a tangentially-related point, while stubbornly refusing to address or even acknowledge the flaws and limitations of the study. He just keeps repeating "...but the study's title says...!" and extrapolating from a ten-second delay at one hypothetical intersection to an overall worsening of traffic, something the content of the study does not address.
Title: Re: Self-Driving Cars
Post by: teethering on March 29, 2016, 01:30:48 PM
Holy crap.  I don't think I've encountered anyone this unhinged.

The best part of this exchange is that I actually just wanted him to cite a study that actually supports his claim.  That's all.  And one of the reasons why I wanted that is because I do think that driverless cars could make congestion worse, all other things being equal *.  This is one of those things that's really difficult to predict, because it (the traffic ecosystem) is very complex and there are lots of non-linear effects.

One way I thought it could be possible is because driverless cars reduce the cost of being stuck in traffic (i.e. you could be working, reading a book, instead of trying to focus on stop-and-go driving) people might choose to live even further from work for the sake of cheaper suburbian housing.  This sort of effect is difficult to predict and even worse, just like the 4-way intersection study, it's difficult to figure out the interplay with other effects, like reduced number of accidents, which are undoubtedly one of the main advantages of automated cars and one of the ways congestion would be reduced, and make a general prediction.

So I wanted to see what sort of theory of how the ecosystem would be affected our learned friend was citing.  But Jesus Christ on a pogo stick....




* All other things being equal is actually a key point here.  The main driver of road congestion is city planning, specifically suburban sprawl and lack of transportation.  Changing who or what drives the car is a small thing compared to building more dense cities with more public transportation, these are the policies that address congestion directly and effectively.
Title: Re: Self-Driving Cars
Post by: stonesean on March 29, 2016, 02:03:45 PM
I always use "Rupert Sheldrake" as my litmus test to decide if I want to bother or not.....

Low post count is the green light turning yellow, then the inevitable "Rupert Sheldrake" is the flashing red light with the siren.....
Title: Re: Self-Driving Cars
Post by: Fast Eddie B on March 29, 2016, 02:23:31 PM
Just from using Waze, I've seen how real-time, "intelligent" apps can save time, routing a user (me) around traffic jams and slowdowns.

When virtually all vehicles have that sort of app/algorithm built in, it seems obvious to me that traffic could be moved more efficiently, not less.
Title: Re: Self-Driving Cars
Post by: teethering on March 29, 2016, 02:31:07 PM
Just from using Waze, I've seen how real-time, "intelligent" apps can save time, routing a user (me) around traffic jams and slowdowns.

When virtually all vehicles have that sort of app/algorithm built in, it seems obvious to me that traffic could be moved more efficiently, not less.

I don't think that actually scales.  It might, but I really doubt it.

These sorts of apps are very helpful when there are accidents that you could avoid, which is why when they work you really notice.  But most of the congestion problems aren't like that, they're chronic issues with too much traffic volume and not enough transportation options.  And that's where self-driving cars could potentially make things worse, though like I said, it's really hard to weigh all the pros and cons.
Title: Re: Self-Driving Cars
Post by: Morvis13 on March 29, 2016, 02:32:29 PM
Just from using Waze, I've seen how real-time, "intelligent" apps can save time, routing a user (me) around traffic jams and slowdowns.

When virtually all vehicles have that sort of app/algorithm built in, it seems obvious to me that traffic could be moved more efficiently, not less.

unless your car is configured to slow down like a train apparently. Adding 10 seconds to your wait time at stop signs is a game breaker.
<\sarcasm>
Title: Re: Self-Driving Cars
Post by: SkeptiQueer on March 29, 2016, 03:00:22 PM
Just from using Waze, I've seen how real-time, "intelligent" apps can save time, routing a user (me) around traffic jams and slowdowns.

When virtually all vehicles have that sort of app/algorithm built in, it seems obvious to me that traffic could be moved more efficiently, not less.

I don't think that actually scales.  It might, but I really doubt it.

These sorts of apps are very helpful when there are accidents that you could avoid, which is why when they work you really notice.  But most of the congestion problems aren't like that, they're chronic issues with too much traffic volume and not enough transportation options.  And that's where self-driving cars could potentially make things worse, though like I said, it's really hard to weigh all the pros and cons.
A large portion of congestion comes from designing roads one way, like to merge in a left-right pattern at the point of convergence, and drivers doing the opposite, like cutting around the line or merging super-early, or using the left exit lane at a passing lane, or blocking intersections...
Title: Re: Self-Driving Cars
Post by: phooey on March 29, 2016, 04:10:26 PM
The skeptic bandwagon has made the sharp left turn from saying there are no studies which say driverless cars would make traffic congestion worse, how can you say such lies, to your new automated response,  "Oh, that study is totally flawed, can't you see that!"

I think you are using performance art to replicate driverless brains. 

Title: Re: Self-Driving Cars
Post by: brilligtove on March 29, 2016, 04:11:45 PM
Teethering raises a good point about urban design. In big American (and Canadian) cities that were planned for vehicles there may be more flexibility. But London? Madrid? Rome? The roads are not even close to a grid in most of the world, and in many places the roads were not designed for large cars. That's part of why scooters are so popular in parts of Europe.

A single-person SDV with a relatively small footprint might make sense in those places, especially in a fleet. Semi-public transit exists in lots of places now; it could work in a lot of new places in the SDV future.
Title: Re: Self-Driving Cars
Post by: SkeptiQueer on March 29, 2016, 04:30:41 PM
The skeptic bandwagon has made the sharp left turn from saying there are no studies which say driverless cars would make traffic congestion worse, how can you say such lies, to your new automated response,  "Oh, that study is totally flawed, can't you see that!"

I think you are using performance art to replicate driverless brains.
Nobody said there were none, just that you hadn't cited any. Still haven't.

We shouldn't have to specify that the studies examine driverless cars operating in a logical manner under reasonable road conditions (no volcanos spewing aliens on the turnpike) using physics and materials common to this planet in this century in this dimension.

Then, when evidence is presented, we examine the evidence to make sure that it actually supports the premise and doesn't contain any glaring errors the way Sheldrake's study did.

It would be highly unskeptical to just look at the headline of one article and presume that it was definite proof.
Title: Re: Self-Driving Cars
Post by: teethering on March 29, 2016, 04:38:21 PM
The other thing to note re: congestion is that it has two negative effects, one is economic impact and the other is making commuters miserable because they're stuck in traffic.  If someone else is driving even if your commute is longer you can work (less economic impact) or entertain yourself (less misery).

Traffic congestion increase in of itself, IF it's something that driverless cars could make worse, would not necessarily tell us whether we're actually worse off.

Incidentally this is also why a lot of people take public transport even when it's slower than driving yourself.

Ultimately I think cars are a terrible thing at the scale at which they're used and dependent upon in North America and automated driving might ease some of the pain, but the bigger issue is getting urban planning right.
Title: Re: Self-Driving Cars
Post by: daniel1948 on March 29, 2016, 04:56:50 PM
The skeptic bandwagon has made the sharp left turn from saying there are no studies which say driverless cars would make traffic congestion worse, how can you say such lies, to your new automated response,  "Oh, that study is totally flawed, can't you see that!"

I think you are using performance art to replicate driverless brains. 



No. We are saying that the study in question does not show or say what you claim. Only the headline says it "could." Even the headline does not say that it WILL, only that it COULD. And the study itself does not even say that.

You are misquoting the only study that has been cited on the thread, and it was not even you who cited it. You have yet to cite ANY evidence for your thesis.

Cite a real study, conducted in a decently scientific manner, or everyone else on this thread will continue to view your posts as meaningless drivel. And even then, you'll have to convince us that the study you cite is not an outlier or a rogue study. You seem blissfully ignorant of the fact that no single study by itself constitutes a general scientific conclusion. Scientific consensus happens when there is a preponderance of evidence based on a large number of properly-conducted and well-replicated studies.

This thread has examined one study and found it to be deeply flawed AND found that it does not even address the claim you are making except obliquely.
Title: Re: Self-Driving Cars
Post by: Morvis13 on March 30, 2016, 09:06:04 AM
If you keep the discussion on topic and not attack anyone personally then you should be fine. I know the mods will be fair and warn any one that breaks the rules. It's not just you.

There is also no need to air your dirty laundry. You stumbled so get up and get back on the horse so to speak.

Now back on topic.

Is there anything that will change your mind on the self driving car? what do you see it will take to get the robot on the road? Why can't technology over come any of the shortcomings you've expressed so far?



Title: Re: Self-Driving Cars
Post by: Fast Eddie B on March 30, 2016, 10:41:37 AM

I don't think that actually scales.  It might, but I really doubt it.

These sorts of apps are very helpful when there are accidents that you could avoid, which is why when they work you really notice.  But most of the congestion problems aren't like that, they're chronic issues with too much traffic volume and not enough transportation options.  And that's where self-driving cars could potentially make things worse, though like I said, it's really hard to weigh all the pros and cons.

Don't see why it would not scale.

I imagine the algorithm might run something like...

"IF the inner beltway is moving less than 30 mph, AND the outer beltway is moving at least 10 mph faster, THEN route traffic to the outer beltway.

IF both the inner AND outer beltways are moving less than 30 mph, THEN exit traffic at Exit X and route on backroads..."

And so on. If the entire infrastructure is inadequate, no algorithm will eliminate delays. Pretty clear to me it could at least reduce delays.

Like I said, seems like it could only help - if done intelligently.
Title: Re: Self-Driving Cars
Post by: Guillermo on March 30, 2016, 11:01:22 AM
I think that the reason it doesn't scale is because the algorithm (at least it seems to me) is focused on the individual and not on the whole. It will route you through the shortest path at the time, but will route everyone based on the current shortest path without taking into account how routing those will affect the overall traffic and how the shortest path might change because of this. It doesn't predict changes in the traffic after a period of time. IE: 30 minutes from now, the shortest path might not be what you originally said it was. This is no problem when you are routing a small portion of the overall traffic for an individual  target.

So, this algorithm cannot really scale. You'd have to change the algorithm considerably in a way that it would not be recognizable as the Waze algorithm.

Still, I am very confident that there are ways to create algorithms that take into account traffic throughput and traffic routing to obtain a minimum average time on the road per vehicle. This is likely an NP Hard problem so you'd need a pretty robust server to handle it.
Title: Re: Self-Driving Cars
Post by: amysrevenge on March 30, 2016, 11:03:54 AM
This is a solved problem.  It is a distributed networking optimization problem, with cars instead of packets of data.

You can do wonders with localized rules providing network-wide performance.
Title: Re: Self-Driving Cars
Post by: Fast Eddie B on March 30, 2016, 11:40:15 AM
...but will route everyone based on the current shortest path without taking into account how routing those will affect the overall traffic and how the shortest path might change because of this. It doesn't predict changes in the traffic after a period of time. IE: 30 minutes from now, the shortest path might not be what you originally said it was. This is no problem when you are routing a small portion of the overall traffic for an individual  target.

So, this algorithm cannot really scale.

You're making assumptions above about what the algorithm will or won't do.

Why would not a programmer take into account "...how routing those will affect the overall traffic and how the shortest path might change..." Or, "...predict changes in the traffic after a period of time. IE: 30 minutes from now..."

These seem patently obvious factors any programmer or team of programmers would place high on their list of priorities.

Why the heck not?
Title: Re: Self-Driving Cars
Post by: Johnny Slick on March 30, 2016, 12:41:18 PM
I also think that a self-driving car would be far better equipped to plug into the larger network of automobiles and engage in activities that would cause any individual driver to get to their destination slightly slower but, if done by every vehicle on the road, would make the overall experience better for everyone, including the people who might be inconvenienced by any single "business rule" that a vehicle follows. A classic example of this is the zipper merge. IRL nobody wants to allow other cars into the merging lane; it's in that whole "justice" aspect of human/primate nature not to want to allow people not already in the lane to get in. But if you allow two lanes to merge into one, everyone gets where they are going faster. The only people who are really "inconvenienced" are the small percentage of drivers who behave like absolute jackasses on the road. I guess there's also some scenarios where regular drivers would want to bend/break rules of the road (for instance, getting a woman whose water just broke to the hospital) which would not be able to be broken, perhaps rightly.
Title: Re: Self-Driving Cars
Post by: Guillermo on March 30, 2016, 12:44:08 PM
...but will route everyone based on the current shortest path without taking into account how routing those will affect the overall traffic and how the shortest path might change because of this. It doesn't predict changes in the traffic after a period of time. IE: 30 minutes from now, the shortest path might not be what you originally said it was. This is no problem when you are routing a small portion of the overall traffic for an individual  target.

So, this algorithm cannot really scale.

You're making assumptions above about what the algorithm will or won't do.

Why would not a programmer take into account "...how routing those will affect the overall traffic and how the shortest path might change..." Or, "...predict changes in the traffic after a period of time. IE: 30 minutes from now..."

These seem patently obvious factors any programmer or team of programmers would place high on their list of priorities.

Why the heck not?
True, I don't know the exact algorithm, but I did say as much. It's based on my assumptions based on my experience with the application. I'm also using a bit of educated guesswork. I doubt that they will make an effort in prediction when the program will change the route you when traffic heats up 30 minutes later, and the user will be none the wiser that he would have arrived several minutes earlier had he taken another route 30 minutes ago, when it wasn't the shortest path. So, no, I doubt that routing traffic prediction is high on their list of priorities, when that would be to expensive (R&D) compared to checking every x amount of time if a shorter path is available.

So, yes I am making assumptions, but I don't think I'm too far from the reality of what those algorithms can and cannot do.
Title: Re: Self-Driving Cars
Post by: teethering on March 30, 2016, 01:09:49 PM
The problem isn't with the efficiency of the algorithm.  The problem is with the throughput of pipes.  No matter how efficiently you distribute between the pipes (i.e. the highways and roads) there's just a simple problem of volume vs. throughput when it comes to traffic in large cities.  Everyone traveling into downtown in the morning and then out into the suburbs in the evening, no matter how efficiently they're spread out on the various routes are using very limited resources in an inefficient manner because it's mostly one person per car.  This is why I very much doubt that it would scale, I think the perceived gains from everyone using the most efficient algorithm are an illusion based on individual experiences of sometimes being able to beat the traffic.

What does scale is public transport and building denser cities.
Title: Re: Self-Driving Cars
Post by: amysrevenge on March 30, 2016, 01:41:51 PM
There are still efficiencies to be teased out even with a limited bandwidth.  The more I think about it the more this seems like a network optimization problem. 

One that I've even worked on myself in simulation.  In particular, something like using graph theory to optimize the "I know where my destination is, but I don't know or care how I get there - just get me there" problem by making optimal intersection-by-intersection decisions locally without even considering the wider network.  The solutions to local congestion ripple outward to the rest of the network so that the local optimizations end up providing a global optimization.  (In the communication case, the parameters were that no individual node knew or cared about the existence of any nodes it couldn't contact directly - it only made routing decisions based on local information.)
Title: Re: Self-Driving Cars
Post by: PANTS! on March 30, 2016, 01:51:47 PM
Yeah.  I imagine the system should have the ability to devote as much bandwidth as possible.  Right now half (or more) of a given streets bandwidth is inaccessible without some sort of intervention.  Under a networked system, that half of the street that is empty in the AM will simply be reassigned.  That turn lane that no one is using?  Becomes a lane now.
Title: Re: Self-Driving Cars
Post by: brilligtove on March 30, 2016, 02:38:57 PM
We're starting to get into the far(ther) future where the entire traffic system is made up of networked SDVs. What about the short term, with a gradually declining number of human drivers who are not networked in? What about competing networks from different organizations?

I suspect we'll be able to increase the humans per m2 on the road by making single occupant vehicles. If these are designed to form a small herd while their route is the same, they could behave like a bus as far as the routing system is concerned. Cars could swap in and out of the virtual bus as appropriate. Spitballing it, I guess we have 1h/12m2 in very slow moving bumper to bumper traffic today. With the single passenger cars you might be able to get 4 to six people into the same road area (1h/3m2) or more, depending on safety tolerances. Each lane today could become 2 or even three lanes for those miniminiSDVs.
Title: Re: Self-Driving Cars
Post by: amysrevenge on March 30, 2016, 02:54:26 PM
We're starting to get into the far(ther) future where the entire traffic system is made up of networked SDVs. What about the short term, with a gradually declining number of human drivers who are not networked in? What about competing networks from different organizations?

You can do it without common networks.

Consider something like this:  https://en.wikipedia.org/wiki/Geographic_routing only with cars choosing routes rather than data packets.  Each car decides at each intersection how it is going to proceed, based on the conditions in its immediate vicinity.  In grad school I built a face routing simulation in MatLab from scratch; it was probably the most difficult thing I pulled off, but I was just one schlub tapping away on my homework.

The "length" (ie. cost) of individual connections can be weighted by universally available information on conditions, but no over-seeing network knows everything about every car (or even anything about any car, other than at a traffic-level sense of how many cars are where).  Or, competing providers could give different qualities of condition information (shit, my buddy's Toyota network knew about the accident on that block, but my Mazda network didn't find out until I made the turn, too late to avoid it).
Title: Re: Self-Driving Cars
Post by: brilligtove on March 30, 2016, 06:11:49 PM
I suspect having total network oversight would likely have a lot of benefits to local anaonymous routing as you describe. The network starts to get biological in the sense that it has different promoters and inhibitors operating at different scales, all contributing to the final behaviour of the individual cells cars.

It is astonishing how complex the overall system behaviour can be when you're dealing with just simple rules. I've been fascinated by that stuff since that first game of life sim.
Title: Re: Self-Driving Cars
Post by: arthwollipot on March 30, 2016, 06:13:48 PM
According to this article (http://www.driverless-future.com/?page_id=774), thinking of self-driving cars using classical if-then algorithms is fallacious anyway.

Quote
To summarize, we should avoid conceptualizing self-driving vehicles as machines which are controlled by a detailed, exactly specified and in principle comprehensible software program. Instead we should conceptualize their behavior as being the result of a long and varied program of learning. The capability of such cars can be analyzed through simulation and testing but not just by examining its source code.
Title: Re: Self-Driving Cars
Post by: phooey on March 30, 2016, 08:49:31 PM
First you have to get cities to accept that they are going to lose all of that fine ticket revenue (along with their shortened yellow light camera contractors) as well as the right to pull over any blacks they don't like for random phantom traffic violations, so they can check them for warrants on overdue video rentals.

Right now traffic control doesn't seem to be such a great priority for communities (they would rather install more cameras), because there is tons more they could be doing with regards to linking up green lights, and counting the number of cars at a light and adjusting lights better, which most places aren't even doing, or changing stop signs to roundabouts. At this day and age why should you ever have to sit at a red light when there are no cars going the other way?  And yet we still do. 

Making money seems a much higher priority to making things right.   
Title: Re: Self-Driving Cars
Post by: Morvis13 on March 30, 2016, 10:37:12 PM
First you have to get cities to accept that they are going to lose all of that fine ticket revenue (along with their shortened yellow light camera contractors) as well as the right to pull over any blacks they don't like for random phantom traffic violations, so they can check them for warrants on overdue video rentals.

Right now traffic control doesn't seem to be such a great priority for communities (they would rather install more cameras), because there is tons more they could be doing with regards to linking up green lights, and counting the number of cars at a light and adjusting lights better, which most places aren't even doing, or changing stop signs to roundabouts. At this day and age why should you ever have to sit at a red light when there are no cars going the other way?  And yet we still do. 

Making money seems a much higher priority to making things right.

It is really too bad municipal bodies governing all this doesn't make cars. Companies like google are going to put the kids counting cars out of summer jobs. There is already technology to do this anyway. Buying a camera for a couple grand is cheaper and easier than hiring a student anyway. That is still not going to stop self driving cars.

Just like shell doesn't want battery powered cars because they don't make money off them. It really is unfortunate for them economies change and will hopefully make them go the way of the dinosaur.

That still doesn't answer the question of why we have to scrap new technology just because of a limitation.
Title: Re: Self-Driving Cars
Post by: starnado on March 31, 2016, 06:13:14 AM
According to this article (http://www.driverless-future.com/?page_id=774), thinking of self-driving cars using classical if-then algorithms is fallacious anyway.

Quote
To summarize, we should avoid conceptualizing self-driving vehicles as machines which are controlled by a detailed, exactly specified and in principle comprehensible software program. Instead we should conceptualize their behavior as being the result of a long and varied program of learning. The capability of such cars can be analyzed through simulation and testing but not just by examining its source code.

I think Arth has it just right. Complex dynamic systems exhibit strange and unpredictable effects that are often not amenable to such 'situation based' if-then algorithms. I thought google et al were using deep learning systems and other forms of AI that learn through experience to create flexible fuzzy algorithms*. I don't think that any programmer or team of programmers can realistically program a car that can respond to every situation. I remember from that great book 'Chaos' by James Gleick that genetically evolved algorithms (similar to AI learning but more basic) can be incredibly efficient - I imagine that the algorithms the cars will use will be almost incomprehensible on some levels to us mere humans although we will have a general understanding of what each bit of stuff will be doing.

If individual car units will be using such systems, surely the overall network managers will be running similar flexible fuzzy systems. Seems that the laws and norms of driving are the boundary conditions or limits of the system but that a flexible AI 'network process manager'** will be able to find unexpected or counter-intuitive ways to minimize congestion and reduce the risk of harm.

* is that even a thing? You know what I mean, anyway.
** See above. Sorry for the poor use of terminology that is above my rank and grade.
Title: Re: Self-Driving Cars
Post by: daniel1948 on March 31, 2016, 08:46:02 AM
So Phooey has flipped from "Self-driving cars won't work," and "Mothers won't trust an SDV with her baby's safety even if the SDV is a safer driver than she is," to "Cities won't allow them because they'd undermine the cities' nefarious ways of making a few extra bucks."

This seems like a tacit agreement that computer-driven cars will work, if the argument against them is that they'd be so efficient that cities would lose traffic-fine revenues. So which does Phooey hate more: Computer-driven cars, or traffic fines? We may never know.
Title: Re: Self-Driving Cars
Post by: Fast Eddie B on March 31, 2016, 11:03:58 AM

I think Arth has it just right. Complex dynamic systems exhibit strange and unpredictable effects that are often not amenable to such 'situation based' if-then algorithms.

I'm not a programmer. So when I came up with a cartoonish "IF>THEN>ELSE" example, it was just to illustrate that an algorithm could take into account a LOT of things that it was being assumed they would not.

Stipulated it might not be a simple algorithm, but a "genetically evolved algorithm" is still an algorithm.

And, ultimately, no matter how complex or "fuzzy" or how derived, in the end is it not reducible to one's and zero's?
Title: Re: Self-Driving Cars
Post by: brilligtove on March 31, 2016, 11:32:41 AM

I think Arth has it just right. Complex dynamic systems exhibit strange and unpredictable effects that are often not amenable to such 'situation based' if-then algorithms.

I'm not a programmer. So when I came up with a cartoonish "IF>THEN>ELSE" example, it was just to illustrate that an algorithm could take into account a LOT of things that it was being assumed they would not.

Stipulated it might not be a simple algorithm, but a "genetically evolved algorithm" is still an algorithm.

And, ultimately, no matter how complex or "fuzzy" or how derived, in the end is it not reducible to one's and zero's?


Yes and no.

At this level of complexity the program is more like biology than traditional coding. So yes, it can be read, and could even be parsed by a human after years of study (for a complex algoithm). A human can understand it in parts, and we can develop tools to do amazing things to manipulate the system at various levels of complexity / detail. What no one human could do is really grok the entirety of the learned AI code, any more than a person could read out and understand the entire genome of a creature. A prodigious feat of memorization might be possible, to recall all the information, and a capacity to understand how any portion of that information functions can be developed. But being able to look at the code and grok what the code will do? It's too complex for human comprehension at that level.

The kind of AI we can understand relatively easily is called an expert system. These are rules-based systems that walk through a fairly narrow decision tree to provide expert advice in a specific domain. These are really useful tools for experts to use when doing diagnostics (and perhaps in other pursuits). In a sense they are intelligent checklists.

The last few decades of AI work has been all about creating systems that can learn about their environment by exposure to that environment. This is analogous to having an embodied consciousness. Today's best AIs are still quite narrow, and can learn very limited domains. So far we have narrow AIs that do natural language recognition, play Go, walk, understand a visual environment, navigate, and so on. Even with the narrow AIs we have today, the algorithms they generate for themselves - the code - is fairly inscrutable.

What we don't have is one AGI - Artificial General Intelligence - that can do all the things that humans can do.
Title: Re: Self-Driving Cars
Post by: amysrevenge on March 31, 2016, 12:03:36 PM
At this level of complexity the program is more like biology than traditional coding.

Yeah, that's sort of how it goes.  You start from small scale local decisions, based on small scale local information, and you get this emergent large-scale behaviour.
Title: Re: Self-Driving Cars
Post by: teethering on March 31, 2016, 01:52:35 PM
I don't find the distinction useful to be honest.  The learning algorithms is just an implementation of a system that can still be described as "if then", it's just too cumbersome to actually describe the full detail of the "if then" chains because a) they're largely hidden in a generated data structure and b) that's kind of the point of using learning algorithms to bypass all of that nonsense of writing down rules and get to the desired behaviour by having an algorithm construct them implicitly.

Perhaps the mistaken impression that people get is that the "if thens" are easy to correct by just adding more rules, maybe that's what the point of that section in the article was.  That would definitely be a wrong impression to have.  It would also be wrong to expect that you can inspect your system and figure out precisely what rules it follows.  That's not really feasible.  But conceptually thinking of it as just a very complicated chain of "if thens" seems ok to me.
Title: Re: Self-Driving Cars
Post by: Johnny Slick on March 31, 2016, 02:30:00 PM
But what if they implement a switch/case system instead?? ?? ?? HOW WILL WE SURVIVE
Title: Re: Self-Driving Cars
Post by: brilligtove on March 31, 2016, 05:13:38 PM
Johnny Slick, you ragamuffin, you.

Teethering, I'm not sure that a mass of if/then statements really is a good analogy. I mean, sure, you might be able to encode it that way, but that doesn't capture the sense of competing algorithms producing vectored potentials for an action, with the final behaviour of the system being a consequence of all those potentials combined. Right this minute I am experiencing one movitation to tear off my clothes because it's effing hot with no breeze, and other motivations in other directions. The combination of all of these has me undoing the buttons of my shirt and drinking lots of ice water icy cold beer. :)

Edited to correct the drink.
Title: Re: Self-Driving Cars
Post by: Johnny Slick on March 31, 2016, 05:27:13 PM
I mean, we already have neural networks, which is basically "we dump a bunch of data into the program, it does something inside of a black box that we don't really know about, and then it spits out an answer". Those are best used for yes/no questions but the gestalt level decision making can surely be boiled down to that. The problem, I guess, with neural networks is that when it comes to a conclusion we don't agree with it can be hard to ferret out the reasoning behind said conclusion.
Title: Re: Self-Driving Cars
Post by: brilligtove on March 31, 2016, 05:37:53 PM
Sometimes the output is just wrong, so more training / data can help correct the system's poor judgement. Sometimes the output is right but not what a human would do. These are interesting, because they can expose ways our cognitive biases make us so we are - and point to other ways of thinking.

In the case of SDVs, I'm certain that they are going to make some decisions that are not at all what humans would do, and that will take some getting used to. Maybe even a new portion of driver's Ed when you renew your license. A module on "Adapting to new safety technologies in your car and in the vehicles around you" or some such.
Title: Re: Self-Driving Cars
Post by: teethering on March 31, 2016, 11:08:52 PM
Johnny Slick, you ragamuffin, you.

Teethering, I'm not sure that a mass of if/then statements really is a good analogy. I mean, sure, you might be able to encode it that way, but that doesn't capture the sense of competing algorithms producing vectored potentials for an action, with the final behaviour of the system being a consequence of all those potentials combined. Right this minute I am experiencing one movitation to tear off my clothes because it's effing hot with no breeze, and other motivations in other directions. The combination of all of these has me undoing the buttons of my shirt and drinking lots of ice water icy cold beer. :)

Edited to correct the drink.

I feel like this is a bit of "but that's not how I implemented it" sort of argument, when we're actually talking about a very high level description of how system behaves, i.e. the requirements.  I don't see a reason to care about the distinction in how we implement a very complex system that has to deterministically map inputs to desired behaviour.  It certainly is easier to conceptualize the emergent functionality as really complicated "if thens" rules than as an implicit map between input and output generated by an algorithm through learning.  Now if someone could point out an actual functional difference between a model of "if thens" and a learning algorithm model, then I'd understand.  But ultimately how you implement the mapping (which is really what we're talking about) is unimportant to thinking about it for anyone who isn't actually writing code or tests for this system.

This is why I tried to think about all the pitfalls that could arise from not keeping in mind how the AI is implemented and I couldn't think of any that actually matter aside from where the difficulties in the implementation actually are, as opposed to where you might imagine them to be.

As for competing algorithms that also does not matter, because at some point there has to be an arbiter that either strictly decides between them or aggregates their output in a deterministic fashion.  Again, it's "if-thens" all the way down in the sense of being able to model it in this way.

Anyway, that passage sort of bothered me because I couldn't figure out in what way that "misconception" actually matters to how we talk about self-driving cars.
Title: Re: Self-Driving Cars
Post by: arthwollipot on March 31, 2016, 11:22:29 PM
First you have to get cities to accept that they are going to lose all of that fine ticket revenue (along with their shortened yellow light camera contractors) as well as the right to pull over any blacks they don't like for random phantom traffic violations, so they can check them for warrants on overdue video rentals.

Right now traffic control doesn't seem to be such a great priority for communities (they would rather install more cameras), because there is tons more they could be doing with regards to linking up green lights, and counting the number of cars at a light and adjusting lights better, which most places aren't even doing, or changing stop signs to roundabouts. At this day and age why should you ever have to sit at a red light when there are no cars going the other way?  And yet we still do. 

Making money seems a much higher priority to making things right.

Self-driving vehicles will make traffic lights entirely obsolete.
Title: Re: Self-Driving Cars
Post by: brilligtove on March 31, 2016, 11:57:52 PM
Adding vectors that represent potentially conflicting motivations in a multidimensional space is definitely a different analogy than lots of if-then statements. It isn't about how I would or want to implement it: it's that no human is putting any kind of if-then anywhere near these systems.

I'm suggesting that there is a much better to think about SDVs and other AIs: they have motivations and desires, not if-then rules. This means that they are not guaranteed to behave according to any one rule any more than a domesticated animal would. I mean sure, there are rules my dog Walter won't break. I'm utterly certain that he will die before he bites me or a child under any circumstance, for example. Still, there are rules that he breaks all the time, given sufficient motivation. Launching his nearly 100lb self after squirrels while still attached to my wife's wrist is one of the worst. :)

To understand Wally's behaviour - and that of most modern AIs - it makes more sense to think about motivations. With my wife, the motivation to chase the live furry toy overwhelms the motivation to walk with her. (It happens with me too, but I'm heavier so the "getting flopped on my back when I'm pulled up short" motivation is stronger for Wally when I'm holding the leash.)

Now let's do a thought experiment (so assume ideal materials for ropes etc., that are weightless and can't break and so on).

Imagine a 10m tall flexible pole. Attached to the pole right at the top are 360 (weightless, unbreakable) cords. Each cord tethers a horse to the pole. The horses are arrayed in a perfect circle - one for each degree - but the horses are not the same. Some are big, some small; some are stronger or weaker; some have endurance and others tire rapidly.

Your Self Driving Vehicle is designed to accelerate in the direction the pole is pointing, with a force to the occupants based on how much force the pole is experiencing. For example, if one horse pulls the pole forward at 100N (newtons) the SDV will accelerate forward hard enough to put 100N of you against your seat. If the pole points to 120° with 1000N you're going to be breaking fairly hard and pulling right (say, to avoid a collision with a mattress that just flew off that hatchback up ahead).

There are 360 horses, but remember the each have different characteristics. When they start to pull (based on whatever motivates them) the combination of all their efforts will bend the pole in only one direction, and with only one amount of force.

In the AI scenario we're talking about with SDVs, there are likely a lot more 'horses' and many more than two dimensions that they are pulling in. Even so, some vector is the result of the combined efforts off all these algorithms. Neural networks in our brains are similar, btw, with some neurons firing only when (for example) the combination of many inputs crosses a threshold.

You can attempt to turn this concept of competing motives into a complex of if-then statements, but I don't think that perspective serves you well at this level of complexity. Neural networks (and other machine learning AI systems) process information in a manner that doesn't easily translate into those kind of simple rules. Expert systems use if-then statements that humans can understand and encode, and these are extraordinarily useful. It doesn't mean that it is the most effective way to think about this kind of computation however.
Title: Re: Self-Driving Cars
Post by: brilligtove on March 31, 2016, 11:58:30 PM
First you have to get cities to accept that they are going to lose all of that fine ticket revenue (along with their shortened yellow light camera contractors) as well as the right to pull over any blacks they don't like for random phantom traffic violations, so they can check them for warrants on overdue video rentals.

Right now traffic control doesn't seem to be such a great priority for communities (they would rather install more cameras), because there is tons more they could be doing with regards to linking up green lights, and counting the number of cars at a light and adjusting lights better, which most places aren't even doing, or changing stop signs to roundabouts. At this day and age why should you ever have to sit at a red light when there are no cars going the other way?  And yet we still do. 

Making money seems a much higher priority to making things right.

Self-driving vehicles will make traffic lights entirely obsolete.

How will you handle those pesky pedestrians?
Title: Re: Self-Driving Cars
Post by: DG on April 01, 2016, 12:38:56 AM

Self-driving vehicles will make traffic lights entirely obsolete.

How will you handle those pesky pedestrians?

My personal preference would be to provide over and under passes for pedestrian access. As it is I try to walk as far as I can from one end of Sydney CBD to the other without having to cross roads. Generally I'd rather the longer walk than a more direct route that takes just as long because you're always stopping and waiting at lights.

Alternatively, rather than crossing at intersections, crossings should be well away from intersections where possible. Meaning that pedestrians need only worry about vehicles approaching along the road that they are crossing and not those joining the road. Of course this would be a huge culture change and, since it appears that people are unwilling to wait for the lights to change now, I can't imagine that moving the crossings would help.
Title: Re: Self-Driving Cars
Post by: brilligtove on April 01, 2016, 12:59:08 AM
I was going to say that overpasses are not gonna fly when there is any other option because you have to walk eight times the distance to cross the road. This shows my ignorance of pedestrian overpasses based on a limited sample size. The only one or two I recalled are very looooooong ramps with several switchbacks, rather than a (relatively) short staircase and bridge.

On city streets though, overpasses are not going to fly. Having a 30s scramble between cycles of lights would address most pedestrians, I would think.
Title: Re: Self-Driving Cars
Post by: DG on April 01, 2016, 01:24:36 AM
I was going to say that overpasses are not gonna fly when there is any other option because you have to walk eight times the distance to cross the road. This shows my ignorance of pedestrian overpasses based on a limited sample size. The only one or two I recalled are very looooooong ramps with several switchbacks, rather than a (relatively) short staircase and bridge.

On city streets though, overpasses are not going to fly. Having a 30s scramble between cycles of lights would address most pedestrians, I would think.

The model that I like is something in which a network akin to a subway exists for pedestrian access. Walkways could follow the path of existing roads (I am assuming infinite financial resources here  ;D)
Title: Re: Self-Driving Cars
Post by: daniel1948 on April 01, 2016, 09:48:38 AM
How will you handle those pesky pedestrians?

Walk/Don't-Walk lights for the pedestrians, to let them know when it is safe to cross. As long as there are pedestrians crossing streets, the cars will have to stop sometimes. The computer network controls the cars and knows when a cross button has been pressed, and manages the cars accordingly, stopping them if needed, or altering their speed to create gaps for the pedestrians. In congested areas, the cars will have to stop for cross traffic anyway, sometimes. The network will decide when traffic must stop, and when coordinated interlacing is possible.

Apparently the Tesla Model S has lane-keeping and lane-changing and parking already. I'm sure lots of other automakers do as well. The future is on our doorstep.
Title: Re: Self-Driving Cars
Post by: phooey on April 01, 2016, 10:19:21 AM
If you are talking about a time when there are no longer any human driven vehicles, and everything is controlled by computer, you are now so far into the future, that you don't even need to consider roads.  people will be using flying cars by then. 

Plus right now cities don't even have enough money to fix potholes and replace aging bridges, and we are going to suddenly make all streets have pedestrians walkways under them? 

That will cost more than Donald Trumps wall around America. 
Title: Re: Self-Driving Cars
Post by: Johnny Slick on April 01, 2016, 10:24:00 AM
gusy i mae nto haev a coherent argument but i haz kweshtuns!!!!
Title: Re: Self-Driving Cars
Post by: teethering on April 01, 2016, 12:37:42 PM
Adding vectors that represent potentially conflicting motivations in a multidimensional space is definitely a different analogy than lots of if-then statements. It isn't about how I would or want to implement it: it's that no human is putting any kind of if-then anywhere near these systems.

I'm suggesting that there is a much better to think about SDVs and other AIs: they have motivations and desires, not if-then rules. This means that they are not guaranteed to behave according to any one rule any more than a domesticated animal would. I mean sure, there are rules my dog Walter won't break. I'm utterly certain that he will die before he bites me or a child under any circumstance, for example. Still, there are rules that he breaks all the time, given sufficient motivation. Launching his nearly 100lb self after squirrels while still attached to my wife's wrist is one of the worst. :)

To understand Wally's behaviour - and that of most modern AIs - it makes more sense to think about motivations. With my wife, the motivation to chase the live furry toy overwhelms the motivation to walk with her. (It happens with me too, but I'm heavier so the "getting flopped on my back when I'm pulled up short" motivation is stronger for Wally when I'm holding the leash.)

Now let's do a thought experiment (so assume ideal materials for ropes etc., that are weightless and can't break and so on).

Imagine a 10m tall flexible pole. Attached to the pole right at the top are 360 (weightless, unbreakable) cords. Each cord tethers a horse to the pole. The horses are arrayed in a perfect circle - one for each degree - but the horses are not the same. Some are big, some small; some are stronger or weaker; some have endurance and others tire rapidly.

Your Self Driving Vehicle is designed to accelerate in the direction the pole is pointing, with a force to the occupants based on how much force the pole is experiencing. For example, if one horse pulls the pole forward at 100N (newtons) the SDV will accelerate forward hard enough to put 100N of you against your seat. If the pole points to 120° with 1000N you're going to be breaking fairly hard and pulling right (say, to avoid a collision with a mattress that just flew off that hatchback up ahead).

There are 360 horses, but remember the each have different characteristics. When they start to pull (based on whatever motivates them) the combination of all their efforts will bend the pole in only one direction, and with only one amount of force.

In the AI scenario we're talking about with SDVs, there are likely a lot more 'horses' and many more than two dimensions that they are pulling in. Even so, some vector is the result of the combined efforts off all these algorithms. Neural networks in our brains are similar, btw, with some neurons firing only when (for example) the combination of many inputs crosses a threshold.

You can attempt to turn this concept of competing motives into a complex of if-then statements, but I don't think that perspective serves you well at this level of complexity. Neural networks (and other machine learning AI systems) process information in a manner that doesn't easily translate into those kind of simple rules. Expert systems use if-then statements that humans can understand and encode, and these are extraordinarily useful. It doesn't mean that it is the most effective way to think about this kind of computation however.

The idea is that at this level of discussion no one cares about this level of detail in understanding exactly how complex it is.  When we discuss system's behaviour we are talking about specific use cases and we're considering how we expect the system to behave in that scenario, e.g. "human in the street, brake or swerve", which both "if-then" and self-learning algos model homologically.  it's just self-learning algorithms in practice are better at capturing all the complexity you're talking about, but we're talking about it at the level where we bracket all that complexity to zoom in on abstract thought experiments.

All that we really care about is that we understand that computers in the car make decisions based sensor data.  The other thing that we care about specifically is that the mapping of sensor data to behaviour is deterministic.  This is critical to being able to predict what the car will do, which is critical for it to being certified as safe for being on the road.  It is also categorically different from the way your labrador behaves, which is not deterministic.  Which is why that's actually a terrible analogy for this system.
Title: Re: Self-Driving Cars
Post by: SkeptiQueer on April 01, 2016, 01:27:28 PM
If you are talking about a time when there are no longer any human driven vehicles, and everything is controlled by computer, you are now so far into the future, that you don't even need to consider roads.  people will be using flying cars by then. 

What are you basing this claim on?
Title: Re: Self-Driving Cars
Post by: daniel1948 on April 01, 2016, 02:31:57 PM
If you are talking about a time when there are no longer any human driven vehicles, and everything is controlled by computer, you are now so far into the future, that you don't even need to consider roads.  people will be using flying cars by then. 

What are you basing this claim on?

The Jetsons.
Title: Re: Self-Driving Cars
Post by: arthwollipot on April 01, 2016, 07:02:48 PM
As long as there are pedestrians crossing streets, the cars will have to stop sometimes.

No they won't, they'll just have to slow down. Just like getting through intersections where there used to be traffic lights.

Ideally, a pedestrian could simply walk straight out onto the road without even looking and the cars would alter their course and speed in order to avoid them.
Title: Re: Self-Driving Cars
Post by: teethering on April 01, 2016, 07:50:19 PM
As long as there are pedestrians crossing streets, the cars will have to stop sometimes.

No they won't, they'll just have to slow down. Just like getting through intersections where there used to be traffic lights.

Ideally, a pedestrian could simply walk straight out onto the road without even looking and the cars would alter their course and speed in order to avoid them.

I don't think that's feasible on city streets.  A car does have a stopping distance and predicting when a pedestrian would suddenly step from the sidewalk onto the street before you need to engage the brakes isn't possible.
Title: Re: Self-Driving Cars
Post by: brilligtove on April 01, 2016, 08:03:31 PM
As long as there are pedestrians crossing streets, the cars will have to stop sometimes.

No they won't, they'll just have to slow down. Just like getting through intersections where there used to be traffic lights.

Ideally, a pedestrian could simply walk straight out onto the road without even looking and the cars would alter their course and speed in order to avoid them.

I don't think that's feasible on city streets.  A car does have a stopping distance and predicting when a pedestrian would suddenly step from the sidewalk onto the street before you need to engage the brakes isn't possible.

Yeah, a pedestrian could still arrange to get hit by a car pretty easily, intentionally or not. A kid running out from between parked cars directly in front (2m) of a vehicle moving at 40km/h is going to get hit unless the car can launch itself into the air.
Title: Re: Self-Driving Cars
Post by: Alex Simmons on April 02, 2016, 01:29:51 AM
https://www.youtube.com/watch?v=LSZPNwZex9s
Title: Re: Self-Driving Cars
Post by: arthwollipot on April 02, 2016, 01:33:55 AM
As long as there are pedestrians crossing streets, the cars will have to stop sometimes.

No they won't, they'll just have to slow down. Just like getting through intersections where there used to be traffic lights.

Ideally, a pedestrian could simply walk straight out onto the road without even looking and the cars would alter their course and speed in order to avoid them.

I don't think that's feasible on city streets.  A car does have a stopping distance and predicting when a pedestrian would suddenly step from the sidewalk onto the street before you need to engage the brakes isn't possible.

Yeah, a pedestrian could still arrange to get hit by a car pretty easily, intentionally or not. A kid running out from between parked cars directly in front (2m) of a vehicle moving at 40km/h is going to get hit unless the car can launch itself into the air.

Well, I did say "ideally". And a vehicle doesn't necessarily have to stop in order to avoid a pedestrian. When all the cars are networked together, the most in-danger car could swerve, and all the other cars in the area could move to allow the swerve. Being computer-controlled and networked, vehicles could get much closer together at much higher speeds than humans could ever deal with.

I recall reading something like this - it may have been in I, Robot, where a person could walk straight out into the middle of a swarm of robotic vehicles without any risk, because the First Law of Robotics prevented a robot from harming a human or through inaction allowing a human to come to harm. The networked swarm would be able to cope.

But that was fiction.
Title: Re: Self-Driving Cars
Post by: phooey on April 02, 2016, 06:33:16 AM
The insanity continues (cars that swerve and brake for unconcerned pedestrians, oh brother...).

Also, the self driving bike is also obviously not real-though I bet a lot of people who saw that thought it was. 
Title: Re: Self-Driving Cars
Post by: brilligtove on April 02, 2016, 06:42:29 AM
(click to show/hide)

The idea is that at this level of discussion no one cares about this level of detail in understanding exactly how complex it is.  When we discuss system's behaviour we are talking about specific use cases and we're considering how we expect the system to behave in that scenario, e.g. "human in the street, brake or swerve", which both "if-then" and self-learning algos model homologically.  it's just self-learning algorithms in practice are better at capturing all the complexity you're talking about, but we're talking about it at the level where we bracket all that complexity to zoom in on abstract thought experiments.

All that we really care about is that we understand that computers in the car make decisions based sensor data.  The other thing that we care about specifically is that the mapping of sensor data to behaviour is deterministic.  This is critical to being able to predict what the car will do, which is critical for it to being certified as safe for being on the road.  It is also categorically different from the way your labrador behaves, which is not deterministic.  Which is why that's actually a terrible analogy for this system.

AI systems like these often use true random number generation in the decision process. They are not strictly deterministic, because one of the weighted inputs is not. The dog's behavioral constraints still seems like a good fit to me.

Title: Re: Self-Driving Cars
Post by: daniel1948 on April 02, 2016, 09:04:19 AM
As long as there are pedestrians crossing streets, the cars will have to stop sometimes.

No they won't, they'll just have to slow down. Just like getting through intersections where there used to be traffic lights.

Ideally, a pedestrian could simply walk straight out onto the road without even looking and the cars would alter their course and speed in order to avoid them.

On city streets there are just too many cars for this to work. At night, when there are only a few cars and only a few pedestrians, it could. Networked control will greatly improve efficiency, but there will still be a need for cars to stop at some crossings, some of the time.

https://www.youtube.com/watch?v=LSZPNwZex9s

Damn! I missed it by one day. Only available yesterday and I really, really want one of these. ;D
Title: Re: Self-Driving Cars
Post by: Alex Simmons on April 03, 2016, 02:08:32 AM
The insanity continues (cars that swerve and brake for unconcerned pedestrians, oh brother...).

Also, the self driving bike is also obviously not real-though I bet a lot of people who saw that thought it was.
I think most would readily spot the fun of this April fool's joke.
Title: Re: Self-Driving Cars
Post by: Fast Eddie B on April 03, 2016, 08:40:08 AM

On city streets there are just too many cars for this to work.

Seems like that's only limited by processing power.

I imagine a rough analogy is how a web page makes it to your computer - in packets that route pretty much seamlessly behind the scenes, appearing to arrive as a "whole" when in fact multiple servers and routers and switches may have been involved.

The above analogy may be flawed - it's all pretty much magic to me. But I still see the problem of "just too many cars" as solvable with the help of Moore's Law!
Title: Re: Self-Driving Cars
Post by: brilligtove on April 03, 2016, 09:16:44 AM

On city streets there are just too many cars for this to work.

Seems like that's only limited by processing power.

I imagine a rough analogy is how a web page makes it to your computer - in packets that route pretty much seamlessly behind the scenes, appearing to arrive as a "whole" when in fact multiple servers and routers and switches may have been involved.

The above analogy may be flawed - it's all pretty much magic to me. But I still see the problem of "just too many cars" as solvable with the help of Moore's Law!

There are some real limitations to internet packet routing that are not about computational power. They are about the way the protocols work, and what they do when things go wrong. These are dealt with by altering the way servers and routers handle congestion, rather than making bigger pipes or adding more power.

I'm sure there will be analogous situations with SDVs, though the details of the solutions will be different. You can't drop and resent a car when it can't get to its destination (that is, destroy it and fire off a copy after a short wait).

That would reduce the world population pretty quickly though...
Title: Re: Self-Driving Cars
Post by: arthwollipot on April 04, 2016, 09:41:55 PM
On city streets there are just too many cars for this to work.

Ironically, this is the very problem that this solution is intended to address.
Title: Re: Self-Driving Cars
Post by: brilligtove on April 05, 2016, 04:33:00 AM
http://youtu.be/UgNhYGAgmZo

It seems the future is now.
Title: Re: Self-Driving Cars
Post by: phooey on April 05, 2016, 11:20:36 AM
Musk says the biggest challenge for autonomous vehicles is not high-speed motorways, but navigating urban streets safely at speeds of between 15mph and 50mph. “It’s the intermediate that’s hard,” he said.

“It’s too dangerous. You can’t have a person driving a two-tonne death machine,” said Musk during an appearance at Nvidia’s annual developers conference, where he discussed Tesla’s ambitionsfor autonomous-cars.

Musk suggested that any wholesale switch from human-driven to self-driving cars could take 20 years, because of the sheer scale of the challenge in replacing cars already on the roads.

Elon is as aware of this problem as I am. 



Elon Musk: self-driving cars could lead to ban on human drivers   
http://www.theguardian.com/technology/2015/mar/18/elon-musk-self-driving-cars-ban-human-drivers
Title: Re: Self-Driving Cars
Post by: phooey on April 05, 2016, 11:26:03 AM
You’re driving through a tunnel and a child appears at the opening and trips, blocking your exit. You can’t stop, so you’re left with the choice of swerving into a wall to save the child, or running over the child to save yourself. Of 110 people polled, 64 percent said they would continue straight and kill the child.

When asked which entities should determine how an autonomous car responds to the tunnel problem, 44 percent of respondents thought it should be the passenger of the vehicle, while 33 percent thought it should be lawmakers. Twelve percent thought the manufacturers or designers should be burdened with that choice; 11 percent responded “other.”


Determining who will control these “ethical settings” that guide no-win responses is a huge problem that self-driving cars are going to have to face in terms of liability. Because if a car will have to be programmed to choose between two lives, that means whoever decides how the algorithm is going to function is also possibly condemning to death either bystanders or passengers.


This type of “predetermined” action, an algorithm that chooses to spare children over adults, for example, would almost vicariously put the programmer in the driver’s seat, lending truth to the Department of Public Safety’s comment that a self-driven car will always have a “determinable human operator.”


http://techcrunch.com/2016/02/27/self-driving-cars-and-the-kobayashi-maru/
Title: Re: Self-Driving Cars
Post by: amysrevenge on April 05, 2016, 11:43:02 AM
Musk suggested that any wholesale switch from human-driven to self-driving cars could take 20 years, because of the sheer scale of the challenge in replacing cars already on the roads.

Elon is as aware of this problem as I am.

He sure is.  "Could take 20 years" is pretty much what you've been saying all along.  Not "not in 200 years", not "unsolvable", not "never gonna happen", but rather "could take 20 years".  That's the phooey I've been reading the whole time.
Title: Re: Self-Driving Cars
Post by: phooey on April 05, 2016, 11:47:28 AM
https://youtu.be/_8nnhUCtcO8(https://youtu.be/_8nnhUCtcO8)
Title: Re: Self-Driving Cars
Post by: Morvis13 on April 05, 2016, 11:49:27 AM
Another key to that argument is "wholesale". Not everyone switched from beta to VHS and not everyone switched from VHS to DVD. Technology doesn't need to be implemented wholesale. We will see early adopters and a gradual increase. That's not stopping our robot overlords from taking over driving for us.
Title: Re: Self-Driving Cars
Post by: phooey on April 05, 2016, 11:54:59 AM
Musk suggested that any wholesale switch from human-driven to self-driving cars could take 20 years, because of the sheer scale of the challenge in replacing cars already on the roads.

Elon is as aware of this problem as I am.

He sure is.  "Could take 20 years" is pretty much what you've been saying all along.  Not "not in 200 years", not "unsolvable", not "never gonna happen", but rather "could take 20 years".  That's the phooey I've been reading the whole time.


If there’s one issue that has plagued every single Tesla vehicle since the company’s inception, it’s product delays. The Tesla Roadster launch was delayed numerous times, as was the Model S. More recently, the Model X launched late last year after a 2+ year delay. What’s more, even customers who already put orders in for the Model X months ago won’t actually see the car show up on their driveway until mid-late 2016.


If Elon says 20 years, I guess more like at least 60.  He is usually off at least two or three fold.   
Title: Re: Self-Driving Cars
Post by: phooey on April 05, 2016, 12:04:43 PM
Driving in Circles
The autonomous Google car may never actually happen.

http://www.slate.com/articles/technology/technology/2014/10/google_self_driving_car_it_may_never_actually_happen.html




"To be able to handle the everyday stresses and strains of the real driving world, the Google car will require a computer with a level of intelligence that machines won't have for many years, if ever."

...said MIT roboticist John Leonard, who does not expect a full self-driving car in his lifetime (he’s 49).


"We tend to lionize computer researchers, forgetting that they've made some colossally bad predictions over the years. When 2001: A Space Odyssey premiered in 1968, MIT's Marvin Minsky assured the public that machines like HAL would indeed be possible in 30 years. Perhaps one day tech enthusiasts will be able to visit a Museum of the Future That Never Was, where the Jetsons’ hover car and the Google super-robocar will sit side-by-side as showcase exhibits. Expect long lines for both, because the demos will be sensational."


I know, I know, MIT is full of kooks and cranks.  Just more trolls. 
Title: Re: Self-Driving Cars
Post by: phooey on April 05, 2016, 12:07:21 PM
Another key to that argument is "wholesale". Not everyone switched from beta to VHS and not everyone switched from VHS to DVD. Technology doesn't need to be implemented wholesale. We will see early adopters and a gradual increase. That's not stopping our robot overlords from taking over driving for us.


No Morvis, you don't seem to understand what is being suggested.  Elon is saying that you CAN'T do it gradually, because the two types of technology, human driven cars and computer driven cars can't co-exist together.   
Title: Re: Self-Driving Cars
Post by: SkeptiQueer on April 05, 2016, 12:27:01 PM
You’re driving through a tunnel and a child appears at the opening and trips, blocking your exit. You can’t stop, so you’re left with the choice of swerving into a wall to save the child, or running over the child to save yourself. Of 110 people polled, 64 percent said they would continue straight and kill the child.

When asked which entities should determine how an autonomous car responds to the tunnel problem, 44 percent of respondents thought it should be the passenger of the vehicle, while 33 percent thought it should be lawmakers. Twelve percent thought the manufacturers or designers should be burdened with that choice; 11 percent responded “other.”


Determining who will control these “ethical settings” that guide no-win responses is a huge problem that self-driving cars are going to have to face in terms of liability. Because if a car will have to be programmed to choose between two lives, that means whoever decides how the algorithm is going to function is also possibly condemning to death either bystanders or passengers.


This type of “predetermined” action, an algorithm that chooses to spare children over adults, for example, would almost vicariously put the programmer in the driver’s seat, lending truth to the Department of Public Safety’s comment that a self-driven car will always have a “determinable human operator.”


http://techcrunch.com/2016/02/27/self-driving-cars-and-the-kobayashi-maru/

The fastest human in the world will still hit the brakes a half a second slower than the car. To put that in perspective, at 35 miles per hour that car stops at least 51 feet faster than a human, or swerves, or whatever. More realistically, it's 100-125 feet, because most people can't react and begin braking with the same speed that a professional drag racer can punch the accelerator when the tree goes green. Some studies actually put the time to brake for drivers between 1.5 and 2.3 seconds, according to the NHTSA.

Again, coming up with contrived one-off scenarios doesn't tell you anything unless you compare the likelihood of those scenarios and their damages against the stuff that happens every day, like a driver eating who drops their cheeseburger and plows into a light pole at highway speed.

I'm still confused though, why we're pretending like a car is going to be driving through a tunnel where a kid falls in the way and smashing into the side of the tunnel is a reasonable response. Is the car a Model T, one of three cars in the town, where plowing into the side of the tunnel doesn't turn one dead kid into six dead people a several million dollars in private and public property damage and hospital bills?
Title: Re: Self-Driving Cars
Post by: Morvis13 on April 05, 2016, 12:40:51 PM
Another key to that argument is "wholesale". Not everyone switched from beta to VHS and not everyone switched from VHS to DVD. Technology doesn't need to be implemented wholesale. We will see early adopters and a gradual increase. That's not stopping our robot overlords from taking over driving for us.


No Morvis, you don't seem to understand what is being suggested.  Elon is saying that you CAN'T do it gradually, because the two types of technology, human driven cars and computer driven cars can't co-exist together.

Current automated driving systems being developed and already on the roads prove this wrong.
Title: Re: Self-Driving Cars
Post by: teethering on April 05, 2016, 12:57:11 PM
The kid at the tunnel scenario phooey is presenting is actually a legitimate issue.  The issue being of liability and moral responsibility.  I'm not saying that it's a problem with self-driving cars but it is an issue that we have to resolve in order to program no-win scenarios into self-driving cars and that requires some ethical heavy lifting on designers' part.  The fact that these are no-win scenarios that humans also can't resolve makes it so it's not a technical challenge for deployment of self-driving cars, but it is nevertheless one of the key problems with automation of driving.

There will be accidents and there will be injuries and fatalities.  Undoubtedly there would be fewer of them with self-driving cars, but when they would happen how do we determine whether the car did the morally right thing?  If the choice is between running over a pedestrian or crashing the car and potentially seriously injuring passengers, what should the car do?  I don't think this is a simple problem.
Title: Re: Self-Driving Cars
Post by: PANTS! on April 05, 2016, 03:43:43 PM
The kid at the tunnel scenario phooey is presenting is actually a legitimate issue.  The issue being of liability and moral responsibility.  I'm not saying that it's a problem with self-driving cars but it is an issue that we have to resolve in order to program no-win scenarios into self-driving cars and that requires some ethical heavy lifting on designers' part.  The fact that these are no-win scenarios that humans also can't resolve makes it so it's not a technical challenge for deployment of self-driving cars, but it is nevertheless one of the key problems with automation of driving.

There will be accidents and there will be injuries and fatalities.  Undoubtedly there would be fewer of them with self-driving cars, but when they would happen how do we determine whether the car did the morally right thing?  If the choice is between running over a pedestrian or crashing the car and potentially seriously injuring passengers, what should the car do?  I don't think this is a simple problem.

It seems fairly simple to me.  The car should do everything in it's power to not hit the pedestrian, barring injuring its passengers or other cars.  There is such a concept as pedestrian liability.  And while they typically have the right of way, jumping into the road at a tunnel mouth should probably be treated as someone jumping onto train tracks.  In other words, a horrible situation, but not the train driver's (or car programmer's) fault.  You ought not to make perfect the enemy of vastly superior (unless you are Sanders, I guess).

Anyway - this has generally been solved for human drivers (http://www.alllaw.com/articles/nolo/auto-accident/driver-at-fault-pedestrian-car.html).  (See The Driver’s Duty of Care section for some illuminating facts).  I see no reason why a system that reduces accidents and protects pedestrians better should be substantially different on issues of legal liability. 

TL; DR - So long as "The Algorithm" follows what can reasonably be agreed upon as A Reasonable Driver’s Duty of Care, then "The Algorithm" shouldn't be held liable.  Same as a human driver.
Title: Re: Self-Driving Cars
Post by: teethering on April 05, 2016, 05:48:22 PM
The kid at the tunnel scenario phooey is presenting is actually a legitimate issue.  The issue being of liability and moral responsibility.  I'm not saying that it's a problem with self-driving cars but it is an issue that we have to resolve in order to program no-win scenarios into self-driving cars and that requires some ethical heavy lifting on designers' part.  The fact that these are no-win scenarios that humans also can't resolve makes it so it's not a technical challenge for deployment of self-driving cars, but it is nevertheless one of the key problems with automation of driving.

There will be accidents and there will be injuries and fatalities.  Undoubtedly there would be fewer of them with self-driving cars, but when they would happen how do we determine whether the car did the morally right thing?  If the choice is between running over a pedestrian or crashing the car and potentially seriously injuring passengers, what should the car do?  I don't think this is a simple problem.

It seems fairly simple to me.  The car should do everything in it's power to not hit the pedestrian, barring injuring its passengers or other cars.  There is such a concept as pedestrian liability.  And while they typically have the right of way, jumping into the road at a tunnel mouth should probably be treated as someone jumping onto train tracks.  In other words, a horrible situation, but not the train driver's (or car programmer's) fault.  You ought not to make perfect the enemy of vastly superior (unless you are Sanders, I guess).

Anyway - this has generally been solved for human drivers (http://www.alllaw.com/articles/nolo/auto-accident/driver-at-fault-pedestrian-car.html).  (See The Driver’s Duty of Care section for some illuminating facts).  I see no reason why a system that reduces accidents and protects pedestrians better should be substantially different on issues of legal liability. 

TL; DR - So long as "The Algorithm" follows what can reasonably be agreed upon as A Reasonable Driver’s Duty of Care, then "The Algorithm" shouldn't be held liable.  Same as a human driver.

The key is in the words "what can reasonably be agreed upon", which is something we generally don't need to do on behalf of an algorithm at the moment, we argue on behalf of human drivers.  It's a different kind of bar.  It's a bit like putting the decision on whether to pull a plug on someone into the hands of a predetermined policy, rather than in the hands of family members, it creates a difficult to resolve controversy because it's now a fixed policy.

Whenever we insert a policy where we expect humans to make moral judgments there's going to be a lot of difficulty trying to figure out what that policy should be.  A lot of lawyers are going to talk to the nerds who write this code.
Title: Re: Self-Driving Cars
Post by: brilligtove on April 05, 2016, 05:53:13 PM
The kid at the tunnel scenario phooey is presenting is actually a legitimate issue.  The issue being of liability and moral responsibility.  I'm not saying that it's a problem with self-driving cars but it is an issue that we have to resolve in order to program no-win scenarios into self-driving cars and that requires some ethical heavy lifting on designers' part.  The fact that these are no-win scenarios that humans also can't resolve makes it so it's not a technical challenge for deployment of self-driving cars, but it is nevertheless one of the key problems with automation of driving.

There will be accidents and there will be injuries and fatalities.  Undoubtedly there would be fewer of them with self-driving cars, but when they would happen how do we determine whether the car did the morally right thing?  If the choice is between running over a pedestrian or crashing the car and potentially seriously injuring passengers, what should the car do?  I don't think this is a simple problem.

This was answered quite nicely in one of the articles upthread: this problem only exists when we insist on asking the wrong question. It isn't that the car needs to do what is morally right - something humans can't reliably do either - but rather that the car does not do something wrong. (Note that wrong in this context means illegal, not "against Mohammad's teachings.")

That is the test that humans are put to when we encounter no win situations. We are not judged on doing the right thing, bit on not doing the wrong thing to the best of our abilities.

ETA Teethering does this cover your concern above?
Title: Re: Self-Driving Cars
Post by: teethering on April 05, 2016, 06:06:12 PM
The kid at the tunnel scenario phooey is presenting is actually a legitimate issue.  The issue being of liability and moral responsibility.  I'm not saying that it's a problem with self-driving cars but it is an issue that we have to resolve in order to program no-win scenarios into self-driving cars and that requires some ethical heavy lifting on designers' part.  The fact that these are no-win scenarios that humans also can't resolve makes it so it's not a technical challenge for deployment of self-driving cars, but it is nevertheless one of the key problems with automation of driving.

There will be accidents and there will be injuries and fatalities.  Undoubtedly there would be fewer of them with self-driving cars, but when they would happen how do we determine whether the car did the morally right thing?  If the choice is between running over a pedestrian or crashing the car and potentially seriously injuring passengers, what should the car do?  I don't think this is a simple problem.

This was answered quite nicely in one of the articles upthread: this problem only exists when we insist on asking the wrong question. It isn't that the car needs to do what is morally right - something humans can't reliably do either - but rather that the car does not do something wrong. (Note that wrong in this context means illegal, not "against Mohammad's teachings.")

That is the test that humans are put to when we encounter no win situations. We are not judged on doing the right thing, bit on not doing the wrong thing to the best of our abilities.

ETA Teethering does this cover your concern above?

Not really, because basically you're saying there won't be a controversy because we all always agree on what is "not wrong".  That is manifestly not true.  Which is why it would be difficult to set any policy down in code, because lots of people will necessarily find it morally wrong.

Again, this is not a technological obstacle, this isn't a problem with AI or difficulty in programming.  This is a difficulty in deciding what is morally right as a set of inflexible rules that a machine can follow.
Title: Re: Self-Driving Cars
Post by: brilligtove on April 05, 2016, 06:32:56 PM
[snip]
(Note that wrong in this context means illegal, not "against Mohammad's teachings.")
[snip]
ETA Teethering does this cover your concern above?

Not really, because basically you're saying there won't be a controversy because we all always agree on what is "not wrong".  That is manifestly not true.  Which is why it would be difficult to set any policy down in code, because lots of people will necessarily find it morally wrong.

Again, this is not a technological obstacle, this isn't a problem with AI or difficulty in programming.  This is a difficulty in deciding what is morally right as a set of inflexible rules that a machine can follow.

I think we have robust mechanisms for determining what is not legal. These are related to but not the same as what is not right. Laws are simplifications of ethical codes that can be enforced in a relatively universal and evenhanded way. We still need judges and lawyers and such because there is a lot of space to maneuver in "relatively" but even so.

We agree that this is not a technological issue at it's core. Society makes some allowances for trolley-car scenarios where there is no 'right' answer by asking if the person acted in a manner that was not illegal. Different scenarios have different measures of illegality and obligation, of course, but the responsibilities and obligations of drivers do have a lot of precident, as I understand it.
Title: Re: Self-Driving Cars
Post by: kvuo75 on April 05, 2016, 06:59:20 PM
in a no win situation just have the algorithm flip a coin then..
Title: Re: Self-Driving Cars
Post by: Fast Eddie B on April 05, 2016, 07:10:22 PM

The fastest human in the world will still hit the brakes a half a second slower than the car. To put that in perspective, at 35 miles per hour that car stops at least 51 feet faster than a human, or swerves, or whatever. More realistically, it's 100-125 feet, because most people can't react and begin braking with the same speed that a professional drag racer can punch the accelerator when the tree goes green. Some studies actually put the time to brake for drivers between 1.5 and 2.3 seconds, according to the NHTSA.


And it's not just reaction time.

I spent about a year as an accident investigator, and part of investigations was measuring skid marks.

Those skid marks would often stop well short of the Point Of Impact. People can feel so uncomfortable with extreme braking they can reflexively let up on the brakes, where continued braking could have stopped them short of the accident, or at the very lead to less speed at that point.

I assume algorithms would be great at "keeping their foot in it" right up to the Point Of Impact, certainly saving lives and minimizing injury.
Title: Re: Self-Driving Cars
Post by: daniel1948 on April 05, 2016, 07:15:00 PM
The kid at the tunnel scenario phooey is presenting is actually a legitimate issue.  The issue being of liability and moral responsibility.  I'm not saying that it's a problem with self-driving cars but it is an issue that we have to resolve in order to program no-win scenarios into self-driving cars and that requires some ethical heavy lifting on designers' part.  The fact that these are no-win scenarios that humans also can't resolve makes it so it's not a technical challenge for deployment of self-driving cars, but it is nevertheless one of the key problems with automation of driving.

There will be accidents and there will be injuries and fatalities.  Undoubtedly there would be fewer of them with self-driving cars, but when they would happen how do we determine whether the car did the morally right thing?  If the choice is between running over a pedestrian or crashing the car and potentially seriously injuring passengers, what should the car do?  I don't think this is a simple problem.

It seems fairly simple to me.  The car should do everything in it's power to not hit the pedestrian, barring injuring its passengers or other cars.  There is such a concept as pedestrian liability.  And while they typically have the right of way, jumping into the road at a tunnel mouth should probably be treated as someone jumping onto train tracks.  In other words, a horrible situation, but not the train driver's (or car programmer's) fault.  You ought not to make perfect the enemy of vastly superior (unless you are Sanders, I guess).

Anyway - this has generally been solved for human drivers (http://www.alllaw.com/articles/nolo/auto-accident/driver-at-fault-pedestrian-car.html).  (See The Driver’s Duty of Care section for some illuminating facts).  I see no reason why a system that reduces accidents and protects pedestrians better should be substantially different on issues of legal liability. 

TL; DR - So long as "The Algorithm" follows what can reasonably be agreed upon as A Reasonable Driver’s Duty of Care, then "The Algorithm" shouldn't be held liable.  Same as a human driver.

The key is in the words "what can reasonably be agreed upon", which is something we generally don't need to do on behalf of an algorithm at the moment, we argue on behalf of human drivers.  It's a different kind of bar.  It's a bit like putting the decision on whether to pull a plug on someone into the hands of a predetermined policy, rather than in the hands of family members, it creates a difficult to resolve controversy because it's now a fixed policy.

Whenever we insert a policy where we expect humans to make moral judgments there's going to be a lot of difficulty trying to figure out what that policy should be.  A lot of lawyers are going to talk to the nerds who write this code.

We should not lose sight of the fact that in a situation such as the tunnel dilemma above, no human driver is ever going to make a reasoned decision. It is ridiculous to argue against computer-controlled cars on the grounds that they'll be unable to make the "right" decision when there is no right decision, and it is actually idiotic to argue that computer-cars must be required to follow a higher standard than we set for human drivers; i.e. that the computer must be capable of rational decision-making in a situation where no human ever makes rational decisions.

The only thing that's relevant in the tunnel problem as it relates to computer-controlled cars is that the computer will be able to apply the brakes faster than a human driver and prevent anyone from being hurt in the much more common situation when the child jumps out a second or two earlier, too close for a human to react but with enough time for the computer to react.

Three scenarios:

1. The child jumps out too close to the car for even a computer to react. Someone gets hurt, either the child or the driver, or both. Doesn't matter whether a human or a computer is driving.

2. The child jumps out far enough ahead for a human to stop. An alert human avoids anybody's getting hurt. A computer avoids anybody getting hurt. A distracted human hits the child. Better to have a computer driving than a human, because humans can get distracted, but if the human driver is paying attention the outcome is the same as for a computer.

3. The child jumps out closer than (2) above, but not as close as (1). A human driver either hits the child or swerves and hurts himself. Possibly the human tries to swerve, loses control of the car, and both are hurt or killed. A computer, with its faster reaction time, avoids anybody's getting hurt.

The choice is clear: The computer is the better choice.

And if the next car back is computer controlled it will stop as quickly as the first car, avoiding a secondary collision.

Phooey is grabbing at straws in a desperate effort to construct a situation where a computer would fail. But these are situations where a human driver fails also. Meanwhile, the superior reaction time of a computer reduces accidents across the board. A computer is not a philosopher, to solve trolly dilemmas in real time. But even philosophers, when they are behind the wheel, do not perform a deep philosophical analysis in a split-second crisis decision.
Title: Re: Self-Driving Cars
Post by: teethering on April 05, 2016, 07:46:28 PM
[snip]
(Note that wrong in this context means illegal, not "against Mohammad's teachings.")
[snip]
ETA Teethering does this cover your concern above?

Not really, because basically you're saying there won't be a controversy because we all always agree on what is "not wrong".  That is manifestly not true.  Which is why it would be difficult to set any policy down in code, because lots of people will necessarily find it morally wrong.

Again, this is not a technological obstacle, this isn't a problem with AI or difficulty in programming.  This is a difficulty in deciding what is morally right as a set of inflexible rules that a machine can follow.

I think we have robust mechanisms for determining what is not legal. These are related to but not the same as what is not right. Laws are simplifications of ethical codes that can be enforced in a relatively universal and evenhanded way. We still need judges and lawyers and such because there is a lot of space to maneuver in "relatively" but even so.

We agree that this is not a technological issue at it's core. Society makes some allowances for trolley-car scenarios where there is no 'right' answer by asking if the person acted in a manner that was not illegal. Different scenarios have different measures of illegality and obligation, of course, but the responsibilities and obligations of drivers do have a lot of precident, as I understand it.

No, we do not have robust mechanisms for determining what is not legal, this is why our legal system can't be left up to computers and is decided by humans and it would be a monumental challenge to make it computable by machines.  We are talking about a very limited subset of the legal code here that does away with a lot of the complexity of litigation, but the core problem is still the same.  We haven't yet tried to automate moral or legal decisions when human lives are at stake on this kind of scale.  This is not a technically challenging issue, but it is a challenging issue nonetheless.

in a no win situation just have the algorithm flip a coin then..


Whatever the proposed solution is the mechanism for figuring out if the algorithm is right is no longer merely establishing its technical correctness, but also whether it is ethically right.  Now we code with some ethics-lite considerations all the time, but putting human lives in the hands of an algorithm on this scale of deployment is, to my knowledge, unprecedented.  It requires considerable non-technical work.

We should not lose sight of the fact that in a situation such as the tunnel dilemma above, no human driver is ever going to make a reasoned decision. It is ridiculous to argue against computer-controlled cars on the grounds that they'll be unable to make the "right" decision when there is no right decision, and it is actually idiotic to argue that computer-cars must be required to follow a higher standard than we set for human drivers; i.e. that the computer must be capable of rational decision-making in a situation where no human ever makes rational decisions.


I agree with the premise (i.e. human drivers don't get this right, so our bar should be set accordingly), I disagree with the conclusion.

The way to think of it is imagine instead of automating driving we're cloning one particular human driver and putting his clones in the driver's seat of lots and lots of cars.  Now imagine we strongly disagree with his decision making when it comes to things like the trolley dilemma.  This would be a big problem.  This is exactly equivalent to what we need to do with self-driving cars, we need to figure out what a good-enough human moral behaviour in these situations is and program that.  Settling on the "good-enough moral behaviour" is not a technical problem, it's an ethical one and computer engineers are possibly the worst kind of people (speaking as a computer engineer) to settle these sorts of questions because of the way we tend to think in breaking down problems (i.e. we tend to ignore or not be aware that non-utilitarian points of view exist).

I'm not saying phooey is right in the way he's pretending this problem is an insurmountable impediment in developing self-driving cars (which I think he's actually doing in a Gish-gallop-y kind of way of throwing all the difficulties at the wall and pretending they all magically sum up to "impossible"), I'm saying of the problems he's so far thrown up this one is legitimately difficult.
Title: Re: Self-Driving Cars
Post by: brilligtove on April 05, 2016, 08:13:07 PM
[snip]
(Note that wrong in this context means illegal, not "against Mohammad's teachings.")
[snip]
ETA Teethering does this cover your concern above?

Not really, because basically you're saying there won't be a controversy because we all always agree on what is "not wrong".  That is manifestly not true.  Which is why it would be difficult to set any policy down in code, because lots of people will necessarily find it morally wrong.

Again, this is not a technological obstacle, this isn't a problem with AI or difficulty in programming.  This is a difficulty in deciding what is morally right as a set of inflexible rules that a machine can follow.

I think we have robust mechanisms for determining what is not legal. These are related to but not the same as what is not right. Laws are simplifications of ethical codes that can be enforced in a relatively universal and evenhanded way. We still need judges and lawyers and such because there is a lot of space to maneuver in "relatively" but even so.

We agree that this is not a technological issue at it's core. Society makes some allowances for trolley-car scenarios where there is no 'right' answer by asking if the person acted in a manner that was not illegal. Different scenarios have different measures of illegality and obligation, of course, but the responsibilities and obligations of drivers do have a lot of precident, as I understand it.

No, we do not have robust mechanisms for determining what is not legal, this is why our legal system can't be left up to computers and is decided by humans and it would be a monumental challenge to make it computable by machines.  We are talking about a very limited subset of the legal code here that does away with a lot of the complexity of litigation, but the core problem is still the same.  We haven't yet tried to automate moral or legal decisions when human lives are at stake on this kind of scale.  This is not a technically challenging issue, but it is a challenging issue nonetheless.



Let's take this in parts, starting with what a robust mechanism for determining illegality means. It does not mean an algorithm programmed into a Vic20 (or supercomputer). It means we have a vast system of legislators, and a justice system, and an enforcement system. These are currently staffed almost completely by humans, with some computer enhanced capabilities. This vast (human) machinery applies it's collective 'wisdom' and 'intelligence' to judging whether any given action by a driver is illegal or not. These systems have significant problems and faults, but I think 'robust' is a reasonable term to use to describe them.


in a no win situation just have the algorithm flip a coin then..


Whatever the proposed solution is the mechanism for figuring out if the algorithm is right is no longer merely establishing its technical correctness, but also whether it is ethically right.  Now we code with some ethics-lite considerations all the time, but putting human lives in the hands of an algorithm on this scale of deployment is, to my knowledge, unprecedented.  It requires considerable non-technical work.

Since we disagree on whether the appropriate premise is that SDVs should be measured against 'right' or vs 'illegal' I don't really have a comment, except that randomness is often directly incorporated into machine learning systems. This renders them non-deterministic in a practical sense.

We should not lose sight of the fact that in a situation such as the tunnel dilemma above, no human driver is ever going to make a reasoned decision. It is ridiculous to argue against computer-controlled cars on the grounds that they'll be unable to make the "right" decision when there is no right decision, and it is actually idiotic to argue that computer-cars must be required to follow a higher standard than we set for human drivers; i.e. that the computer must be capable of rational decision-making in a situation where no human ever makes rational decisions.


I agree with the premise (i.e. human drivers don't get this right, so our bar should be set accordingly), I disagree with the conclusion.

The way to think of it is imagine instead of automating driving we're cloning one particular human driver and putting his clones in the driver's seat of lots and lots of cars.  Now imagine we strongly disagree with his decision making when it comes to things like the trolley dilemma.  This would be a big problem.  This is exactly equivalent to what we need to do with self-driving cars, we need to figure out what a good-enough human moral behaviour in these situations is and program that.  Settling on the "good-enough moral behaviour" is not a technical problem, it's an ethical one and computer engineers are possibly the worst kind of people (speaking as a computer engineer) to settle these sorts of questions because of the way we tend to think in breaking down problems (i.e. we tend to ignore or not be aware that non-utilitarian points of view exist).

I'm not saying phooey is right in the way he's pretending this problem is an insurmountable impediment in developing self-driving cars (which I think he's actually doing in a Gish-gallop-y kind of way of throwing all the difficulties at the wall and pretending they all magically sum up to "impossible"), I'm saying of the problems he's so far thrown up this one is legitimately difficult.

Meet Gav. Gav has two billion clones, and they are all identical - transporter duplicates, not vat grown and trained. The Gavs are now the only people licensed to drive motor vehicles.

Gavs like to mow down old ladies at crosswalks. This is illegal. Gavs are punished for this behaviour when they exhibit it.

Gavs like to behave in a way that many people disagree with when confronted with a trolley problem. If this behaviour is likely illegal, the current system of enforcement catches Gavs that exhibit this behaviour . The justice system evaluates these actions case by case, perhaps based on precedent, perhaps setting precedent. If it's determined to be legal behaviour, the Gavs are free to continue driving. If not, no more license for the Gav that broke the law. Now if the legislators disagree with the courts, they can change the law to allow or disallow the thing the Gavs did.

At this point there is no difference between normal human drivers and our SDV/Gavs.

If enforcers, the judiciary, and legislators all agree that a behaviour of the Gavs condition [trolley] is wrong, we use the same magic that created the Gavs to replace them with Gav Mark II, who do not make that decision that way.

That's new, and an advantage in my books: we can rapidly upgrade the behaviours and decision making capacities of all the SDVs everywhere.

Title: Re: Self-Driving Cars
Post by: phooey on April 05, 2016, 11:21:46 PM
Yes right, I am "grabbing at straws", "gish-galloping", creating "contrived, ridiculous scenarios"...

Yes, just like the MIT engineers and reporters who actually wrote the articles! 

This is predictably like the last time you all tried to find every rationalizing in the book to say that an article that I claimed says driverless cars could make traffic much worse, (which is titled Driverless Cars Could Make Traffic Dramatically Worse) doesn't claim that driverless cars could make traffic worse.

When one is willing to argue as disingenuously as that, how can you take them seriously? 

I mean, even when I showed that the leading figures in the technology believe that humans and self driving cars can't mix together well, there are people here that still want to argue that point by saying, "Oh, what are you talking about, you are nuts, they are doing it already, what a bunch of nonsense, yadda, yadda...." Its laughable, that's Google and Elon Musk saying those things, not me, but you still try your bullshit levels of disagreement.   

Argue with the authors of the articles all you want, I don't care.  "The article doesn't mean what you say it means!" even when I don't say what it means, I just show you.  This is a microcosm of the skeptical argument techniques used throughout the internet.  You all use the same playbook. 

Its laughable.   
Title: Re: Self-Driving Cars
Post by: phooey on April 05, 2016, 11:31:55 PM
Quote
" Driving in the United States is actually incredibly safe, with fatal crashes occurring once every roughly 3 million hours of driving. Driverless vehicles will need to be even safer than that, Shladover said.

Given existing software, "that is amazingly difficult to do," he said.   That's because no software in laptops, phones or other modern devices is designed to operate for extended periods without freezing, crashing or dropping a call — and similar errors would be deadly in a car. Right now, Google's self-driving cars avoid this by having both a backup driver and a second person as a monitor, who can shut off the system at the first hint of a glitch. But coming up with safety-critical, fail-safe software for completely driverless cars would require reimagining how software is designed, Shladover said.

"There is no current process to efficiently develop safe software," Shladover said.

Send your hate mail to Steven Shladover, at the Partners for Advanced Transportation Technology at the University of California, Berkeley. 

He is an idiot, he is gish-galloping, he is making shit up, this problem has already been addressed, livescience magazine just likes to write sensationalizing articles, the source is not credible, the study is flawed!, he doesn't mean what he is saying he means....


http://www.livescience.com/50841-future-of-driverless-cars.html
Title: Re: Self-Driving Cars
Post by: phooey on April 05, 2016, 11:36:07 PM
"But Urmson admits that Google's autonomous car won't recognize a pothole in the road—or worse, an open manhole—unless it's marked off with traffic cones. Yeesh."

http://gizmodo.com/6-simple-things-googles-self-driving-car-still-cant-han-1628040470


They don't even have a solution for potholes right now.  Ho ho ho. 

Be careful with your coffee during your commute there, skeptiqueer. 
Title: Re: Self-Driving Cars
Post by: SkeptiQueer on April 06, 2016, 04:15:48 AM
"But Urmson admits that Google's autonomous car won't recognize a pothole in the road—or worse, an open manhole—unless it's marked off with traffic cones. Yeesh."

http://gizmodo.com/6-simple-things-googles-self-driving-car-still-cant-han-1628040470


They don't even have a solution for potholes right now.  Ho ho ho. 

Be careful with your coffee during your commute there, skeptiqueer.
I'm confused as to how "it's not doable right this second" translates to "not doable for 200 years."

I'll happily seconded an email to Berkeley to explain that my car doesn't run 24/7, but rather shuts down when I park, usually for a number of hours.

How many drivers will recognize an open manhole cover during a routine commute?
Title: Re: Self-Driving Cars
Post by: AtheistApotheosis on April 06, 2016, 08:07:21 AM
If a self-driving car gets in an accident, who is to blame?  The person behind the wheel?  The software developer?  The vehicle manufacturer?  Until those problems are solved, almost definitely by legislative intervention, widescale adoption is unlikely.

Don't existing liability laws covering this possibility. If a manufacturing fault results in a fatal accident, the liability comes back to the manufacturer. The manufacturer also may be required to repair the fault free of charge to the owner or the vehicle will be recalled and the owner compensated. existing laws shouldn't be any different. If the accident is the fault of the human driver then the current laws would most likely still apply. Accidents are going to happen, in fact there have already been accidents involving self driving cars. Fortunately nothing fatal, but it's almost certainly going to happen. At least until all vehicles become self driving. Progress is rarely without risk, but usually the benefits outweigh the risks. Only time will tell.
Title: Re: Self-Driving Cars
Post by: SkeptiQueer on April 06, 2016, 08:48:36 AM
If a self-driving car gets in an accident, who is to blame?  The person behind the wheel?  The software developer?  The vehicle manufacturer?  Until those problems are solved, almost definitely by legislative intervention, widescale adoption is unlikely.

Don't existing liability laws covering this possibility. If a manufacturing fault results in a fatal accident, the liability comes back to the manufacturer. The manufacturer also may be required to repair the fault free of charge to the owner or the vehicle will be recalled and the owner compensated. existing laws shouldn't be any different. If the accident is the fault of the human driver then the current laws would most likely still apply. Accidents are going to happen, in fact there have already been accidents involving self driving cars. Fortunately nothing fatal, but it's almost certainly going to happen. At least until all vehicles become self driving. Progress is rarely without risk, but usually the benefits outweigh the risks. Only time will tell.

So, manufacturer's won't want to take liability. The solution is either no-fault insurance, or building limited-liability into the DOT regulations.
Title: Re: Self-Driving Cars
Post by: AtheistApotheosis on April 06, 2016, 09:27:12 AM
If a self-driving car gets in an accident, who is to blame?  The person behind the wheel?  The software developer?  The vehicle manufacturer?  Until those problems are solved, almost definitely by legislative intervention, widescale adoption is unlikely.

Don't existing liability laws covering this possibility. If a manufacturing fault results in a fatal accident, the liability comes back to the manufacturer. The manufacturer also may be required to repair the fault free of charge to the owner or the vehicle will be recalled and the owner compensated. existing laws shouldn't be any different. If the accident is the fault of the human driver then the current laws would most likely still apply. Accidents are going to happen, in fact there have already been accidents involving self driving cars. Fortunately nothing fatal, but it's almost certainly going to happen. At least until all vehicles become self driving. Progress is rarely without risk, but usually the benefits outweigh the risks. Only time will tell.

So, manufacturer's won't want to take liability. The solution is either no-fault insurance, or building limited-liability into the DOT regulations.

Manufacturers never want to take liability for their products. No-fault insurance and limited liability could reduce the incentive to improve the technology and make it safer. It happens in other industries. If the car manufacturer is too risk averse they may simply choose not to invest in the technology and lose out to the competition. Do we have to compromise on safety, if so how far do we go. We certainly don't with the airline industry.
Title: Re: Self-Driving Cars
Post by: SkeptiQueer on April 06, 2016, 09:36:44 AM
If a self-driving car gets in an accident, who is to blame?  The person behind the wheel?  The software developer?  The vehicle manufacturer?  Until those problems are solved, almost definitely by legislative intervention, widescale adoption is unlikely.

Don't existing liability laws covering this possibility. If a manufacturing fault results in a fatal accident, the liability comes back to the manufacturer. The manufacturer also may be required to repair the fault free of charge to the owner or the vehicle will be recalled and the owner compensated. existing laws shouldn't be any different. If the accident is the fault of the human driver then the current laws would most likely still apply. Accidents are going to happen, in fact there have already been accidents involving self driving cars. Fortunately nothing fatal, but it's almost certainly going to happen. At least until all vehicles become self driving. Progress is rarely without risk, but usually the benefits outweigh the risks. Only time will tell.

So, manufacturer's won't want to take liability. The solution is either no-fault insurance, or building limited-liability into the DOT regulations.

Manufacturers never want to take liability for their products. No-fault insurance and limited liability could reduce the incentive to improve the technology and make it safer. It happens in other industries. If the car manufacturer is too risk averse they may simply choose not to invest in the technology and lose out to the competition. Do we have to compromise on safety, if so how far do we go. We certainly don't with the airline industry.

By limited liability, I mean limiting the scope of what's considered to be manufacturer's responsibility, otherwise every incident is the manufacturer's fault, even if the vehicle didn't do anything wrong but did the best it could under the circumstances.
Title: Re: Self-Driving Cars
Post by: brilligtove on April 06, 2016, 09:39:35 AM
If a self-driving car gets in an accident, who is to blame?  The person behind the wheel?  The software developer?  The vehicle manufacturer?  Until those problems are solved, almost definitely by legislative intervention, widescale adoption is unlikely.

Don't existing liability laws covering this possibility. If a manufacturing fault results in a fatal accident, the liability comes back to the manufacturer. The manufacturer also may be required to repair the fault free of charge to the owner or the vehicle will be recalled and the owner compensated. existing laws shouldn't be any different. If the accident is the fault of the human driver then the current laws would most likely still apply. Accidents are going to happen, in fact there have already been accidents involving self driving cars. Fortunately nothing fatal, but it's almost certainly going to happen. At least until all vehicles become self driving. Progress is rarely without risk, but usually the benefits outweigh the risks. Only time will tell.

So, manufacturer's won't want to take liability. The solution is either no-fault insurance, or building limited-liability into the DOT regulations.

Manufacturers never want to take liability for their products. No-fault insurance and limited liability could reduce the incentive to improve the technology and make it safer. It happens in other industries. If the car manufacturer is too risk averse they may simply choose not to invest in the technology and lose out to the competition. Do we have to compromise on safety, if so how far do we go. We certainly don't with the airline industry.

I think the manufacturers are going to have to take on some liability - especially the software and the sensor/controls tech. I say this because I explicitly do not own the software I purchase (at least I don't today). If I did I could make copies of it, modify it, etc.. Assuming Google and Tesla are selling me a licence and not the software itself, it's going to be hard to pin responsibility on me as the "owner" of an SDV.

Yes right, I am "grabbing at straws", "gish-galloping", creating "contrived, ridiculous scenarios"

It seems Phoobot may have recognized what we've been telling it for some time. Good for it's coders.
Title: Re: Self-Driving Cars
Post by: SkeptiQueer on April 06, 2016, 09:51:45 AM
If a self-driving car gets in an accident, who is to blame?  The person behind the wheel?  The software developer?  The vehicle manufacturer?  Until those problems are solved, almost definitely by legislative intervention, widescale adoption is unlikely.

Don't existing liability laws covering this possibility. If a manufacturing fault results in a fatal accident, the liability comes back to the manufacturer. The manufacturer also may be required to repair the fault free of charge to the owner or the vehicle will be recalled and the owner compensated. existing laws shouldn't be any different. If the accident is the fault of the human driver then the current laws would most likely still apply. Accidents are going to happen, in fact there have already been accidents involving self driving cars. Fortunately nothing fatal, but it's almost certainly going to happen. At least until all vehicles become self driving. Progress is rarely without risk, but usually the benefits outweigh the risks. Only time will tell.

So, manufacturer's won't want to take liability. The solution is either no-fault insurance, or building limited-liability into the DOT regulations.

Manufacturers never want to take liability for their products. No-fault insurance and limited liability could reduce the incentive to improve the technology and make it safer. It happens in other industries. If the car manufacturer is too risk averse they may simply choose not to invest in the technology and lose out to the competition. Do we have to compromise on safety, if so how far do we go. We certainly don't with the airline industry.

I think the manufacturers are going to have to take on some liability - especially the software and the sensor/controls tech. I say this because I explicitly do not own the software I purchase (at least I don't today). If I did I could make copies of it, modify it, etc.. Assuming Google and Tesla are selling me a licence and not the software itself, it's going to be hard to pin responsibility on me as the "owner" of an SDV.

Yes right, I am "grabbing at straws", "gish-galloping", creating "contrived, ridiculous scenarios"

It seems Phoobot may have recognized what we've been telling it for some time. Good for it's coders.

I had completely forgotten that angle. End-user modification would clearly shift liability, so I'm sure we'll have to have DOT sit down with a bunch of lawyers and nerds and suss out how much that is.
Title: Re: Self-Driving Cars
Post by: daniel1948 on April 06, 2016, 10:08:24 AM
We should not lose sight of the fact that in a situation such as the tunnel dilemma above, no human driver is ever going to make a reasoned decision. It is ridiculous to argue against computer-controlled cars on the grounds that they'll be unable to make the "right" decision when there is no right decision, and it is actually idiotic to argue that computer-cars must be required to follow a higher standard than we set for human drivers; i.e. that the computer must be capable of rational decision-making in a situation where no human ever makes rational decisions.


I agree with the premise (i.e. human drivers don't get this right, so our bar should be set accordingly), I disagree with the conclusion.

The way to think of it is imagine instead of automating driving we're cloning one particular human driver and putting his clones in the driver's seat of lots and lots of cars.  Now imagine we strongly disagree with his decision making when it comes to things like the trolley dilemma.  This would be a big problem.  This is exactly equivalent to what we need to do with self-driving cars, we need to figure out what a good-enough human moral behaviour in these situations is and program that.  Settling on the "good-enough moral behaviour" is not a technical problem, it's an ethical one and computer engineers are possibly the worst kind of people (speaking as a computer engineer) to settle these sorts of questions because of the way we tend to think in breaking down problems (i.e. we tend to ignore or not be aware that non-utilitarian points of view exist).

I'm not saying phooey is right in the way he's pretending this problem is an insurmountable impediment in developing self-driving cars (which I think he's actually doing in a Gish-gallop-y kind of way of throwing all the difficulties at the wall and pretending they all magically sum up to "impossible"), I'm saying of the problems he's so far thrown up this one is legitimately difficult.

Computers cannot and will not make moral decisions. Computer programmers are unlikely to ever be able to program a computer to make moral decisions. What programmers and computers will be able to do is calculate likely outcomes, with very large error margins, and execute braking and/or avoidance maneuvers in a fraction of an eye-blink. The result will be that when accidents occur there will be much hand-wringing over the computer's decision, but accidents will happen far less often, and when they do happen the injuries will be less severe and there will be fewer deaths. We will be far better off, but we'll have to accept that human "morality" factors are often arbitrary and serve no useful purpose in the overall goal of reducing injury and death.

As for "Gav the driver" and his clones, computers will not all be running the same program. The first computer-controlled car will be safer for its occupants and for the drivers of nearby cars and for pedestrians than any but the most skilled race-trained human drivers, and with each car model using different software, car makers and programmers will compare the driving behaviors and outcomes of the different programs and will make improvements.

One of the reasons I want a Tesla Model 3 is for the optional auto-pilot features. It won't be a self-driving car yet, but it will make lane-changing safer and will make it easier to follow the flow of traffic. I.e., it will have a few of the advantages of a self-driving car.
Title: Re: Self-Driving Cars
Post by: brilligtove on April 06, 2016, 10:36:25 AM
We should not lose sight of the fact that in a situation such as the tunnel dilemma above, no human driver is ever going to make a reasoned decision. It is ridiculous to argue against computer-controlled cars on the grounds that they'll be unable to make the "right" decision when there is no right decision, and it is actually idiotic to argue that computer-cars must be required to follow a higher standard than we set for human drivers; i.e. that the computer must be capable of rational decision-making in a situation where no human ever makes rational decisions.


I agree with the premise (i.e. human drivers don't get this right, so our bar should be set accordingly), I disagree with the conclusion.

The way to think of it is imagine instead of automating driving we're cloning one particular human driver and putting his clones in the driver's seat of lots and lots of cars.  Now imagine we strongly disagree with his decision making when it comes to things like the trolley dilemma.  This would be a big problem.  This is exactly equivalent to what we need to do with self-driving cars, we need to figure out what a good-enough human moral behaviour in these situations is and program that.  Settling on the "good-enough moral behaviour" is not a technical problem, it's an ethical one and computer engineers are possibly the worst kind of people (speaking as a computer engineer) to settle these sorts of questions because of the way we tend to think in breaking down problems (i.e. we tend to ignore or not be aware that non-utilitarian points of view exist).

I'm not saying phooey is right in the way he's pretending this problem is an insurmountable impediment in developing self-driving cars (which I think he's actually doing in a Gish-gallop-y kind of way of throwing all the difficulties at the wall and pretending they all magically sum up to "impossible"), I'm saying of the problems he's so far thrown up this one is legitimately difficult.

Computers cannot and will not make moral decisions. Computer programmers are unlikely to ever be able to program a computer to make moral decisions. What programmers and computers will be able to do is calculate likely outcomes, with very large error margins, and execute braking and/or avoidance maneuvers in a fraction of an eye-blink. The result will be that when accidents occur there will be much hand-wringing over the computer's decision, but accidents will happen far less often, and when they do happen the injuries will be less severe and there will be fewer deaths. We will be far better off, but we'll have to accept that human "morality" factors are often arbitrary and serve no useful purpose in the overall goal of reducing injury and death.

As for "Gav the driver" and his clones, computers will not all be running the same program. The first computer-controlled car will be safer for its occupants and for the drivers of nearby cars and for pedestrians than any but the most skilled race-trained human drivers, and with each car model using different software, car makers and programmers will compare the driving behaviors and outcomes of the different programs and will make improvements.

One of the reasons I want a Tesla Model 3 is for the optional auto-pilot features. It won't be a self-driving car yet, but it will make lane-changing safer and will make it easier to follow the flow of traffic. I.e., it will have a few of the advantages of a self-driving car.

Yeah, I don't expect there to be only one SDV brain program for all vehicles - but that was the frame of the thought experiment so I went with it. I expect there will be several distinct autopilot models from different companies. Also, as you go higher in price you might get autopilots that have more training and computational power. Still, I expect there will only be a few dozen core autopilots, with carmakers regularly upgrading the entire fleet with improved safety and performance enhancements. (And fixes and patches, of course.)
Title: Re: Self-Driving Cars
Post by: Jeremy's Sea on April 06, 2016, 12:02:53 PM
When one is willing to argue as disingenuously as that, how can you take them seriously? 
The fact anyone still engages you is shocking. You heavily cherry pick those industry leaders who agree with your assessment and you ignore those who think it's feasible and then scream that everyone else ignores industry leaders amidst a serious discussion about challenges and concerns. I really hoe you're awesome220, I'd really hate to think there's 2 of him out there...
Title: Re: Self-Driving Cars
Post by: phooey on April 06, 2016, 12:44:49 PM
Oh really Jeremy?  Its cherry picking to point out article that agree with my assessment? 

If that's cherry picking dude, then EVERY adversarial discussion is by definition cherry picking, because each side is presenting their side of the discussion.  That is a preposterous use of the term "cherry picking". 

Its like someone showing articles that supposedly support the idea of Darwinian Natural Selection (ok, obviously there are no such articles, but hypothetically, for this analogy), and then someone responding, "Well, that's cherry picking, why aren't you showing articles from people who don't believe in natural selection!" 

That is a patently stupid premise.   
Title: Re: Self-Driving Cars
Post by: SkeptiQueer on April 06, 2016, 01:13:59 PM
Oh really Jeremy?  Its cherry picking to point out article that agree with my assessment? 

If that's cherry picking dude, then EVERY adversarial discussion is by definition cherry picking, because each side is presenting their side of the discussion.  That is a preposterous use of the term "cherry picking". 

Its like someone showing articles that supposedly support the idea of Darwinian Natural Selection (ok, obviously there are no such articles, but hypothetically, for this analogy), and then someone responding, "Well, that's cherry picking, why aren't you showing articles from people who don't believe in natural selection!" 

That is a patently stupid premise.

I'll ask again, how does someone saying "twenty years" support your assertion that it will take two-hundred?

It's cherry picking to take one person and decide that because they agree with you, everyone else is wrong. If there was a consensus of experts, it would be a different ballgame, but there's not. Further, none of the experts agree with your two-century predictions, and the most conservative outliers are two decades with the most generous hype being two years.

It's cherry picking when you pick out one person, decide that's right, and then ignore or dodge any challenges and refuse to consider any dissent as remotely valid or worthy of consideration. Make sense?
Title: Re: Self-Driving Cars
Post by: phooey on April 06, 2016, 01:30:27 PM
Do you mean the way you are considering dissent, skeptiqueer? 

And Elon Musk is the most conservative outlier?  Very funny. 

How about the people who are predicting it will never happen, what are they? 

Wait, you are cherry picking. 
Title: Re: Self-Driving Cars
Post by: teethering on April 06, 2016, 01:46:33 PM
[snip]
(Note that wrong in this context means illegal, not "against Mohammad's teachings.")
[snip]
ETA Teethering does this cover your concern above?

Not really, because basically you're saying there won't be a controversy because we all always agree on what is "not wrong".  That is manifestly not true.  Which is why it would be difficult to set any policy down in code, because lots of people will necessarily find it morally wrong.

Again, this is not a technological obstacle, this isn't a problem with AI or difficulty in programming.  This is a difficulty in deciding what is morally right as a set of inflexible rules that a machine can follow.

I think we have robust mechanisms for determining what is not legal. These are related to but not the same as what is not right. Laws are simplifications of ethical codes that can be enforced in a relatively universal and evenhanded way. We still need judges and lawyers and such because there is a lot of space to maneuver in "relatively" but even so.

We agree that this is not a technological issue at it's core. Society makes some allowances for trolley-car scenarios where there is no 'right' answer by asking if the person acted in a manner that was not illegal. Different scenarios have different measures of illegality and obligation, of course, but the responsibilities and obligations of drivers do have a lot of precident, as I understand it.

No, we do not have robust mechanisms for determining what is not legal, this is why our legal system can't be left up to computers and is decided by humans and it would be a monumental challenge to make it computable by machines.  We are talking about a very limited subset of the legal code here that does away with a lot of the complexity of litigation, but the core problem is still the same.  We haven't yet tried to automate moral or legal decisions when human lives are at stake on this kind of scale.  This is not a technically challenging issue, but it is a challenging issue nonetheless.



Let's take this in parts, starting with what a robust mechanism for determining illegality means. It does not mean an algorithm programmed into a Vic20 (or supercomputer). It means we have a vast system of legislators, and a justice system, and an enforcement system. These are currently staffed almost completely by humans, with some computer enhanced capabilities. This vast (human) machinery applies it's collective 'wisdom' and 'intelligence' to judging whether any given action by a driver is illegal or not. These systems have significant problems and faults, but I think 'robust' is a reasonable term to use to describe them.

Their robustness is inherent in the "human driver" of the legal system.  In other words, we rely on various humans to establish what's "right" and what's "wrong" post factum.  This isn't available for programming behaviour a priori.  We can program "something" and then see how it stacks up and then adjust it when it fucks up.  The problem with this approach, however, is that the stakes are really really high, both legal liabilities and human lives at risk.  So if you're going to deploy this en masse no one will take this approach, no sane company will take those kinds of risks, their lawyers won't let them and for actually pretty legitimate reasons.

My only point is that this is legitimately difficult, as opposed to the rest of the bullshit phooey throws up.
Title: Re: Self-Driving Cars
Post by: PANTS! on April 06, 2016, 02:11:23 PM
Eh....  I think there will be a way to help with liability.  Like the special Vax court.  Or how we resolve workman's comp.
Title: Re: Self-Driving Cars
Post by: Jeremy's Sea on April 06, 2016, 02:35:20 PM
It's cherry picking to take one person and decide that because they agree with you, everyone else is wrong. If there was a consensus of experts, it would be a different ballgame, but there's not. Further, none of the experts agree with your two-century predictions, and the most conservative outliers are two decades with the most generous hype being two years.

It's cherry picking when you pick out one person, decide that's right, and then ignore or dodge any challenges and refuse to consider any dissent as remotely valid or worthy of consideration. Make sense?
This, exactly. For the [self proclaimed] One True Skeptic on this forum, you need to better understand the ground rules.
ETA - referring to phooey, not referring to SkeptiQueer
Title: Re: Self-Driving Cars
Post by: daniel1948 on April 06, 2016, 03:08:42 PM
One time out of a hundred thousand the renegade turns out to be right. This gives renegades everywhere complete and utter confidence that they are right and everyone else is wrong. And justifies them in citing the outlier in support of their position. What's kind of amusing, in a pathetic sort of way, is how Phooey cites Elon Musk saying it will take 20 years for a new technology to completely replace the existing fleet, asserts that Musk's predictions are always off by an order of magnitude, since the three cars they've built had some delays at the beginning, and concludes that Musk is supporting his (Phooey's) assertion that self-driving cars are 200 years away. We shouldn't be too hard on Phooey, though. Youth often suffers from delusions of omniscience.

Note: Any other company would have sold defective cars on schedule. Tesla Motors prefers to miss its goal dates and deliver cars that meet the highest standards. This leads some folks who are desperate to find something to say against EVs or Tesla, to cite these delays in their neo-Luddite arguments. The delays show commitment to quality, and do justify taking projected dates with a grain of salt. The 18-month promise for the Model 3 could well be 2 years. Not the 15 years that Phooey's creative math would suggest.
Title: Re: Self-Driving Cars
Post by: brilligtove on April 06, 2016, 06:07:53 PM
Teethering, you say that we don't have any a priori way to program a sense of "right" and "wrong" behaviours into an SDV. This may be true in an abstract philosophical sense, but it is not true in practice. Our societies have already defined what is legal and illegal as a proxy for what is right and wrong, and SDVs are being trained (and programmed) to comply to those laws. We have also defined what is acceptable and not acceptable driving conventions outside of the legal and regulatory requirements. Much of the existing approach to judging the performance of these vehicles is based on the SDV operating in a manner that is within specified behavioural parameters. This is also how we judge the performance of human drivers.
Title: Re: Self-Driving Cars
Post by: arthwollipot on April 06, 2016, 07:14:22 PM
This was answered quite nicely in one of the articles upthread...

I'm willing to lay bets that phooey didn't read that article. He's still making many of the arguments that were debunked - specifically Misconceptions 2, 3, 4 and 6.
Title: Re: Self-Driving Cars
Post by: teethering on April 06, 2016, 07:16:43 PM
Teethering, you say that we don't have any a priori way to program a sense of "right" and "wrong" behaviours into an SDV. This may be true in an abstract philosophical sense, but it is not true in practice. Our societies have already defined what is legal and illegal as a proxy for what is right and wrong, and SDVs are being trained (and programmed) to comply to those laws. We have also defined what is acceptable and not acceptable driving conventions outside of the legal and regulatory requirements. Much of the existing approach to judging the performance of these vehicles is based on the SDV operating in a manner that is within specified behavioural parameters. This is also how we judge the performance of human drivers.

But this isn't true.  We judge the performance of human drivers not via an algorithm, but via a system made up of humans.  That is a key crucial difference.  We do not replace this evaluation with a blind computer calculation of rules.  So the effort to actually codify this into a program, that is to have the driving algorithm reflect not just the rules written, but how the legal system behaves, which is a lot of different evaluations performed by human actors in an actual legal system that we currently have, that's a complicated and challenging endeavor.

I am not a lawyer, but I'm pretty sure that it is incorrect to presume that you can encode written laws alone in an algorithm that would determine if something is legal or not.  That is not how our legal system functions, nor is it necessarily that such an algorithm would even be coherent, because the human actors in our legal system smooth out a lot of its contradictions and ambiguities.
Title: Re: Self-Driving Cars
Post by: phooey on April 06, 2016, 07:29:01 PM
arthwollipot,

Well, you think that we are going to be able to eliminate traffic lights, and just have pedestrians walk into the middle of the street without looking, and the computer programs will just adjust traffic, swerve, brake or whatever to avoid the people walking out. 

So I am willing to bet, you couldn't understand the articles, even if you did try to read them. 
Title: Re: Self-Driving Cars
Post by: arthwollipot on April 06, 2016, 08:12:43 PM
arthwollipot,

Well, you think that we are going to be able to eliminate traffic lights, and just have pedestrians walk into the middle of the street without looking, and the computer programs will just adjust traffic, swerve, brake or whatever to avoid the people walking out. 

So I am willing to bet, you couldn't understand the articles, even if you did try to read them.

I'm willing to bet that you didn't notice that when I mentioned those things I was referencing a fictional novel, so there's that.
Title: Re: Self-Driving Cars
Post by: DG on April 06, 2016, 08:43:50 PM
Moderator Comment
... and things were going so well.

Given the general tenor of in this thread, I would suggest that all comments be directed at ideas and not at individuals.

If at any time you are inclined to comment in any way refers to an individual (or group of individuals), ask yourself this: Is the point of this post to reference the other person's contribution, or to reference the contributor? If the latter, please reconsider the necessity of the post at all.

Please, discuss ideas and reconsider the posting of anything that could be considered insulting or bickering, or that is otherwise unnecessarily disruptive to the exchange of ideas relevant to the topic.

We continue to monitor this thread.


For clarity:

The above is not a revision of the rules, but a simple attempt to encourage constructive discussion.
Title: Re: Self-Driving Cars
Post by: Morvis13 on April 06, 2016, 09:26:23 PM
On my commute today I was thinking how frustrating it would be to have a computer driving the speed limit while everyone else on the road is easily doing 10-20 kph over. I have read this is also one drawback of current robot drivers. I also suspect once robots take over then speed limits can go away. I wish the future would hurry up and get here.
Title: Re: Self-Driving Cars
Post by: DG on April 06, 2016, 09:39:44 PM
On my commute today I was thinking how frustrating it would be to have a computer driving the speed limit while everyone else on the road is easily doing 10-20 kph over. I have read this is also one drawback of current robot drivers. I also suspect once robots take over then speed limits can go away. I wish the future would hurry up and get here.

Somewhat ironically, I have the same concern, but my frustration is that they are doing so without incurring sanctions. Your concern could be achieved by having the SDC report the number plate of any vehicle passing that does so in excess of the speed limit. Problem solved ;)
Title: Re: Self-Driving Cars
Post by: Morvis13 on April 06, 2016, 10:21:08 PM
On my commute today I was thinking how frustrating it would be to have a computer driving the speed limit while everyone else on the road is easily doing 10-20 kph over. I have read this is also one drawback of current robot drivers. I also suspect once robots take over then speed limits can go away. I wish the future would hurry up and get here.

Somewhat ironically, I have the same concern, but my frustration is that they are doing so without incurring sanctions. Your concern could be achieved by having the SDC report the number plate of any vehicle passing that does so in excess of the speed limit. Problem solved ;)

No, no, no. have the robot go with the flow of traffic or give it a setting for +10 over the limit.
Title: Re: Self-Driving Cars
Post by: DG on April 06, 2016, 10:22:15 PM
Then why not just increase the speed limit by 10?
Title: Re: Self-Driving Cars
Post by: Fast Eddie B on April 06, 2016, 11:01:38 PM
"But Urmson admits that Google's autonomous car won't recognize a pothole in the road—or worse, an open manhole—unless it's marked off with traffic cones. Yeesh."


Sorry, I don't see why that's not easily solvable with software.

A human will have his retina tickled by something dark in the road, try to process whether its a shadow or an oil stain or a pothole or an open manhole or an object in the road, make a decision as to what evasive action to take, if any, and then deal with an interminable reaction time. Most times at highway speed a driver will simply take no corrective action and run over/into whatever it was.

I have zero doubt an intelligent car could be programmed to process the information and act on it far more quickly and appropriately than a human driver. Either directly through algorithms or through "learning".

It's going to happen, the only question is when.
Title: Re: Self-Driving Cars
Post by: daniel1948 on April 06, 2016, 11:06:10 PM
I think maybe it was George Carlin (?) who said something to the effect of "Anyone who drives slower than me is as asshole and anyone who drives faster than me is an idiot."

The almost universal contempt for speed limits among human drivers is an issue we will need to address with computer-controlled cars. At present in most jurisdictions the speed laws are not strictly enforced. A positive step would be to re-assess speed limits, decide what is to be allowed, and then strictly enforce that limit. Then computer-controlled cars and human-driven cars would not be in conflict. Otherwise programmers could write programs to have the car go with the flow of traffic, or to strictly obey the speed limit. Some folks will be unhappy with either choice, but the problem is not with computer-controlled cars, it's with the widespread contempt for speed limits and the concomitant lack of enforcement by the authorities.

If the cops are not going to ticket cars until they're going 5 mph over the limit, that should be clearly stated. If it's 10 over, that should be clearly stated. Then computer programs could drive accordingly. Disrespect for the laws is a reason to prohibit humans from driving cars, not a reason to oppose computer-driven cars.
Title: Re: Self-Driving Cars
Post by: Fast Eddie B on April 06, 2016, 11:18:28 PM
I think maybe it was George Carlin (?) who said something to the effect of "Anyone who drives slower than me is as asshole and anyone who drives faster than me is an idiot."

I remember that, but think slower=idiot and faster=maniac.

Which is very true, of course!
Title: Re: Self-Driving Cars
Post by: phooey on April 07, 2016, 12:26:57 AM
Daniel,

Trying to enforce the speed limit to the exact speed is not practical for any number of reasons.  First, whose speedometer or radar is most accurate?  The cars?  The police?  Some camera? 

Second, when you are driving, not only can you not be staring at the speedometer for every second of your drive, but even if you could, think how hard it would be to keep the car at exactly the speed you want. Plus there are hills, bumps, etc.  So how fast BELOW the current speed limit would you want drivers to realistically go? 10?  15 below the limit, so as to make sure to NEVER go 1 mile an hour over the limit?  So on a 55 mph speedway, you want traffic averaging 45? or 40? And on a 25 mph zone, maybe they would need to go 15, just to be sure. 

Fortunately for now, in other countries they are not as speed ticket obsessed as in America, so you can drive in a reasonable manner without worrying that a cop is waiting for you around every bush, treating everyone who wants to go to the grocery store for milk as a feared criminal. The America model is a failed one when it comes to this.  No wonder everyone hates the police.   
Title: Re: Self-Driving Cars
Post by: DG on April 07, 2016, 01:48:02 AM
Daniel,

Trying to enforce the speed limit to the exact speed is not practical for any number of reasons.  First, whose speedometer or radar is most accurate?  The cars?  The police?  Some camera? 

Accuracy is not at point, the issue is convincing the court that one has evidence that the person was speeding (this is the threshold for offences relating to the driving a vehicle at a speed in excess of the signposted speed. Ordinarily it is not done by reference to the 'exact speed', but by reference to a range i.e "person exceeded the speed limit by not more than 15kph", or "person exceeded the speed limit by at least 15kph, but not more than 30kph".

While a police officer may tell you "You were doing 65 in a 60" - the actual offence is for exceeding the speed limit by an amount greater than permitted by law (I note that some jurisdictions allow for a margin of error).

In many jurisdictions there is a rebuttable presumption that the speed camera of the police is evidence of the speed that the vehicle was travelling (with a margin of error).

The situation identified is not an issue in a legal sense, the situation is presently dealt with by the courts.

*it is also worth noting that, at least in this jurisdiction, for a car to be road worthy the speedometer must provide a reading of the speed of the vehicle that is accurate to "with 10% of, and not less than the actual speed of the vehicle" - that's the margin of error. If you speed reports that you are doing 95kph when travelling at 97kph, your vehicle is not (under current law) road worthy - the penalties for driving such a vehicle far outweigh the penalties for speeding, so I certainly wouldn't want to rely on that as a defence. 

Quote

Second, when you are driving, not only can you not be staring at the speedometer for every second of your drive, but even if you could, think how hard it would be to keep the car at exactly the speed you want. Plus there are hills, bumps, etc.  So how fast BELOW the current speed limit would you want drivers to realistically go? 10?  15 below the limit, so as to make sure to NEVER go 1 mile an hour over the limit?  So on a 55 mph speedway, you want traffic averaging 45? or 40? And on a 25 mph zone, maybe they would need to go 15, just to be sure. 


One need not keep the vehicle at a particular speed, They must not exceed a posted speed.

The intent of the post was to advocate compliance with the law. I was somewhat ironically suggesting that, no matter the signposted speed, people will travel over that speed because, well, they can do so (largely without consequence). The reason being that they rely on the fact that they drive with disregard for the posted speed limit. Electing to apply their own judgement instead.

As for your examples, yes, I am far more comfortable sitting behind a learner driver travelling 30% below the posted speed than I am being passed by a person who displays a disregard for the speed limit. Don't even get me started on the "but everyone does it" defence.

If driving slowly is an issue then the solution is to legislate a minimum speed (something along the lines of: a person not must drive at less than 50% of the posted speed unless, (a) the driver is entering or exiting the road, (b) the driver is travelling at a speed approximately equal to that of the vehicle immediately in front of them, or (c) it is unsafe to drive at a speed that is greater than 50% of the posted speed limit).

Quote

Fortunately for now, in other countries they are not as speed ticket obsessed as in America, so you can drive in a reasonable manner without worrying that a cop is waiting for you around every bush, treating everyone who wants to go to the grocery store for milk as a feared criminal. The America model is a failed one when it comes to this.  No wonder everyone hates the police.

I'm not from the USA, or anywhere in the Americas.

I'm actually not aware of any country that does not (at least ostensibly) have speed limits, or issue tickets for breaches of the applicable road rules\driving codes. I would be interested to know which country does not have such a law. I understand it was the case the the Northern Territory did (at least for a while) have unrestricted speeds in the outback, however I understand that has been revised to a maximum speed of 130kph. Equally, I understand that the German roads that were famously "unrestricted", now have maximum speed limits.

While different jurisdictions have 'informal' tolerance for driver error, I am concerned that this invites corruption and discriminatory application of the laws (i.e a police officer may elect to give a ticket to the rich, middle aged man from the ethnic majority driving a Ferrari at 5kph over the limit, but let the apparently disadvantaged young person from an ethnic minority off with a warning for travelling at 35kph over the limit in his/her rust bucket).

My point is, whatever the speed limit, it should be enforced, or abandoned. Consistency in the application and enforcement of speed limits may, while unpopular, address (at least in part) the problem referred to above where those who want to speed are being "frustrated" by those who are complying with their legal obligation to drive at or below the posted speed limit.
Title: Re: Self-Driving Cars
Post by: phooey on April 07, 2016, 03:17:58 AM
Dg,

I am not so clear what you are saying.  First, You think people should be ticketed for going one mile an hour over the speed limit?  And then fight it out in court over the accuracy of the ticket? 

Second, I never suggested there were other countries without speed limits.  What I said was that most other places are not as obsessed with writing as many tickets as possible, just because someone is innocently trying to get somewhere at a reasonable speed.  Its a question of how strict and how worthwhile it is to regulate the people of a community to the nth degree.  Where I live, you might get a ticket from a camera if you are really exceeding the limit, but it is not something you have to necessarily be cautious of every time you get into your car.  Which is actually a good thing for the society, and I will explain why.

Where I live, people don't hate the police, they actually like them for the most part.  They are treated respectfully in general, and most people's interaction with the law is not all about someone trying to hit them with more fines based on bullshit speed limits, which may or may not be reasonable for the road conditions. Speeding ten miles and hour over a speed limit is not the biggest threat to a society.  But you wouldn't know that in America, where cops treat people like captured servants, asking them to bow in front of the police for their terrible indiscretions ("Keep your hands on the wheel, don't turn around! Where are you going this evening?").   This is not a good thing for society in any way shape or form.  I would argue it is this constant hassling, and bureaucratic nonsense that causes a great deal of the strive that occurs in American cities.  Think of the guy who is just trying to drive home from his late shift at the auto factory.  He gets pulled over because he didn't notice that he is going through a supposed school zone where the speed suddenly dips extra low, and now he has to pay an extra $200 for a ticket.  He already has a late tax bill he is trying to deal with, the college where his daughter is going just raised their fees, and just last week he got a $100 parking ticket.  On top of that the officer is trying to lecture him about obeying authority.  The guy doesn't really want to hear this from the cop, and asked if he can give him a break, but instead the cop says, you know, I don't like your attitude, I am going to go ahead and give you another ticket for your upper brake light not working properly.  Maybe that will teach you some manners next time...

This is what America is like in a nutshell.  If you think the biggest problem facing American cities is people going 5 miles an hour over the speed limit, I can tell you that is way wrong.  Giving people back some of their dignity, maybe a little less in the way of constant regulations, and using some common sense instead of coercion at the threat of fines and other punishment goes a long way to making a better society, not more government control. 

 

Title: Re: Self-Driving Cars
Post by: DG on April 07, 2016, 03:46:41 AM
Dg,

I am not so clear what you are saying.  First, You think people should be ticketed for going one mile an hour over the speed limit?  And then fight it out in court over the accuracy of the ticket? 
[\quote]

No.

Quote
Second, I never suggested there were other countries without speed limits.  What I said was that most other places are not as obsessed with writing as many tickets as possible, just because someone is innocently trying to get somewhere at a reasonable speed.  Its a question of how strict and how worthwhile it is to regulate the people of a community to the nth degree.  Where I live, you might get a ticket from a camera if you are really exceeding the limit, but it is not something you have to necessarily be cautious of every time you get into your car.  Which is actually a good thing for the society, and I will explain why.

Where I live, people don't hate the police, they actually like them for the most part.  They are treated respectfully in general, and most people's interaction with the law is not all about someone trying to hit them with more fines based on bullshit speed limits, which may or may not be reasonable for the road conditions. Speeding ten miles and hour over a speed limit is not the biggest threat to a society.  But you wouldn't know that in America, where cops treat people like captured servants, asking them to bow in front of the police for their terrible indiscretions ("Keep your hands on the wheel, don't turn around! Where are you going this evening?").   This is not a good thing for society in any way shape or form.  I would argue it is this constant hassling, and bureaucratic nonsense that causes a great deal of the strive that occurs in American cities.  Think of the guy who is just trying to drive home from his late shift at the auto factory.  He gets pulled over because he didn't notice that he is going through a supposed school zone where the speed suddenly dips extra low, and now he has to pay an extra $200 for a ticket.  He already has a late tax bill he is trying to deal with, the college where his daughter is going just raised their fees, and just last week he got a $100 parking ticket.  On top of that the officer is trying to lecture him about obeying authority.  The guy doesn't really want to hear this from the cop, and asked if he can give him a break, but instead the cop says, you know, I don't like your attitude, I am going to go ahead and give you another ticket for your upper brake light not working properly.  Maybe that will teach you some manners next time...


I am not in the USA, nor purporting to speak about the USA, nor any particular jurisdiction.

I agree that the issue with the cop complaining about the drivers attitude is a problem.

 I would suggests that the solution is less discretion for the officer, not more. By removing the officer's discretion (in this instance), the improper use of power is denied.

As mentioned, discretion is demonstrably used in a discriminatory way - however I fear we are straying from the topic of SDV and speed limits in to a social commentary on law enforcement. Again I suspect we agree that law enforcement can be 'variable' depending on the "mood" of the officer, and we would agree that this is harmful to the institution of law enforcement, respect for that instution, and ongoing compliance with the law. However, again, this is off the topic of SDV. 

Quote

This is what America is like in a nutshell.  If you think the biggest problem facing American cities is people going 5 miles an hour over the speed limit, I can tell you that is way wrong.  Giving people back some of their dignity, maybe a little less in the way of constant regulations, and using some common sense instead of coercion at the threat of fines and other punishment goes a long way to making a better society, not more government control.

What, in your estimation, is the role of a speed limit?
Title: Re: Self-Driving Cars
Post by: Alex Simmons on April 07, 2016, 07:04:54 AM
https://www.reddit.com/r/gifs/comments/4dejx0/for_my_phd_i_teach_robots_how_to_avoid_obstacles/

One can only wonder at how much better machines will ultimately be at avoiding collisions and reducing collision rates dramatically.
Title: Re: Self-Driving Cars
Post by: Fast Eddie B on April 07, 2016, 07:50:12 AM
Fortunately for now, in other countries they are not as speed ticket obsessed as in America, so you can drive in a reasonable manner without worrying that a cop is waiting for you around every bush, treating everyone who wants to go to the grocery store for milk as a feared criminal. The America model is a failed one when it comes to this.  No wonder everyone hates the police.

Last time I was in Canada, they seemed at least as "speed ticket obsessed" as the U.S.

Unless that's changed.
Title: Re: Self-Driving Cars
Post by: phooey on April 07, 2016, 07:59:22 AM
DG,

I am still not clear what you are suggesting.  So 1 mile an hour over the speed limit?  2? 

A camera on every street? 
Title: Re: Self-Driving Cars
Post by: Morvis13 on April 07, 2016, 08:02:02 AM
When you live in 'Merica there is no other country.

It has been a while since I was on the Autobahn in Germany but that was fun. Speed limits are no more than a community agreed upon safe velocity in a given area. On a long straight enclosed road with controlled entrance and no pedestrians the limit is higher than a city street near a school. Road conditions also effect speed since traffic moves faster on bright sunny days than during a winter snow storm.

I still think robot could do a better job than humans and once we have enough self driving cars we can remove speed limits and let the computer decide on the safe velocity. For now we force the car to obey the posted limit.
Title: Re: Self-Driving Cars
Post by: SkeptiQueer on April 07, 2016, 08:15:35 AM
DG,

I am still not clear what you are suggesting.  So 1 mile an hour over the speed limit?  2? 

A camera on every street?

How many over is allowable? 5? If so, why not just raise the speed limit by 5mph?

Speed limits are supposed to reflect the safe speed. Much like BAC the true safe limit varies by person, but as we don't keep a full competency profile on each person and keep each driver RFID tagged and have their license plate reflect their relative safe speeds, the actionable alternative is that you enforce a flat maximum speed for everyone, and you do it as best you can.


Can you come up with a reason why you need to be allowed to exceed the speed limit?
Title: Re: Self-Driving Cars
Post by: daniel1948 on April 07, 2016, 08:17:38 AM
I recall, some decades ago, somebody visiting Mexico and getting arrested for possession of marijuana, and complaining bitterly about the condition of Mexican jails. My thought was, when you visit a foreign country you obey their laws or you pay the price. If you don't like their laws or how they enforce them, you can stay home.

The whole point of speed limits is you are supposed to go no faster than the limit. You may go slower. You are supposed to drive as much below the limit as you must in order to ensure that you do not exceed the limit. Fooey's argument, that you should not be required to know the speed you are driving, is equivalent to arguing that there should be no speed limit at all.

There is a very real purpose to speed limits: They save lives, if enforced and obeyed. I knew the sheriff of a rural Oregon town. He regularly had to go to accident sites where there were dead and injured people. What he saw was something physics also tells us: energy increases as the square of speed, and the energy of a crash, transferred to a soft, non-elastic body, turns that body into splatter.

This is relevant to this thread because obeying the speed limit saves lives. The purpose of speeding tickets is to convince drivers to obey the law. And the advantage of computer-controlled cars is that they can be programmed to obey the speed limits. If this annoys human drivers (or passengers) they should lobby for higher speed limits, not complain about the computer or the little old lady who actually obey the law, or the cop who enforces it.

Sadly, in our very racist country, cops often harass and abuse people of color, but this has nothing to do with the issue of speed limits. And Fooey's story of being harassed by a cop at a speeding stop has nothing to do with speed limits or self-driving cars and therefore does not even belong in this thread. It belongs in the politics forum in a thread about police misconduct.
Title: Re: Self-Driving Cars
Post by: phooey on April 07, 2016, 08:20:46 AM
https://www.reddit.com/r/gifs/comments/4dejx0/for_my_phd_i_teach_robots_how_to_avoid_obstacles/

One can only wonder at how much better machines will ultimately be at avoiding collisions and reducing collision rates dramatically.

Interesting video.  As most of the commenters mentioned, I wonder why he never tries to actually hit it. 
Title: Re: Self-Driving Cars
Post by: SkeptiQueer on April 07, 2016, 08:35:57 AM
https://www.reddit.com/r/gifs/comments/4dejx0/for_my_phd_i_teach_robots_how_to_avoid_obstacles/

One can only wonder at how much better machines will ultimately be at avoiding collisions and reducing collision rates dramatically.

Interesting video.  As most of the commenters mentioned, I wonder why he never tries to actually hit it.

Physics, for one. He, planted on the ground and using his body as leverage, can accelerate that foil far faster than the quadcopter can move. In fact, he can swing it faster than the quadcopter can fall out of the air, much less laterally accelerate. The purpose of the exercise is not for it to be a training drone a la Star Wars but to teach the software what an approaching object looks like. It's akin to playing catch with a toddler; you don't try to smoke the kid with a fastball, you're just helping the kid's brain and eyes understand what the ball looks like as it flies. When I was teaching martial arts to kids I never went full-contact and full-speed, as their software was still learning what an incoming punch looked like.
Title: Re: Self-Driving Cars
Post by: Morvis13 on April 07, 2016, 09:19:13 AM
https://www.reddit.com/r/gifs/comments/4dejx0/for_my_phd_i_teach_robots_how_to_avoid_obstacles/

One can only wonder at how much better machines will ultimately be at avoiding collisions and reducing collision rates dramatically.

Interesting video.  As most of the commenters mentioned, I wonder why he never tries to actually hit it.

Lets see, you are a starving university student that spent hundreds if not thousands of dollars and many, many hours making this robot and you're a couple days into testing. Here is a good idea. Try to destroy the whole thing.

Now we need the jedi training sphere.
Title: Re: Self-Driving Cars
Post by: Guillermo on April 07, 2016, 09:57:05 AM
On my commute today I was thinking how frustrating it would be to have a computer driving the speed limit while everyone else on the road is easily doing 10-20 kph over. I have read this is also one drawback of current robot drivers. I also suspect once robots take over then speed limits can go away. I wish the future would hurry up and get here.

Somewhat ironically, I have the same concern, but my frustration is that they are doing so without incurring sanctions. Your concern could be achieved by having the SDC report the number plate of any vehicle passing that does so in excess of the speed limit. Problem solved ;)
I have no frustration. I'd be reading a book or playing my DS.
Title: Re: Self-Driving Cars
Post by: brilligtove on April 07, 2016, 11:05:45 AM
Teethering, you say that we don't have any a priori way to program a sense of "right" and "wrong" behaviours into an SDV. This may be true in an abstract philosophical sense, but it is not true in practice. Our societies have already defined what is legal and illegal as a proxy for what is right and wrong, and SDVs are being trained (and programmed) to comply to those laws. We have also defined what is acceptable and not acceptable driving conventions outside of the legal and regulatory requirements. Much of the existing approach to judging the performance of these vehicles is based on the SDV operating in a manner that is within specified behavioural parameters. This is also how we judge the performance of human drivers.

But this isn't true.  We judge the performance of human drivers not via an algorithm, but via a system made up of humans.  That is a key crucial difference.  We do not replace this evaluation with a blind computer calculation of rules.  So the effort to actually codify this into a program, that is to have the driving algorithm reflect not just the rules written, but how the legal system behaves, which is a lot of different evaluations performed by human actors in an actual legal system that we currently have, that's a complicated and challenging endeavor.

I am not a lawyer, but I'm pretty sure that it is incorrect to presume that you can encode written laws alone in an algorithm that would determine if something is legal or not.  That is not how our legal system functions, nor is it necessarily that such an algorithm would even be coherent, because the human actors in our legal system smooth out a lot of its contradictions and ambiguities.

I think our disconnect lies in programming vs training. Broadly speaking, SDVs are not programmed in the traditional sense: they are trained like an animal is trained.

In that paradigm, the trainers are not just coding if then rules, like "IF [speed] >= [speed_limit - 5kph] THEN Speed(decelerate);". They're training the embodied AI to navigate based on goals (a destination, occupant safety, collision avoidance, etc.).

The questions of what is right and wrong are defined by the trainers, and the nAI learns to work within those boundaries.
Title: Re: Self-Driving Cars
Post by: phooey on April 07, 2016, 12:26:43 PM
"Can you come up with a reason why you need to be allowed to exceed the speed limit?"

I just gave some reasons, did you not even bother to read?

So you want speeding tickets for one mile an hour over the speed limit.  Do you want cameras on every street as well then? 

They just started giving out tickets in North Carolina for 1 mile an hour over.  People don't like it.  Its not very good having policies that people don't like.  It doesn't make for a very content and harmonious society.  Maybe its one of the reasons why Americans are not nearly as happy as people in other countries. 

So how about it, cameras on every single street? 
Title: Re: Self-Driving Cars
Post by: daniel1948 on April 07, 2016, 12:43:22 PM
Sounds like what you want is for people to be allowed to break the law with impunity. Remember, those speed limit laws were passed by elected officials. Sure, some people don't like those laws. But I'll bet a lot of people are exceedingly happy to see the reckless speeders ticketed.

And if a lot of people are speeding, then I say, sure: speed cameras everywhere. There is no presumption of privacy when you are on a public road operating a six-thousand-pound death machine. Driving is a privilege, not a right. Obey the law, or change the law, or lose the privilege.
Title: Re: Self-Driving Cars
Post by: phooey on April 07, 2016, 12:49:33 PM
Yea man, go after all those reckless soccer mom's going 46 in 45 mph zones.

You skeptics are some amazing thinkers. 
Title: Re: Self-Driving Cars
Post by: phooey on April 07, 2016, 12:54:53 PM
Better still, I say give them jail time.  That ought to make them think next time. 

"Sorry honey, mommy won't be around to make you lunch this month.  I am a criminal and I need to pay for my lawless ways.  Don't worry dear, its only a medium security prison, since its only my second offense. Its mostly just with petty drug dealers." 
Title: Re: Self-Driving Cars
Post by: SkeptiQueer on April 07, 2016, 01:32:08 PM
"Can you come up with a reason why you need to be allowed to exceed the speed limit?"

I just gave some reasons, did you not even bother to read?

So you want speeding tickets for one mile an hour over the speed limit.  Do you want cameras on every street as well then? 

They just started giving out tickets in North Carolina for 1 mile an hour over.  People don't like it.  Its not very good having policies that people don't like.  It doesn't make for a very content and harmonious society.  Maybe its one of the reasons why Americans are not nearly as happy as people in other countries. 

So how about it, cameras on every single street?

So again, why should people be allowed to go faster if they can tell their robot car to go the speed limit?

If not going one-over is too hard, aim for one-under. It's a maximum, not a required speed.
Title: Re: Self-Driving Cars
Post by: phooey on April 07, 2016, 01:36:54 PM
https://www.washingtonpost.com/news/true-crime/wp/2016/04/07/after-fairfax-police-ticket-cars-awaiting-state-inspection-va-passes-law-banning-it/


"The story of Bruce Redwine and the Chantilly Service Center made people mad. For years, he and other auto shop owners watched with outrage as a Fairfax County parking enforcement officer ticketed cars that were awaiting state inspection or repair. One shop owner estimated his customers had been hit with $60,000 worth of fines and fees for expired inspection stickers or tags over six years. Redwine got so angry he snatched one ticket out of the parking officer’s hand, only to be charged with felony assault on a police officer."



http://jalopnik.com/north-carolina-is-going-to-be-a-real-jerk-about-speedin-1766936094


"As part of Gov. McCrory’s multi-headed plan to make the once-great state of North Carolina suck just a little bit more, this Friday will mark the start of the North Carolina Department of Transportation’s new campaign, Obey the Sign or Pay the Fine. Uuuuuggghhh....Thanks, Gov. McCrory. Remember, nobody likes you."



Do these make you proud to be an American Daniel? 
Title: Re: Self-Driving Cars
Post by: phooey on April 07, 2016, 01:39:20 PM
Yea I gotcha skeptiqueer, I see your vision for the world. 

Put em in jail.  Also, they should start writing more tickets for not putting on turn signals 150 yards before a turn, or not keeping the proper number of cars lengths away from the car in front of you.  Crack down, that will make America and even better place! 

Skeptics rule!  Cameras on every street!!
Title: Re: Self-Driving Cars
Post by: SkeptiQueer on April 07, 2016, 01:41:40 PM
You're doing a red herring. Stop it. If you have a problem with enforcement or punishment, address that in a separate thread instead of shitting up this one.

So again, why do the self-driving cars need to be capable of exceeding the speed limit on a day to day basis? Why do people need to exceed the speed limit on a day to day basis?
Title: Re: Self-Driving Cars
Post by: phooey on April 07, 2016, 01:47:33 PM
Its not a red herring.  Don't tell me how to reply. 

You, along with Daniel and DG have stated what you see as a reasonable vision for how traffic violations should be handled.  You want as much enforcement as possible, you want as many cameras as possible.  You want strict following of the letter of the law, because to you that is a good America.  A Strong America!  USA, USA, USA!

People here should be allowed to judge the reasonableness of the skeptic mindset here.   
Title: Re: Self-Driving Cars
Post by: Morvis13 on April 07, 2016, 02:05:04 PM
You're doing a red herring. Stop it. If you have a problem with enforcement or punishment, address that in a separate thread instead of shitting up this one.

So again, why do the self-driving cars need to be capable of exceeding the speed limit on a day to day basis? Why do people need to exceed the speed limit on a day to day basis?

This kinda my fault as I said I wanted my robot to speed. I know some people don't like having a slow driver in front of them. Right now the robots are programmed to obey the speed limits and people are not. If it is a bright sunny day on a highway then why can't you speed? As mentioned the limits are there for safety. Mostly because human drivers have terrible reaction times and poor judgement of the vehicle handling at high speeds. Robots would be better at this so speed limits should increase as automated driving systems gain experience and the erratic humans vacate the roads.

As for reasons to speed. I can think of many. Passing, merging into traffic, Terrible zoning, Lack of other traffic, Increased traffic flow, Avoiding red lights, Avoiding accidents and going around slower traffic like buses and trucks.

I do not think cops should ticket for 1 mile over. I would fight that right on the side of the road. Then again I'm likely to have set my cruise control so I won't have to constantly watch the speedometer and I will be doing the limit when out of my home area.
Title: Re: Self-Driving Cars
Post by: phooey on April 07, 2016, 02:08:16 PM
Scofflaw!
Title: Re: Self-Driving Cars
Post by: SkeptiQueer on April 07, 2016, 02:10:24 PM
Accelerating to make a light, passing, and merging are all frequent causes of accidents. If people stop trying to get that tiny edge in traffic, we'd have fewer accidents. People think they're better drivers than they are, so they push the limits.

I don't want to get into the logistics of enforcement in this thread. If there's a case to be made for increasing the speed limit in an area, that's got to be on a pretty specific basis, but I don't accept that people should view cars following the speed limit and driving safely as a hazard. Y'all are the hazard, and you need to clam your tots and start planning in some buffer time for traffic instead of trying to cut around in traffic.

If you'll excuse me, I just got a massive erection imagining the RAGE of people who will be stuck in a car that won't speed or cut people off.

Sent from my SM-G530T using Tapatalk

Title: Re: Self-Driving Cars
Post by: phooey on April 07, 2016, 02:22:52 PM
Skeptiqueers  (and Daniel and DG's) solution to a better transportation model is duly noted.  More tickets, more cameras, more fines, more punishment.  The law is the law, get tough on crime.  If you go one mile an hour over the speed limit, you are a criminal. 

This will make the world a better place.  Oh, and don't forget to carry a glock. 
Title: Re: Self-Driving Cars
Post by: daniel1948 on April 07, 2016, 05:01:19 PM
The late and very much lamented Tom Magliozzi often advocated a 35 mph speed limit. He said that people are in too much of a hurry. In Thomas Jefferson's day, nearly all transportation moved at about three mph.

I repeat: The laws in this country are made by citizens who have been elected by other citizens. If you don't like the speed limit, change it. And if you are not a citizen, then you're a guest when you are here and should obey the rules of the house.

I must admit that I don't understand why Phooey, who said he does not live here, has his knickers in such a twist over our speed limit laws and their enforcement, or what it really has to do with the issue of computer-controlled cars, which can be programmed to drive any way the company instructs the programmer to do. Once we actually have computer-controlled cars that we can buy, that will be the time to discuss how we want them to set their velocity. But since that can easily be set however we like, it's absolutely irrelevant to any argument about whether or not to have these cars.
Title: Re: Self-Driving Cars
Post by: daniel1948 on April 07, 2016, 05:02:31 PM
... and don't forget to carry a glock. 

Where the fuck did that come from?
Title: Re: Self-Driving Cars
Post by: arthwollipot on April 07, 2016, 08:52:19 PM
Daniel,

Trying to enforce the speed limit to the exact speed is not practical for any number of reasons.  First, whose speedometer or radar is most accurate?  The cars?  The police?  Some camera? 

Second, when you are driving, not only can you not be staring at the speedometer for every second of your drive, but even if you could, think how hard it would be to keep the car at exactly the speed you want. Plus there are hills, bumps, etc.  So how fast BELOW the current speed limit would you want drivers to realistically go? 10?  15 below the limit, so as to make sure to NEVER go 1 mile an hour over the limit?  So on a 55 mph speedway, you want traffic averaging 45? or 40? And on a 25 mph zone, maybe they would need to go 15, just to be sure. 

Fortunately for now, in other countries they are not as speed ticket obsessed as in America, so you can drive in a reasonable manner without worrying that a cop is waiting for you around every bush, treating everyone who wants to go to the grocery store for milk as a feared criminal. The America model is a failed one when it comes to this.  No wonder everyone hates the police.

Oddly enough, I am a professional driver. In my city, the vast majority of drivers drive at the speed limit. They have little difficulty maintaining the correct speed, and there is absolutely zero problem with comparing their speedos with the fixed or mobile traffic cameras that are frequently seen on our roads. They seem to be able to pay attention to traffic while simultaneously maintaining their speed - presumably they do so the same way I do, by occasional glances at the speedo that do not distract from paying attention to the road.

Your objections are utterly spurious. And I say that from my position as a professional driver - one who will be out of a job when the Robot Car Revolution arrives.
Title: Re: Self-Driving Cars
Post by: arthwollipot on April 07, 2016, 08:57:23 PM
https://www.reddit.com/r/gifs/comments/4dejx0/for_my_phd_i_teach_robots_how_to_avoid_obstacles/

One can only wonder at how much better machines will ultimately be at avoiding collisions and reducing collision rates dramatically.

Interesting video.  As most of the commenters mentioned, I wonder why he never tries to actually hit it.

He is trying to hit it. He's trying to hit it fencing-style, which has a particular technique that he is demonstrating quite well. You'd have to know a few things about fencing to realise that, though. :)

I wonder how it'd go with a cut-and-thrust sword. When he approaches it from a low guard, it seems to overreact, which could be a problem.
Title: Re: Self-Driving Cars
Post by: Alex Simmons on April 07, 2016, 09:36:41 PM
One guy doing a PhD can make a flying object to learn and respond very quickly to avoid a fencing foil. That's really the point. Add in a shite load of $ and programming resource and one can imagine what's possible with a vehicle that essentially operates in 2 dimensions, layered onto already pretty sophisticated self steering, braking and throttle control systems. Always vigilant, always with optimal response time.
Title: Re: Self-Driving Cars
Post by: arthwollipot on April 07, 2016, 09:38:39 PM
And more importantly, does anyone seriously think that this drone was programmed to react using a simple series of nested if-then statements?
Title: Re: Self-Driving Cars
Post by: Andrew Clunn on April 07, 2016, 10:21:08 PM
And more importantly, does anyone seriously think that this drone was programmed to react using a simple series of nested if-then statements?

No dawg.  It be a case statement caught up in some iterative looping shit.
Title: Re: Self-Driving Cars
Post by: Alex Simmons on April 08, 2016, 01:51:17 AM
Anyway, there is always this story about fleets of self driving trucks making successful long journeys across Europe:

http://www.iflscience.com/technology/six-platoons-self-driving-trucks-just-drove-thousands-kilometers-across-europe

But no, self driving vehicles could never work  ::)
Title: Re: Self-Driving Cars
Post by: daniel1948 on April 08, 2016, 08:09:02 AM
Anyway, there is always this story about fleets of self driving trucks making successful long journeys across Europe:

http://www.iflscience.com/technology/six-platoons-self-driving-trucks-just-drove-thousands-kilometers-across-europe

But no, self driving vehicles could never work  ::)

Wow. That is totally amazing! Yep, the future is here even earlier than I thought. Thanks for posting that link.
Title: Re: Self-Driving Cars
Post by: brilligtove on April 08, 2016, 11:26:57 AM
You beat me to it! I had the link ready and everything. :)
Title: Re: Self-Driving Cars
Post by: phooey on April 08, 2016, 11:44:25 AM
Well, this really is amazing!  They were able to self-drive these trucks on closed highways, with well marked European roads, driving in formation so as not to allow other cars to intervene in their chain, and with professionals drivers monitoring the trucks and taking control during difficult sections.  I can't believe it! 

Next thing you know we are going to have airplanes with autopilot! 

I just hope no one accuses you of cherry picking.   
Title: Re: Self-Driving Cars
Post by: Johnny Slick on April 08, 2016, 11:52:06 AM
Guys, the reason this worked is simple. There are no space ninjas in Europe. How you so-called "skeptics" failed to understand that simple fact proves that I am the only true skeptic here. ::)
Title: Re: Self-Driving Cars
Post by: bimble on April 08, 2016, 05:03:32 PM
Well, this really is amazing!  They were able to self-drive these trucks on closed highways,   

Where does it say the roads were closed?? I can't think of why, or even how, they would close thousands of kilometers of highway across several countries to run a test to see if the trucks could complete the trip.
Title: Re: Self-Driving Cars
Post by: Johnny Slick on April 08, 2016, 05:07:01 PM
Well, this really is amazing!  They were able to self-drive these trucks on closed highways,   

Where does it say the roads were closed?? I can't think of why, or even how, they would close thousands of kilometers of highway across several countries to run a test to see if the trucks could complete the trip.
It was a secret plan, dumas, clearly designed to further the agendas of Big Automated Car. ::)
Title: Re: Self-Driving Cars
Post by: Alex Simmons on April 08, 2016, 05:47:23 PM
Well, this really is amazing!  They were able to self-drive these trucks on closed highways, with well marked European roads, driving in formation so as not to allow other cars to intervene in their chain, and with professionals drivers monitoring the trucks and taking control during difficult sections.  I can't believe it! 

Next thing you know we are going to have airplanes with autopilot! 

I just hope no one accuses you of cherry picking.   

These were not closed roads. It was done under normal traffic conditions.

https://youtu.be/G-xOEJb4LEU?t=783

And did you expect that such a test at this early stage of development would be permitted without human fail safe in place? One cannot expect perfection instantly - but you must admit this is pretty impressive and it runs counter to much of what you've been saying. This stuff is possible and will happen. And it'll make the biggest difference to commercial transport to begin with.

This is the project's website BTW:
https://www.eutruckplatooning.com/home/default.aspx

Note the partner's page - it's a collaboration of transportation interests from governments, industry bodies, research institutions, unions, vehicle manufacturers, shipping companies and ports.
Title: Re: Self-Driving Cars
Post by: phooey on April 08, 2016, 07:01:54 PM
Alex,

Where in what I wrote did I say the road was closed?  I said it was on limited access highways, and they were driving in a chain, with human drivers, and we have no idea how much input they had.  They did nothing that a Tesla can't do already.

This doesn't run counter at all to anything I am saying.  Its no different than an airplane running on autopilot.  Its not dealing with the problems I have discussed all along, pedestrians, bicycles, unmarked roads, bad weather, no steering wheel to save it, no road construction, wild animals running around, cars weaving in and out, stop lights, balloons, skateboards, pot holes, grannies with walkers, motorbikes, plastic shops bags blowing in the wind, parking lots,...

Volvo has been trying to push their sdv technology for a while.  I know the limitations they require right now to attempt any self driving.  You have no idea how much human input was even involved in this trip.  Yet you all accuse me of "cherry picking" when I quote people's whose very job it is to push this technology as saying they are still skeptical that it can be done anytime soon. 
Title: Re: Self-Driving Cars
Post by: Alex Simmons on April 08, 2016, 07:13:43 PM
Alex,

Where in what I wrote did I say the road was closed?

When you said this:
Well, this really is amazing!  They were able to self-drive these trucks on closed highways, with well marked European roads, driving in formation so as not to allow other cars to intervene in their chain, and with professionals drivers monitoring the trucks and taking control during difficult sections.  I can't believe it! 

The highways used were not closed. The routes involved more than just highways, and if you took a little time to look into the project some more, you'll see them working on the technology and how it being developed and tested to manage interactions with other road users, e.g. with cars interspersing themselves between trucks, making risky manoeuvres around the convoy and so on.

As Rachel Hunter said, it won't happen overnight but it will happen.
Title: Re: Self-Driving Cars
Post by: Alex Simmons on April 08, 2016, 07:15:42 PM
Here we see testing examples of the systems in place.

https://youtu.be/fHibsfiIrhk?t=49

Tech that link the trucks so that braking manoeuvres can be done far more quickly, and tech that manages interactions with other vehicles in the convoy.
There are plenty of videos about to look at examples of more variable behaviour by drivers of other vehicles.

We've already seen in this thread examples of Google's car detect and make allowances for cyclists.
Title: Re: Self-Driving Cars
Post by: teethering on April 08, 2016, 07:20:20 PM
The freight transportation adoption seems like the most obvious early use case, as you get most bang/buck and you can narrow the scope to minimize extra complicated downtown traffic driving.

The other thing I was thinking about how the tech could enter wide use is as a taxi service, a kind of extension of Uber.  The nice thing about this is also the limited scope in terms of area of service, since that appears to be an important factor in scaling.
Title: Re: Self-Driving Cars
Post by: arthwollipot on April 08, 2016, 08:37:18 PM
Uber is already rumoured (I believe all but confirmed) to have invested massively in self-driving cars. That's not just speculation - it's happening.

And SDVs are supposed to travel in phalanx. They're designed to. SDVs can travel closer to other SDVs than human drivers will because all are networked and being controlled by the same systems. That's one reason that SDVs will reduce traffic congestion, by packing more vehicles into a smaller area so that you have smaller blocks that move as one.
Title: Re: Self-Driving Cars
Post by: brilligtove on April 09, 2016, 12:11:11 AM
This thread is so much better when the Phoobot is blocked.

I have some family who are in the blue-collar range of jobs (most are white collar, and some are academics). One got into being a trucker last year and I was both happy and sad for him. I mean, I'm thrilled he got a job that he enjoys and is good at and that pays reasonably well. I'm not so thrilled that he's likely going to be out of work in a couple years because the SDVs will do it better and cheaper.
Title: Re: Self-Driving Cars
Post by: phooey on April 09, 2016, 01:58:30 PM
http://finance.yahoo.com/news/experts-caution-self-driving-cars-173124507.html


"Experts caution self-driving cars aren't ready for roads
Self-driving cars are more likely to hurt than help public safety because of unsolved technical issues, engineers and safety advocates told the government Friday

—Poorly marked pavement, including parking lots and driveways, could foil the technology, which relies on clear lane markings.

—Bad weather can interfere with vehicle sensors.

—Self-driving cars can't take directions from a policeman.

—Inconsistent traffic-control devices — horizontal versus lateral traffic lights, for example."


Don't cherry pick brillingtrove.  (Have you been banned)?)
Title: Re: Self-Driving Cars
Post by: Alex Simmons on April 09, 2016, 05:30:22 PM
http://finance.yahoo.com/news/experts-caution-self-driving-cars-173124507.html


"Experts caution self-driving cars aren't ready for roads
Self-driving cars are more likely to hurt than help public safety because of unsolved technical issues, engineers and safety advocates told the government Friday

—Poorly marked pavement, including parking lots and driveways, could foil the technology, which relies on clear lane markings.

—Bad weather can interfere with vehicle sensors.

—Self-driving cars can't take directions from a policeman.

—Inconsistent traffic-control devices — horizontal versus lateral traffic lights, for example."


Don't cherry pick brillingtrove.  (Have you been banned)?)
These may be problems currently, but why are they insurmountable in future? Consider the progress that has been made to date. it's as if you think that there will never be any further improvements.

So, do you admit that you said the highway was closed in that Euro platoon test, even though they were on open roads?
Title: Re: Self-Driving Cars
Post by: phooey on April 09, 2016, 06:48:31 PM
No Alex, what I was saying was that they are closed highways, meaning they are closed to outside traffic like pedestrian, carts, motorbikes, etc.  Where I live we call these limited access roads.  People are not allowed to take walks on them, there are no traffic lights, you can't stop on the side of the road, etc. 

Furthermore, it looks like the routes were all checked out beforehand, and its on very modern European roads.  They also drove in tandem, so there was no human driven vehicles with different route intentions between the trucks. And you don't have any idea how much human input there was during these trips.  I have maintained all along that this is not the challenge to self-driving vehicles, any more than flying a plane in a sky because a plane does not need to constantly decide about what defensive maneuvers in must take in unpredictable environments. 

So if I am accused of cherry picking by pointing out all the difficulties with SDV, what should your side be accused of? 
Title: Re: Self-Driving Cars
Post by: SkeptiQueer on April 09, 2016, 06:52:12 PM
So again, why are the issues mentioned in the yahoo article insurmountable, and what basis do you have for declaring your 200 year timeline?

Sent from my SM-G530T using Tapatalk

Title: Re: Self-Driving Cars
Post by: Alex Simmons on April 10, 2016, 01:22:15 AM
No Alex, what I was saying was that they are closed highways, meaning they are closed to outside traffic like pedestrian, carts, motorbikes, etc.  Where I live we call these limited access roads.  People are not allowed to take walks on them, there are no traffic lights, you can't stop on the side of the road, etc.

Closed means closed to all traffic. These were not closed roads.

Furthermore, it looks like the routes were all checked out beforehand, and its on very modern European roads.  They also drove in tandem, so there was no human driven vehicles with different route intentions between the trucks. And you don't have any idea how much human input there was during these trips.  I have maintained all along that this is not the challenge to self-driving vehicles, any more than flying a plane in a sky because a plane does not need to constantly decide about what defensive maneuvers in must take in unpredictable environments. 
Given the nature of this initial experiment in the public domain, it is to be completely expected that there will be a level of human control and support. So what? It was a test of what's possible now, not a failure to undertake total automation since that was not the project's aim. You seem to think that the challenges are insurmountable, yet the sort of things being achieved now would have been considered insurmountable not so long ago.

So if I am accused of cherry picking by pointing out all the difficulties with SDV, what should your side be accused of?
It seems you have decided not to read on and investigate further into this particular project to see that in fact they did deal with regular traffic on the roads, that they have implemented automated systems to deal with cars driving in and around them, in between the trucks, and making wild lane pass manoeuvres etc.

I'm not cherry picking, nor on any "side". I'm just pointing out examples of the sort of things that are being achieved today, and challenges that have been addressed so far - as it's indicative of the collective ability of those involved in these projects to take on other newer and complex challenges involved with driving automation.

Will full automated self-drive happen? I really don't know and neither do you - but one thing we do know is that quite impressive advances are being made and it would be foolish to completely rule out such a possibility since impressive advances will continue. Industry applications will lead - goods transportation and mining are the prime initial candidates (mine trucks in Australia are self drive now). These sorts of technologies that are available now will save lives and they will save money, fuel and reduce pollution. They will also continue to get better.
Title: Re: Self-Driving Cars
Post by: Guillermo on April 10, 2016, 11:15:56 AM
The funny things happening on the way to singularity (http://techcrunch.com/2016/04/09/the-funny-things-happening-on-the-way-to-singularity/)

Quote
Should machines break the law?

It’s an important question to address as machines become more autonomous; instinctively, we’d say no, they shouldn’t break the law. But how about this example. You’re riding in an autonomous car and it’s getting on the freeway. The manually driven vehicles are all driving 75 MPH in a 55 zone. Does the autonomous car speed up to merge, and in the process break the law, or does it attempt to merge at the speed limit. If the latter, I imagine the potential for road rage goes up significantly.

I’ve talked to more than a few people who suggest that autonomous vehicles would have their own lanes, or be geo-fenced somehow. While that may work in the long term, there is a massive transition between then and now. The average car is on the road for 11 years. It will be at least 20 years before autonomous cars outnumber human-driven vehicles.
Title: Re: Self-Driving Cars
Post by: phooey on April 10, 2016, 12:21:09 PM
Where I live lanes are often unpredictable, the inner lanes can suddenly turn into a turning lane, and you really can't know when this is the case until you are already into the turn lane.  If you haven't had time to merge right, you got stuck with the cars turning left.  The only way out is often to make a quick dart to the right, in between breaks in traffic. Its a tricky move, and only possible because cars coming up on the outside lanes sort of know what you are doing and they also swerve further right, usually into another cars lane, or onto the shoulder.  Its sort of a nightmare of chain reactions, that people are used to, and they just accommodate by not following any rules at all.   

An automated car would be stuck there in paralysis, with a line of traffic behind it, beeping their horns like crazy, or throwing bottles of water at it.   
Title: Re: Self-Driving Cars
Post by: Enkidu on April 10, 2016, 12:29:29 PM
It seems pretty obvious that the first self-driving cars will be freight trucks which will do 99% of their driving where freight trucks normally do 99% of their driving - on interstate highways, avoiding large urban centers. Yes some roads are unequivocally going to be more difficult for AI to navigate reliably - but it is not all or nothing. Just as tractors and horse and buggies are forbidden from driving on the interstate, it's likely AI will be, at least initially, restricted from driving off of the interstate.
Title: Re: Self-Driving Cars
Post by: phooey on April 10, 2016, 12:54:01 PM
Right, but then what?  What is the next evolution? 
Title: Re: Self-Driving Cars
Post by: Enkidu on April 10, 2016, 12:55:26 PM
Better AI. It's not going to come out in one fell swoop. There will be incremental improvements. One step at a time.
Title: Re: Self-Driving Cars
Post by: Alex Simmons on April 10, 2016, 04:38:48 PM
Where I live lanes are often unpredictable, the inner lanes can suddenly turn into a turning lane, and you really can't know when this is the case until you are already into the turn lane.  If you haven't had time to merge right, you got stuck with the cars turning left.  The only way out is often to make a quick dart to the right, in between breaks in traffic. Its a tricky move, and only possible because cars coming up on the outside lanes sort of know what you are doing and they also swerve further right, usually into another cars lane, or onto the shoulder.  Its sort of a nightmare of chain reactions, that people are used to, and they just accommodate by not following any rules at all.   

An automated car would be stuck there in paralysis, with a line of traffic behind it, beeping their horns like crazy, or throwing bottles of water at it.   
Well there are cars now that have a "traffic jam mode" (my term) - they automatically manage the car in stop/start traffic, so I see no reason why they can't eventually deal with such dynamic scenarios in future, indeed they will likely eventually manage it better than human drivers who have a tendency to let ego get in the road of good group decision making.

And one can easily imagine intelligent vehicles will be better able to anticipate such things because they (a) have detailed mapping to start with so are able to anticipate and deal with such situations as well as make the merging smoother for all, and (b) vehicles record and upload temporary/new road blockage or obstructions that are deviations from the mapping data that are available to all networked vehicles in the area. When the obstruction is cleared, the next automated vehicle to pass that location can then upload the new "all clear" status. My $100 GPS mapping unit already gives me prior warning of busy sections of road ahead, the length of the delay, and suggests alternative routes if I want, or advises me that I'm still on the fastest route but to expect a delay in that section.
Title: Re: Self-Driving Cars
Post by: DG on April 10, 2016, 07:09:40 PM
Where I live lanes are often unpredictable, the inner lanes can suddenly turn into a turning lane, and you really can't know when this is the case until you are already into the turn lane.  If you haven't had time to merge right, you got stuck with the cars turning left.  The only way out is often to make a quick dart to the right, in between breaks in traffic. Its a tricky move, and only possible because cars coming up on the outside lanes sort of know what you are doing and they also swerve further right, usually into another cars lane, or onto the shoulder. Its sort of a nightmare of chain reactions, that people are used to, and they just accommodate by not following any rules at all.   

An automated car would be stuck there in paralysis, with a line of traffic behind it, beeping their horns like crazy, or throwing bottles of water at it.   

Wouldn't it be awesome if there were some way for the cars to talk to each other and cause the vehicles in one lane to create the requisite gap in traffic for the merging traffic (in a zipper formation for example, or whichever algorithm is most efficient at the time) without relying on some human agent deciding whether they will or will not let the car join the traffic?
Title: Re: Self-Driving Cars
Post by: Fast Eddie B on April 10, 2016, 07:22:53 PM
Wouldn't it be awesome if there were some way for the cars to talk to each other and cause the vehicles in one lane to create the requisite gap in traffic for the merging traffic (in a zipper formation for example, or whichever algorithm is most efficient at the time) without relying on some human agent deciding whether they will or will not let the car join the traffic?

Well, that's just crazy talk.

No way computers could ever talk to each other without wires!
Title: Re: Self-Driving Cars
Post by: arthwollipot on April 10, 2016, 08:00:25 PM
Where I live lanes are often unpredictable, the inner lanes can suddenly turn into a turning lane, and you really can't know when this is the case until you are already into the turn lane.  If you haven't had time to merge right, you got stuck with the cars turning left.  The only way out is often to make a quick dart to the right, in between breaks in traffic. Its a tricky move, and only possible because cars coming up on the outside lanes sort of know what you are doing and they also swerve further right, usually into another cars lane, or onto the shoulder.  Its sort of a nightmare of chain reactions, that people are used to, and they just accommodate by not following any rules at all.   

An automated car would be stuck there in paralysis, with a line of traffic behind it, beeping their horns like crazy, or throwing bottles of water at it.   

Nope. An automated car would be preprogrammed with the layout of the road so that this didn't occur. Unlike a human driver, a SDV could change lanes earlier in order to be in the proper lane for its destination. At worse, it could take the right turn and update to find a new route to the destination - just like GPS systems do today.
Title: Re: Self-Driving Cars
Post by: arthwollipot on April 10, 2016, 08:06:45 PM
There you go - several different solutions to yet another non-problem.
Title: Re: Self-Driving Cars
Post by: SkeptiQueer on April 10, 2016, 09:11:24 PM
And yet another example of a case where self-driving vehicles aren't the problem, asshole drivers are.

Sent from my SM-G530T using Tapatalk

Title: Re: Self-Driving Cars
Post by: lubbarin on April 10, 2016, 09:36:42 PM

I may be way off, but it seems like there's an elephant in the room.
Just a quick browse through the Bureau of Labor Statistics
Long haul truck drivers approx. 1.8 mil median pay 39,500
Bus drivers 665,000 median pay 30k
Taxi drivers 233,700 median pay 22k
Misc delivery truck drivers 1.33 mil median pay 27.4k

From the United States Postal Service
Around 620k employees averaging 70k, conservative estimate 1/2 are drivers.

Round about 300k ups and FedEx making between 40k and 90k


I'm no luddite, but holy crap. Delivery drones and driverless cars appear at a glance to have the potential to make around 4 million jobs obsolete (in theory). These range from lower-middle to middle class jobs.


There is absolutely, positively no stopping it.
So...  what? Labor demand to build and maintain the automated infrastructure will not be a drop in the bucket next to the losses.
Our national attitude towards unneeded labor ranges from punish the unemployed to funnel them into alternative careers, but none of that seems adaquate to cope with the future employment environment.

And that's just one corner of the economy. This story is reflected to one degree or another in almost every sector of widespread employment, and I seriously question whether the process can possibly be slow enough for adjustment.


Let's just go ahead and lay out an outrageous hypothetical.

Next week we get official word, there is overwhelming expert consensus that all transportation and delivery will be fully automated in 5 years. 5-6 million jobs gone, median pay 40k.
Those will be replaced by 200-220k jobs paying 60-90k.
Amazon profits will skyrocket 30%, which they plan to reinvest in programs designed to dismantle, automate, and take over the entirety of food service in the United states.
Top scholars agree it's an atom bomb dropping on the economy. A USA Today poll shows that 85% of all human beings think that this is dreadful news, and they passionately hate every bit of it.

I submit that in that case it would happen anyway
There is no stopping progress.
I don't honestly see it playing out with such dramatic suddenness, but we have serious thinking to do in the coming generations.

The real question, in my opinion, is can we change our fundamental economic model to compensate as this story plays out in one industry after another? Is it even possible?
Title: Re: Self-Driving Cars
Post by: SkeptiQueer on April 10, 2016, 10:50:11 PM
As I noted, the railroad unions have mandated having a man to push the button, even when human operators have caused more crashes, as a way to preserve jobs and also as a sort of liability pressure valve. They require fewer operators than before, and as time goes on certain lines have gone from 4 to 2 or even 1. Very complicated process of bidding and heading involved.

Similarly, I would expect the teamsters to push for the same, which would make trucking much easier job. Still, just as renewable energy puts coal miners out of business we have to be prepared for automation to eat jobs.

We're not prepared now, but we need to be going slash-and-burn through our system to prepare for it. Education/vocational rehabilitation programs, welfare programs, new jobs (rebuilding infrastructure) and tax reform all need to happen. The alternative isn't pretty, but I don't believe a good case can be made for paying people to supervise robots.

Sent from my SM-G530T using Tapatalk

Title: Re: Self-Driving Cars
Post by: Enkidu on April 11, 2016, 02:27:17 AM
You've probably already seen this, but just in case, here it is:

https://www.youtube.com/watch?v=7Pq-S557XQU

If CGP Grey is right, these jobs will never be replaced. We are moving into a world where machines do almost everything of economic value. Capitalism as we understand it (as well as every other existing economic model) is going to be completely obsolete. Our economic models are not prepared to deal with anything but an economy of scarcity.
Title: Re: Self-Driving Cars
Post by: phooey on April 11, 2016, 03:19:47 AM
All of these supposed solutions are only solutions once the entire transportation is only self driving cars (which will never happen in any of our lifetimes).

And skeptiqueer, you have it backwards, the hole reason why the system works is because humans understand what is going on, and they bend the rules to accommodate it. I know you like to lock-up everyone going one mile an hour over the speed limit, but actually sometimes bending the rules is a good thing in society, not a bad one. 

Keep your glock holstered. 
Title: Re: Self-Driving Cars
Post by: SkeptiQueer on April 11, 2016, 03:57:09 AM
But if people weren't shitty drivers, they would allow you to merge, negating the need to dart into traffic...

Sent from my SM-G530T using Tapatalk

Title: Re: Self-Driving Cars
Post by: phooey on April 11, 2016, 04:42:08 AM
They do allow you to merge skeptiqueer, that is the point.  The car on the left is stuck in a line that is turning, and not moving.  The traffic in the right lanes in continuing ahead in a steady stream of cars going through the green light.  So people just turn their wheels towards the right, to try to dart into the stream, and the other cars just swerve out of the way and keep going. 

Now what would a computer do, if it was mixed with human powered vehicles?  Would it be able to nudge into the human powered stream of cars, and cut them off?  I don't think so.  And would a computer driven car be able to slide across a solid line to give more space for the car to turn into?  Or would it have to stop completely, messing up the whole chain? 
Title: Re: Self-Driving Cars
Post by: SkeptiQueer on April 11, 2016, 06:57:14 AM
What would a computer do? Anticipate the traffic patterns to avoid getting stuck in yet another scenario that is avoidable. Unless you're saying the roads just change at random, so on Tuesday it's a regular lane of travel and on Wednesday it's a turn lane because LOL FUCK TRAFFIC.

Maybe you're right. Maybe in your terrifyingly designed roads, which you described as a "nightmare of chain reactions" you'll be stuck hoping the other drivers can react in time.

What''s the rate of fatality or the overall accident rate in your country?

Sent from my SM-G530T using Tapatalk

Title: Re: Self-Driving Cars
Post by: phooey on April 11, 2016, 07:31:45 AM
What's the overall accident rate when cars are allowed to drive without the aid of humans?

Oh wait, we don't know, because they can't. 
Title: Re: Self-Driving Cars
Post by: Fast Eddie B on April 11, 2016, 08:14:55 AM
All of these supposed solutions are only solutions once the entire transportation is only self driving cars (which will never happen in any of our lifetimes)...
Keep your glock holstered.

I call trolling.
Title: Re: Self-Driving Cars
Post by: teethering on April 11, 2016, 01:01:21 PM

I may be way off, but it seems like there's an elephant in the room.
Just a quick browse through the Bureau of Labor Statistics
Long haul truck drivers approx. 1.8 mil median pay 39,500
Bus drivers 665,000 median pay 30k
Taxi drivers 233,700 median pay 22k
Misc delivery truck drivers 1.33 mil median pay 27.4k

From the United States Postal Service
Around 620k employees averaging 70k, conservative estimate 1/2 are drivers.

Round about 300k ups and FedEx making between 40k and 90k


I'm no luddite, but holy crap. Delivery drones and driverless cars appear at a glance to have the potential to make around 4 million jobs obsolete (in theory). These range from lower-middle to middle class jobs.


There is absolutely, positively no stopping it.
So...  what? Labor demand to build and maintain the automated infrastructure will not be a drop in the bucket next to the losses.
Our national attitude towards unneeded labor ranges from punish the unemployed to funnel them into alternative careers, but none of that seems adaquate to cope with the future employment environment.

And that's just one corner of the economy. This story is reflected to one degree or another in almost every sector of widespread employment, and I seriously question whether the process can possibly be slow enough for adjustment.


Let's just go ahead and lay out an outrageous hypothetical.

Next week we get official word, there is overwhelming expert consensus that all transportation and delivery will be fully automated in 5 years. 5-6 million jobs gone, median pay 40k.
Those will be replaced by 200-220k jobs paying 60-90k.
Amazon profits will skyrocket 30%, which they plan to reinvest in programs designed to dismantle, automate, and take over the entirety of food service in the United states.
Top scholars agree it's an atom bomb dropping on the economy. A USA Today poll shows that 85% of all human beings think that this is dreadful news, and they passionately hate every bit of it.

I submit that in that case it would happen anyway
There is no stopping progress.
I don't honestly see it playing out with such dramatic suddenness, but we have serious thinking to do in the coming generations.

The real question, in my opinion, is can we change our fundamental economic model to compensate as this story plays out in one industry after another? Is it even possible?

Fundamentally I think it boils down to this.

Do you believe that it's a coincidence that before in history when we saw industrialization and automation shrink the amount of labour required to make X, that at the same time there was a corresponding  increase in employment elsewhere, so that on average unemployment remained steady?

We've never had a recession due to technological progress.  Which is weird if you believe that automating things causes unemployment.  I don't believe it's a coincidence, I believe that labour in general is just a flexible resource and when demand for it in one area shrinks somewhere else it creates a profitable opportunity.

All that being said, it's not perfectly flexible and there are painful transitions and people suffer from these relatively quick and radical changes.  It's also likely that wages will fall as a result.  But it's also true that it's a net win in terms of how far your dollar goes, as presumably transportation gets significantly cheaper as a result, and that affects everything and helps everyone.  And I think that these gains are probably asymmetrically greater for the employers and the rich than they are for the working class.

So to sum it all up, I think framing these problems as "jobs disappearing" will always appear as if you've incorrectly identifying the problem.  The problem will always look like, regardless of whether it's because of automation or anything else, the rich by virtue of being rich will be able to take advantage of economic growth far more than the working class and without government intervention the wealth gap and the income gap will continue to grow.  So what we really need to be worrying about is wealth re-distribution.
Title: Re: Self-Driving Cars
Post by: Enkidu on April 11, 2016, 01:54:55 PM
I largely agree as far as wealth distribution goes, though I also tend to agree with CGP Grey. This isn't like previous transitions. We are looking at automation taking over just about everything. Within a century it is possible that machines will be able to do everything that people can do, and do it better and cheaper. The list CGP Grey gives of top jobs today vs in the late 1700's is telling - they are basically the same jobs.

New fields that have emerged have not replaced old ones as far as employment numbers - the top 30 job types are the same as they were 100 years ago. If those go, we are in trouble. You do not replace 100 office workers with 100 software engineers to make the software bots that took their jobs - one software bot created by one engineer could replace thousands of people. We have not been here before.

Physical labor, mental labor - even art to some extent - are subject to automation. We are, potentially, looking at the end of human labor. Even if you can come up with new fields for humans to work in, machines will quickly take those over.
Title: Re: Self-Driving Cars
Post by: teethering on April 11, 2016, 04:17:39 PM
I largely agree as far as wealth distribution goes, though I also tend to agree with CGP Grey. This isn't like previous transitions. We are looking at automation taking over just about everything. Within a century it is possible that machines will be able to do everything that people can do, and do it better and cheaper. The list CGP Grey gives of top jobs today vs in the late 1700's is telling - they are basically the same jobs.

New fields that have emerged have not replaced old ones as far as employment numbers - the top 30 job types are the same as they were 100 years ago. If those go, we are in trouble. You do not replace 100 office workers with 100 software engineers to make the software bots that took their jobs - one software bot created by one engineer could replace thousands of people. We have not been here before.

Physical labor, mental labor - even art to some extent - are subject to automation. We are, potentially, looking at the end of human labor. Even if you can come up with new fields for humans to work in, machines will quickly take those over.

Well, I think this is committing the same error as we've been committing every time these improvements in technology rolled around.  People will want people to interact with for lots of activities and will be able to exchange resources they have for these services.  For example right now a lot more people are involved in the fancier food service industry than before.  There are a lot more yoga classes and, consequently, a lot more yoga instructors.  Economy transforms to take advantage of a new rate of exchange between labour and other stuff, what appeared to be luxury before becomes more commonplace and every technological step forward shifts this rate of exchange in favour of labour.

The only thing to worry about is whether "other stuff" is being disproportionately owned or controlled by fewer and fewer people who are able to leverage their wealthy to get better and better deals for themselves.  What we want is a more equitable distribution of economic gains.  I don't think there's any reason to worry about not finding uses for idle hands and minds.
Title: Re: Self-Driving Cars
Post by: arthwollipot on April 11, 2016, 08:24:46 PM
If CGP Grey is right, these jobs will never be replaced.

If he's right. We certainly don't know if he is or not. But even if he is - and in every case in the past it hasn't been that way - it doesn't mean the downfall of civilised society.

New jobs will appear to replace old ones. This is absolutely undeniable. What is debatable is to what extent those new jobs will replace those lost - it may be that more jobs are lost than are created by the new economy. But with governments even today experimenting with unconditional basic income (https://en.wikipedia.org/wiki/Basic_income), unemployment may not be the bugbear in the future that it is today. Under Basic, it is possible for a person to be unemployed and not be a drag on the economy.

But that's for a different thread. :)
Title: Re: Self-Driving Cars
Post by: Enkidu on April 11, 2016, 11:05:43 PM
If CGP Grey is right, these jobs will never be replaced.

If he's right. We certainly don't know if he is or not. But even if he is - and in every case in the past it hasn't been that way - it doesn't mean the downfall of civilised society.

New jobs will appear to replace old ones. This is absolutely undeniable. What is debatable is to what extent those new jobs will replace those lost - it may be that more jobs are lost than are created by the new economy. But with governments even today experimenting with unconditional basic income (https://en.wikipedia.org/wiki/Basic_income), unemployment may not be the bugbear in the future that it is today. Under Basic, it is possible for a person to be unemployed and not be a drag on the economy.

But that's for a different thread. :)

Oh I'm not arguing that it's the downfall of civilization (that would be silly). More that we are creating the technology for the kind of money-less space communism of Star Trek and we just aren't ready to do that on a cultural or emotional level. I have conservative relatives that worry that this is going to just fill the world up with lazy moochers. The only thing that is worse for a conservative than a failed welfare state is a successful one. I think some kind of mandatory minimum income is going to be essential - I just don't believe there will be enough jobs to go around.

We will reach a point where there really isn't anything people can do that a machine can't do better. When Teethering talks about yoga instructors and the human element of commerce - it sounds like we'll all be working in the modern equivalent of Colonial Williamsburg, churning butter and plowing fields with oxen because that's what the customer is going to want. It just sounds silly to me. Yes, there will be a few of those jobs around. Some people will insist on paying extra for coffee made by a human barista or having food cooked by a human - but how many of those jobs are there going to be? If 40 million people get thrown out of work, how many new yoga instructors will there be a demand for? Capitalism isn't going to be able to fix this.

I'm not worried about the end result - I think its going to be awesome. It's the transition I'm worried about, and all of the people who think it is immoral for OTHER people to not have to work.
Title: Re: Self-Driving Cars
Post by: phooey on April 11, 2016, 11:56:21 PM
http://www.theverge.com/2016/4/11/11405414/robot-waitstaff-china-restaurant-closed
Title: Re: Self-Driving Cars
Post by: Morvis13 on April 12, 2016, 08:08:09 AM
if the waitbots were like the picture then I think they built the wrong kind of robot. Like using a wrench to pound in a nail. I'm sure you have a different take on it. Anyway that argument is invalid when it comes to self driving cars.
We won't have that kind of robot driver until 2084

(click to show/hide)
Title: Re: Self-Driving Cars
Post by: Johnny Slick on April 12, 2016, 11:43:32 AM
I largely agree as far as wealth distribution goes, though I also tend to agree with CGP Grey. This isn't like previous transitions. We are looking at automation taking over just about everything. Within a century it is possible that machines will be able to do everything that people can do, and do it better and cheaper. The list CGP Grey gives of top jobs today vs in the late 1700's is telling - they are basically the same jobs.

New fields that have emerged have not replaced old ones as far as employment numbers - the top 30 job types are the same as they were 100 years ago. If those go, we are in trouble. You do not replace 100 office workers with 100 software engineers to make the software bots that took their jobs - one software bot created by one engineer could replace thousands of people. We have not been here before.

Physical labor, mental labor - even art to some extent - are subject to automation. We are, potentially, looking at the end of human labor. Even if you can come up with new fields for humans to work in, machines will quickly take those over.

Well, I think this is committing the same error as we've been committing every time these improvements in technology rolled around.  People will want people to interact with for lots of activities and will be able to exchange resources they have for these services.  For example right now a lot more people are involved in the fancier food service industry than before.  There are a lot more yoga classes and, consequently, a lot more yoga instructors.  Economy transforms to take advantage of a new rate of exchange between labour and other stuff, what appeared to be luxury before becomes more commonplace and every technological step forward shifts this rate of exchange in favour of labour.

The only thing to worry about is whether "other stuff" is being disproportionately owned or controlled by fewer and fewer people who are able to leverage their wealthy to get better and better deals for themselves.  What we want is a more equitable distribution of economic gains.  I don't think there's any reason to worry about not finding uses for idle hands and minds.
I think this is a bit simplistic. No, you can't *exactly* say that economies didn't fall due to technology, but the Industrial Revolution in particular caused a massive amount of *change* in the workforce. Having large families, for instance, swiftly became suboptimal where before it was a great way of running a personal farm (because every child you had, once they got to a certain age, meant one fewer hand you had to hire to mind the fields), and while this and the industrial revolution most definitely had an overall positive effect on the economy at large, it also had a detrimental effect on the lower class worker at first and, arguably, overall as well. The fact that labor movements coincide with the IR is not, as they say, a coincidence. Some of that is due to the lower class getting a taste of what the middle class had and wanting some of that for itself, to be sure, but an awful lot of it too was outrage over 6 year olds working in cotton factories (and often dying of lung diseases before ever reaching adulthood), people being more or less forced to work hard for 10-12 hours a day, 6 or 7 days a week (that describes farmwork as well to some extent, but there are natural breaks in farmwork both in the day and in the seasons that do not exist in factories), and so on.

The Industrial Revolution also served to centralize the economies of countries, which on the one hand meant that there were perhaps fewer instances of regions being beset by famine for several years (although on the other hand, Ireland) but on the other hand made nationwide and worldwide depressions possible in ways they were simply not before. We retired the word "depression" after the Great Depression but before it hit the US for example suffered through a depression (some of which would be reclassified as a "recession" today in the same way that what happened in the early 90s was a recession, some of which really were quite a bit worse) every 5-6 years or so, throwing vast swathes of the workforce onto the street for no other good reason than that we hadn't figured out how to manage economies yet. There were also major economic turndowns following major wars, which isn't *exactly* due to technological improvement but is certainly due to having to re-invent one's economy after making it reliant on new military technology for a few to several years beforehand.

I think the lower class came out of the Industrial Revolution better off overall, in part because of the birth of the labor movement, but I don't think that the phrase "the economy never went south due to technology" succeeds in telling that story. This is applicable here because as automation gets cheaper than manual labor, millions of people will be put out of work and forced to learn and create new niches. I have every confidence that we'll be able to do this but I am equally sure there will be huge growing pains involved.
Title: Re: Self-Driving Cars
Post by: Enkidu on April 12, 2016, 12:08:57 PM
I think the lower class came out of the Industrial Revolution better off overall, in part because of the birth of the labor movement, but I don't think that the phrase "the economy never went south due to technology" succeeds in telling that story. This is applicable here because as automation gets cheaper than manual labor, millions of people will be put out of work and forced to learn and create new niches. I have every confidence that we'll be able to do this but I am equally sure there will be huge growing pains involved.

I agree with pretty much everything you wrote except for that. It would be more accurate to say "as automation gets cheaper than manual and mental labor . . ." Software bots are already writing articles for publications; they are already taking mind-labor jobs. Our desk jobs are not safe, and in fact many of them will be automated before a lot of manual labor. There really is no reason to think that there is anything more than a small fraction of jobs that can't be done better and cheaper by a machine (eventually, and sooner than most of us think).

I think Teethering hit it on the head with the idea that most human jobs will exist solely due to the irrational desire to have human produced goods and services - and I really don't think that demand will be anything nearly strong enough to replace the jobs we are going to loose. It's similar to people who pay more to shop at the local mom and pop store instead of the "big box" store; sure, they exist, but most people just go to Walmart or Home Depot. The economic re-organization that this is going to necessitate will make the industrial revolution look like a cake walk in comparison.
Title: Re: Self-Driving Cars
Post by: teethering on April 12, 2016, 12:56:59 PM
I largely agree as far as wealth distribution goes, though I also tend to agree with CGP Grey. This isn't like previous transitions. We are looking at automation taking over just about everything. Within a century it is possible that machines will be able to do everything that people can do, and do it better and cheaper. The list CGP Grey gives of top jobs today vs in the late 1700's is telling - they are basically the same jobs.

New fields that have emerged have not replaced old ones as far as employment numbers - the top 30 job types are the same as they were 100 years ago. If those go, we are in trouble. You do not replace 100 office workers with 100 software engineers to make the software bots that took their jobs - one software bot created by one engineer could replace thousands of people. We have not been here before.

Physical labor, mental labor - even art to some extent - are subject to automation. We are, potentially, looking at the end of human labor. Even if you can come up with new fields for humans to work in, machines will quickly take those over.

Well, I think this is committing the same error as we've been committing every time these improvements in technology rolled around.  People will want people to interact with for lots of activities and will be able to exchange resources they have for these services.  For example right now a lot more people are involved in the fancier food service industry than before.  There are a lot more yoga classes and, consequently, a lot more yoga instructors.  Economy transforms to take advantage of a new rate of exchange between labour and other stuff, what appeared to be luxury before becomes more commonplace and every technological step forward shifts this rate of exchange in favour of labour.

The only thing to worry about is whether "other stuff" is being disproportionately owned or controlled by fewer and fewer people who are able to leverage their wealthy to get better and better deals for themselves.  What we want is a more equitable distribution of economic gains.  I don't think there's any reason to worry about not finding uses for idle hands and minds.
I think this is a bit simplistic. No, you can't *exactly* say that economies didn't fall due to technology, but the Industrial Revolution in particular caused a massive amount of *change* in the workforce. Having large families, for instance, swiftly became suboptimal where before it was a great way of running a personal farm (because every child you had, once they got to a certain age, meant one fewer hand you had to hire to mind the fields), and while this and the industrial revolution most definitely had an overall positive effect on the economy at large, it also had a detrimental effect on the lower class worker at first and, arguably, overall as well. The fact that labor movements coincide with the IR is not, as they say, a coincidence. Some of that is due to the lower class getting a taste of what the middle class had and wanting some of that for itself, to be sure, but an awful lot of it too was outrage over 6 year olds working in cotton factories (and often dying of lung diseases before ever reaching adulthood), people being more or less forced to work hard for 10-12 hours a day, 6 or 7 days a week (that describes farmwork as well to some extent, but there are natural breaks in farmwork both in the day and in the seasons that do not exist in factories), and so on.

The Industrial Revolution also served to centralize the economies of countries, which on the one hand meant that there were perhaps fewer instances of regions being beset by famine for several years (although on the other hand, Ireland) but on the other hand made nationwide and worldwide depressions possible in ways they were simply not before. We retired the word "depression" after the Great Depression but before it hit the US for example suffered through a depression (some of which would be reclassified as a "recession" today in the same way that what happened in the early 90s was a recession, some of which really were quite a bit worse) every 5-6 years or so, throwing vast swathes of the workforce onto the street for no other good reason than that we hadn't figured out how to manage economies yet. There were also major economic turndowns following major wars, which isn't *exactly* due to technological improvement but is certainly due to having to re-invent one's economy after making it reliant on new military technology for a few to several years beforehand.

I think the lower class came out of the Industrial Revolution better off overall, in part because of the birth of the labor movement, but I don't think that the phrase "the economy never went south due to technology" succeeds in telling that story. This is applicable here because as automation gets cheaper than manual labor, millions of people will be put out of work and forced to learn and create new niches. I have every confidence that we'll be able to do this but I am equally sure there will be huge growing pains involved.

L2read,  lib  ::)

All that being said, it's not perfectly flexible and there are painful transitions and people suffer from these relatively quick and radical changes.  It's also likely that wages will fall as a result.  But it's also true that it's a net win in terms of how far your dollar goes, as presumably transportation gets significantly cheaper as a result, and that affects everything and helps everyone.  And I think that these gains are probably asymmetrically greater for the employers and the rich than they are for the working class.

So to sum it all up, I think framing these problems as "jobs disappearing" will always appear as if you've incorrectly identifying the problem.  The problem will always look like, regardless of whether it's because of automation or anything else, the rich by virtue of being rich will be able to take advantage of economic growth far more than the working class and without government intervention the wealth gap and the income gap will continue to grow.  So what we really need to be worrying about is wealth re-distribution.

In an ideal world we have a robust welfare state that eases the pain of these transitions.  But like I say above, the key issue isn't really the rapid transitions and the welfare state, which is kind of an emergency response system, but a long term policy that makes a more equitable distribution of economic gains from progress its primary goal.  And the reason for this goal is because, as we can readily observe from the data of the last several decades, economic growth isn't lifting all boats, its only lifting the yachts.  It's both unjust and suboptimal in terms of long term health of society.
Title: Re: Self-Driving Cars
Post by: phooey on April 12, 2016, 01:10:31 PM
I think one has to wonder how much different company inventiveness and productivity would be affected if there was a salary cap put on the maximum any one person could be paid in salary in a year.  Say that maximum was 5 million dollars.  Would people really do last, contribute less, if they knew their maximum would be 5 million instead of say 100 million.  I think not.  And if they really would do less, I am not sure it would matter, there would be someone just as capable who would do it for 5 million. 

If we have a minimum wage, why not also a maximum.  I think it would make for a much more healthy world. 
Title: Re: Self-Driving Cars
Post by: SkeptiQueer on April 12, 2016, 02:36:00 PM
Because the people who spend millions on campaign contributions don't want one.

Sent from my SM-G530T using Tapatalk

Title: Re: Self-Driving Cars
Post by: phooey on April 14, 2016, 09:15:53 PM
(https://www.youtube.com/watch?v=MrwxEX8qOxA)

Anyone want to sleep during this commute? 

https://www.youtube.com/watch?v=MrwxEX8qOxA

Title: Re: Self-Driving Cars
Post by: Enkidu on April 14, 2016, 11:02:52 PM
The feature isn't supposed to be used in those driving conditions (country road, tight turns/hills). It's only supposed to be used on the highway. I'm sure it's not perfect, but in this instance it is clear user error.
Title: Re: Self-Driving Cars
Post by: phooey on April 15, 2016, 12:09:03 AM
Yea, he used it where other cars are also on the road.  You clearly don't want to do that!

Apparently it was intending to follow the other car on the road. An unfortunate byproduct of the system, if the car happens to be going in the opposite direction! 
Title: Re: Self-Driving Cars
Post by: Enkidu on April 15, 2016, 01:33:16 AM
1. This is not a self driving car. It's more comparable to cruise control and is intended for interstate use only - you know, where on-coming traffic is separated by a median and doesn't come within a few inches of you.


2. You are never supposed to remove your hands from the steering wheel as this guy does. That's like turning on cruise control and then putting your feet out of the window.


3. The sound you hear at around the 9 second mark is the autopilot telling the driver "I don't know what is going on here, you need to take over" which isn't a problem if you keep your hands on the wheel the way you are supposed to.

Title: Re: Self-Driving Cars
Post by: arthwollipot on April 15, 2016, 04:32:07 AM
Yea, he used it where other cars are also on the road.  You clearly don't want to do that!

Apparently it was intending to follow the other car on the road. An unfortunate byproduct of the system, if the car happens to be going in the opposite direction!

An unfortunate byproduct of a human driver misusing the system. If only there were a way to take human error out of the equation...
Title: Re: Self-Driving Cars
Post by: phooey on April 15, 2016, 06:13:34 AM
arthwollipot,

Yes, the human error of believing this car can drive by itself.  That was the only human error.  We need to elimnate that error.

Enkidu,

"This is not a self driving car."

Yes, I am well aware that this is not a self-driving car.  I think you might need to inform some others though.

Title: Re: Self-Driving Cars
Post by: bimble on April 15, 2016, 07:05:29 AM

Yes, I am well aware that this is not a self-driving car.  I think you might need to inform some others though.

So why are you using it as an example of why self-driving cars won't be any good at driving on the roads??
Title: Re: Self-Driving Cars
Post by: Guillermo on April 15, 2016, 09:03:03 AM
Quote
UPDATE: While using AutoPilot today, my Model S spotted a Tesla Roadster, followed it, and tried to mate with it.
Title: Re: Self-Driving Cars
Post by: Enkidu on April 15, 2016, 11:35:32 AM
Enkidu,

"This is not a self driving car."

Yes, I am well aware that this is not a self-driving car.  I think you might need to inform some others though.

If the guy in that video can't be assed to pay attention to Tesla's own recommendations, why would he listen to me? He's an idiot. If I try to cut a tree down with a paring knife, and the blade snaps off and flys into my eye, whose fault is that? Does it make cutting down a tree a pie-in-the-sky fairy tale that will never happen? Or does it mean that I'm an idiot?
Title: Re: Self-Driving Cars
Post by: brilligtove on April 15, 2016, 11:55:37 AM
This thread is a lot more fun when you block the Phoobot.
Title: Re: Self-Driving Cars
Post by: Morvis13 on April 15, 2016, 11:58:21 AM
Yea, he used it where other cars are also on the road.  You clearly don't want to do that!

Apparently it was intending to follow the other car on the road. An unfortunate byproduct of the system, if the car happens to be going in the opposite direction!

“A common mistake that people make when trying to design something completely foolproof is to underestimate the ingenuity of complete fools.” - Douglas Adams.

Title: Re: Self-Driving Cars
Post by: Johnny Slick on April 15, 2016, 01:51:32 PM
You dummies have no idea what you are talking about! If self driving vehicles are so superior, explain this!

https://www.youtube.com/watch?v=waWM9vAfsIw
Title: Re: Self-Driving Cars
Post by: Alex Simmons on April 15, 2016, 05:30:25 PM

Yes, I am well aware that this is not a self-driving car.  I think you might need to inform some others though.

So why are you using it as an example of why self-driving cars won't be any good at driving on the roads??

It's a clearly fallacious example. It was like phooey's truck platoon argument. These are examples of what these vehicles are and are not currently capable of. They are not however examples of failed totally autonomous vehicles.

Saying something is a failure because is doesn't meet a standard it was never designed for is a fallacious argument.
Title: Re: Self-Driving Cars
Post by: phooey on April 15, 2016, 06:06:27 PM
Except I didn't say that now did I Alex? You are arguing with your own strawman. 

But I will be happy to say that the technology isn't nearly as good as many of you want to claim here. And it also does mean you all have your own internal disagreements when some of you claim self-driving vehicles already exist, and some claim they don't.   
Title: Re: Self-Driving Cars
Post by: Enkidu on April 16, 2016, 01:40:55 AM
Except I didn't say that now did I Alex? You are arguing with your own strawman. 

But I will be happy to say that the technology isn't nearly as good as many of you want to claim here. And it also does mean you all have your own internal disagreements when some of you claim self-driving vehicles already exist, and some claim they don't.

Self driving vehicles already exist. They are not yet where they need to be but are getting close. Anyone here besides phooey disagree with this?
Title: Re: Self-Driving Cars
Post by: phooey on April 16, 2016, 02:11:23 AM
Endiku,

Wait, you already said the Tesla is not a self driving car, because it needs human intervention to keep it safe, and besides it can't even be used anywhere but a well marked freeway, and even then, the human has to be in control. 

So where are the self-driving cars?  Surely you don't mean a Google car, which can't go over 25 mph, and also need a human to prevent it from crashing?  What self driving car are you referring to? 
Title: Re: Self-Driving Cars
Post by: SkeptiQueer on April 16, 2016, 02:22:10 AM
Endiku,

Wait, you already said the Tesla is not a self driving car, because it needs human intervention to keep it safe, and besides it can't even be used anywhere but a well marked freeway, and even then, the human has to be in control. 

So where are the self-driving cars?  Surely you don't mean a Google car, which can't go over 25 mph, and also need a human to prevent it from crashing?  What self driving car are you referring to?
The Google car is self-driving. It's limited by law, not by technology.
Title: Re: Self-Driving Cars
Post by: Enkidu on April 16, 2016, 11:49:39 AM
What SkeptiQueer said. The Tesla is just a car that got a software update. I'm not familiar with the specifics, but I know it doesn't have the lidar thing on the roof spinning around giving it a 360 degree view - it's just using basic cameras to track road stripes. The two aren't really comparable, which is why Tesla warns drivers to keep their hands on the wheels. Honestly, I'm not sure how useful the Tesla feature really is - it seems more like a stunt at this point.
Title: Re: Self-Driving Cars
Post by: brilligtove on April 19, 2016, 02:10:58 PM
http://youtu.be/9I5rraWJq6E

This guys Tesla actively steered him out of an accident.
Title: Re: Self-Driving Cars
Post by: PANTS! on April 19, 2016, 02:26:22 PM
Yay Cleveland!  (https://www.google.com/maps/@41.4254907,-81.6832563,3a,75y,2.2h,85.24t/data=!3m6!1e1!3m4!1s-2GgwvqzQ5ftW5kKDeF5ng!2e0!7i13312!8i6656)
Title: Re: Self-Driving Cars
Post by: Friendly Angel on April 19, 2016, 04:19:19 PM
This guys Tesla actively steered him out of an accident.

Does the Tesla come with an angry honk and virtual finger subroutine?
Title: Re: Self-Driving Cars
Post by: brilligtove on April 19, 2016, 06:52:23 PM
This guys Tesla actively steered him out of an accident.

Does the Tesla come with an angry honk and virtual finger subroutine?

That's an upgrade package. Daniel1948 probably knows the pricing...?
Title: Re: Self-Driving Cars
Post by: phooey on April 20, 2016, 09:26:42 AM
It looks like the Tesla was not willing to cross the line and go over into the shoulder, and almost failed to avoid the crash.  A human would have moved over much further, and been much safer in that case. 
Title: Re: Self-Driving Cars
Post by: Morvis13 on April 20, 2016, 09:28:37 AM
It looks like the Tesla was not willing to cross the line and go over into the shoulder, and almost failed to avoid the crash.  A human would have moved over much further, and been much safer in that case.

It is quite possible the human would have over corrected, hit the shoulder and lost control.
Title: Re: Self-Driving Cars
Post by: brilligtove on April 20, 2016, 09:53:06 AM
I love how Phoobot desperately clings to it's flat-earth type ideas about SDVs, on those occasions when I actually see a Phoobot post in a quote. Would it be fair to call it delusional?
Title: Re: Self-Driving Cars
Post by: SkeptiQueer on April 20, 2016, 10:05:56 AM
It looks like the Tesla was not willing to cross the line and go over into the shoulder, and almost failed to avoid the crash.  A human would have moved over much further, and been much safer in that case.
Not necessarily. I'very seen humans overcorrect, undercorrect, and the ubiquitous failure to recognize what's happening and plow right into it, or try to brake and fail.
Title: Re: Self-Driving Cars
Post by: Johnny Slick on April 20, 2016, 11:14:44 AM
FALSE a human driver would have done a sweet skidout Magnum PI style. Magnum PI was not a robot, dummies.
Title: Re: Self-Driving Cars
Post by: Guillermo on April 20, 2016, 11:22:26 AM
FALSE a human driver would have done a sweet skidout Magnum PI style. Magnum PI was not a robot, dummies.
That you know of.
Title: Re: Self-Driving Cars
Post by: amysrevenge on April 20, 2016, 11:35:21 AM
FALSE a human driver would have done a sweet skidout Magnum PI style. Magnum PI was not a robot, dummies.

I seem to recall he was a ghost for a while though, at the end.  Or am I only dreaming that?
Title: Re: Self-Driving Cars
Post by: Ah.hell on April 20, 2016, 01:55:10 PM
FALSE a human driver would have done a sweet skidout Magnum PI style. Magnum PI was not a robot, dummies.

I seem to recall he was a ghost for a while though, at the end.  Or am I only dreaming that?
I'm pretty sure he rejoined the Navy in the last episode, but it was the 80s, so he probably had an out of body experience during a coma at some point. 
Title: Re: Self-Driving Cars
Post by: amysrevenge on April 20, 2016, 02:04:49 PM
FALSE a human driver would have done a sweet skidout Magnum PI style. Magnum PI was not a robot, dummies.

I seem to recall he was a ghost for a while though, at the end.  Or am I only dreaming that?
I'm pretty sure he rejoined the Navy in the last episode, but it was the 80s, so he probably had an out of body experience during a coma at some point.

OK now you made me look it up.  I was right...

Quote from: wikipedia
At the end of the seventh season, Magnum was to be killed off, which was intended to end the series. The final episode of the season, "Limbo", after seeing Magnum wander around as a ghost for nearly the entire run-time, closes with him appearing to walk off into heaven. However, following outcry from fans, who demanded a more satisfactory conclusion, an eighth, final season was produced to bring Magnum "back to life" and to round off the series.
Title: Re: Self-Driving Cars
Post by: Ah.hell on April 20, 2016, 03:37:03 PM
To be fair we were both right:
Quote
By the final episode of Magnum P.I. in 1988, Thomas had decided to return to active duty in the Navy at the rank Commander (O-5).
Title: Re: Self-Driving Cars
Post by: brilligtove on June 11, 2016, 12:18:42 AM
So yeah, this thread took a hard turn into a wall, unlike SELF DRIVING DINO CARS.

(http://www.qwantz.com/comics/comic2-2997.png) (http://www.qwantz.com/index.php?comic=2991)

...So I'm not entirely sure I'm comfortable with my new robot overcabs anymore.
Title: Re: Self-Driving Cars
Post by: arthwollipot on June 11, 2016, 02:56:59 AM
In the world described by the T-rex, why would jaywalking still be a crime?
Title: Re: Self-Driving Cars
Post by: brilligtove on June 16, 2016, 07:49:56 AM
http://readwrite.com/2016/06/15/millions-self-driving-cars-expected-global-roads-2035-tl4

The global fleet will take quite some time to overturn, but not as long as many would think.
Title: Re: Self-Driving Cars
Post by: daniel1948 on June 16, 2016, 10:20:40 AM
http://readwrite.com/2016/06/15/millions-self-driving-cars-expected-global-roads-2035-tl4

The global fleet will take quite some time to overturn, but not as long as many would think.

Cool! "Get in your car, go to sleep, and let the car take you to your destination, by 2019." That's probably overly optimistic. But the main text suggests they'll be available by 2025. I'll be 77 if I can manage to live that long.
Title: Re: Self-Driving Cars
Post by: DG on June 16, 2016, 09:33:57 PM
So yeah, this thread took a hard turn into a wall, unlike SELF DRIVING DINO CARS.

(http://www.qwantz.com/comics/comic2-2997.png) (http://www.qwantz.com/index.php?comic=2991)

...So I'm not entirely sure I'm comfortable with my new robot overcabs anymore.

Hmmm, this makes me more comfortable with robocars... When a driver of a non-robocar breaks the law, video evidence could be provided to the police immediately. Reduce the size of the police force...smaller government - I am sure that will appeal to some...
Title: Re: Self-Driving Cars
Post by: phooey on June 30, 2016, 10:08:56 PM
Tesla driver killed in crash while using car's 'Autopilot'

"Tesla said on its website that neither the driver nor the Autopilot noticed the white side of the trailer, which was perpendicular to the Model S, against the brightly lit sky, and neither applied the brakes."

Here I was worried about nefarious car-jackers, and kids on skateboards.  Turns out we aren't even ready for sunshine yet. 


https://www.yahoo.com/news/self-driving-car-driver-died-205642937.html
Title: Re: Self-Driving Cars
Post by: phooey on July 01, 2016, 10:06:19 AM
http://abcnews.go.com/Technology/wireStory/tesla-crash-hurt-sentiment-driverless-cars-40271080?yptr=yahoo


Now there is an understatement! 
Title: Re: Self-Driving Cars
Post by: Jeremy's Sea on July 01, 2016, 11:15:32 AM
Sucks that figuring out that (seems obvious now?) limitation cost a life.
Title: Re: Self-Driving Cars
Post by: daniel1948 on July 01, 2016, 12:40:20 PM
Tesla driver killed in crash while using car's 'Autopilot'

"Tesla said on its website that neither the driver nor the Autopilot noticed the white side of the trailer, which was perpendicular to the Model S, against the brightly lit sky, and neither applied the brakes."

Here I was worried about nefarious car-jackers, and kids on skateboards.  Turns out we aren't even ready for sunshine yet. 


https://www.yahoo.com/news/self-driving-car-driver-died-205642937.html

This has nothing to do with self-driving cars. This was one more tragic automobile death, which you are cynically and disgustingly using in your ongoing rant against technology.

Item: Tesla has always said its technology is an aid, not a substitute for your own eyes on the road and hands on the wheel. It makes the car a little more safe. It does not make it perfect.

Item: this is the first death in a Model S in 130 million passenger miles, compared to 94 million U.S. fleet average. Still the safest car on the road. And still not a self-driving car. People will die in cars. Even when there are self-driving cars people will die in them. A few sensationalist reporters are shouting, without any evidence, that this will undermine consumer confidence in self-driving cars, ignoring that the Model S is NOT a self-driving car.

Of course, shameless neo-Luddites will use this tragic death to argue against a technology that has nothing to do with this accident, as if this car were supposed to be self driving, and as if one single death meant the technology was worthless. It's disgusting.

The Tesla Model S is the safest car on the road, but it's still a car, and cars still kill people, and always will. Self-driving cars will kill fewer people than human drivers do. And the technology in the Model S and some other high-end cars, make it and them a little safer. They do not make them perfect. My condolences to the family of the victim, and to the families of the roughly thirty-two thousand people killed in the U.S. by human drivers every year.
Title: Re: Self-Driving Cars
Post by: Fast Eddie B on July 01, 2016, 12:53:48 PM
Sucks that figuring out that (seems obvious now?) limitation cost a life.

Pretty sure there will always be limitations. One can imagine a human being "blind" to a white truck converging against a white background, or a green car lost in surrounding foliage.

I'd think cameras and computers can be tweaked to be "better" than a human, maybe even "seeing" in the ultraviolet and infrared, but to assume zero limitations - i.e. perfection - will ever be reachable is unreasonable.
Title: Re: Self-Driving Cars
Post by: SkeptiQueer on July 01, 2016, 01:01:24 PM
Sucks that figuring out that (seems obvious now?) limitation cost a life.

Pretty sure there will always be limitations. One can imagine a human being "blind" to a white truck converging against a white background, or a green car lost in surrounding foliage.

I'd think cameras and computers can be tweaked to be "better" than a human, maybe even "seeing" in the ultraviolet and infrared, but to assume zero limitations - i.e. perfection - will ever be reachable is unreasonable.

Humans can literally not see a car or person just because they don't expect to see them.
Title: Re: Self-Driving Cars
Post by: Friendly Angel on July 01, 2016, 01:02:42 PM

Item: Tesla has always said its technology is an aid, not a substitute for your own eyes on the road and hands on the wheel. It makes the car a little more safe. It does not make it perfect.

The name "Autopilot" does *sound* like "self-driving"... not a good name for what the Tesla system does.

The deceased driver recorded his Tesla preventing a potentially nasty accident just a couple months ago:

http://jalopnik.com/man-killed-in-self-driving-tesla-recorded-video-of-auto-1782918905

Title: Re: Self-Driving Cars
Post by: teethering on July 01, 2016, 01:13:47 PM
Tesla driver killed in crash while using car's 'Autopilot'

"Tesla said on its website that neither the driver nor the Autopilot noticed the white side of the trailer, which was perpendicular to the Model S, against the brightly lit sky, and neither applied the brakes."

Here I was worried about nefarious car-jackers, and kids on skateboards.  Turns out we aren't even ready for sunshine yet. 


https://www.yahoo.com/news/self-driving-car-driver-died-205642937.html

Did you notice that?  Kind of important.
Title: Re: Self-Driving Cars
Post by: Jeremy's Sea on July 01, 2016, 01:30:06 PM
Sucks that figuring out that (seems obvious now?) limitation cost a life.

Pretty sure there will always be limitations. One can imagine a human being "blind" to a white truck converging against a white background, or a green car lost in surrounding foliage.

I'd think cameras and computers can be tweaked to be "better" than a human, maybe even "seeing" in the ultraviolet and infrared, but to assume zero limitations - i.e. perfection - will ever be reachable is unreasonable.
I wasn't in anyway trying to suggest that there will never be limitations, just commenting that it's too bad someone had to die in the process and it wasn't just a crash. They should also make sure the car doesn't check facebook while it drives, like that idiot who almost ran over my dog in front of my house yesterday as we were walking...

I did find it kind of disturbing that phooey used a tragedy to advance his own agenda, but that's what the ideologically blind (and politicians) do.
Title: Re: Self-Driving Cars
Post by: bimble on July 01, 2016, 01:46:25 PM
Quote
n a statement, Tesla said it appeared the Model S car was unable to recognise "the white side of the tractor trailer against a brightly lit sky" that had driven across the car's path.

So the lorry had hit the car, as opposed to the car hitting the lorry.

Quote
The company said in a statement: "The high ride height of the trailer combined with its positioning across the road and the extremely rare circumstances of the impact caused the Model S to pass under the trailer, with the bottom of the trailer impacting the windshield of the Model S."

Here in the UK and Europe, HGV trailers are designed so you can't pass under the trailer in a car, to stop these very type of accidents.
Title: Re: Self-Driving Cars
Post by: Fast Eddie B on July 01, 2016, 01:46:37 PM
I wasn't in anyway trying to suggest that there will never be limitations, just commenting that it's too bad someone had to die in the process and it wasn't just a crash. They should also make sure the car doesn't check facebook while it drives, like that idiot who almost ran over my dog in front of my house yesterday as we were walking...

I did find it kind of disturbing that phooey used a tragedy to advance his own agenda, but that's what the ideologically blind (and politicians) do.

Understood. And agreed.
Title: Re: Self-Driving Cars
Post by: Friendly Angel on July 01, 2016, 02:22:51 PM
(https://static01.nyt.com/images/2016/07/01/business/tesla-crash-police-report-diagram/tesla-crash-police-report-diagram-superJumbo.png)

What is "FR" ?

Here's the actual intersection:
https://www.google.com/maps/place/NE+140th+Ct,+Williston,+FL+32696/@29.4106381,-82.5397444,281m/data=!3m1!1e3!4m5!3m4!1s0x88e892986f8d8eaf:0x4cebe4b1d7706926!8m2!3d29.4037061!4d-82.539709

Also... this:

http://www.cnbc.com/2016/07/01/truck-driver-involved-with-fatal-tesla-accident-says-driver-was-watching-a-movie.html


The truck driver claimed he heard the video playing a movie when he went to check on the driver but thought it was on the car's touch screen, which it can't do. 
Title: Re: Self-Driving Cars
Post by: wastrel on July 01, 2016, 02:41:33 PM
"FR" = Full Rest or something along those lines I think.
Title: Re: Self-Driving Cars
Post by: phooey on July 01, 2016, 04:39:52 PM
Daniel,

I will tell you what I think is disgusting.  That people have been touting this technology way too soon.   

He was killed because they hadn't considered the effects of sunshine yet?  Are you kidding me.  They are going to get sued, and rightfully so, because if you haven't considered what sunlight will do to your cameras detection system, you have no business installing this in a car and telling people to use it.  If the system requires you to be sitting there with your hands hovering over the steering wheel, and your foot in the air above the brake, what is the point of it?  How much time do you need to give the system to know if it is going to do the right thing?  What kind of way is that to drive a car?

Now, for those of you who are going to say, well, that's just the cost of innovation, well, how many more of these law suits do you think Tesla would be able to stand, before they would decide, hey this isn't what we were bargaining for?  3? 5? 10?

Sunshine, are you fucking kidding me?  And you are going to call me disgusting?  For having sense? 
Title: Re: Self-Driving Cars
Post by: Morvis13 on July 01, 2016, 04:48:32 PM
Rolls Royce has a new driverless car that looks awful.

https://www.youtube.com/watch?v=dGSm082ELXY
Title: Re: Self-Driving Cars
Post by: SkeptiQueer on July 01, 2016, 04:55:16 PM
If the system requires you to be sitting there with your hands hovering over the steering wheel, and your foot in the air above the brake, what is the point of it?

Selling expensive cars to people to finance the next generation of research. The car isn't supposed to driver for you yet, it's supposed to help you drive better.
Title: Re: Self-Driving Cars
Post by: teethering on July 01, 2016, 05:03:03 PM
According to the information so far, the driver didn't see the truck.  It's a bit hard to believe, unless he was asleep or not paying any attention to the road.  However, if the information proves accurate and he didn't see because he had the sun in his eyes, then with or without the autopilot this would have happened.  If it is, in fact, that he was relying on technology and either watching a movie or asleep, then he was using the technology inappropriately, as according to the company it's not ready yet for fully autonomous driving.

What exactly is this proving, phooey?
Title: Re: Self-Driving Cars
Post by: phooey on July 01, 2016, 05:17:50 PM
Its proving that if you have a system called autopilot, and you tell people it can be used, there is no way they are going to be paying as close attention to the situation, as they would if they were driving the car all themselves.

That is the point.  If the system requires you to put in the same amount of mental energy as you would without it, then it is pointless.  In fact, its worse than pointless, its extra dangerous, because now you have two voices, the computers and your own telling you how to react in any situation.  Imagine if you had the passenger always telling you what maneuver to make during a speed through traffic.  That would be insane.  Driving a car, with your hands hovering over the steering wheel, waiting to see if you are needed?  That is no way to pilot a car. 
Title: Re: Self-Driving Cars
Post by: Fast Eddie B on July 01, 2016, 05:32:36 PM

I will tell you what I think is disgusting...

No, dancing on the grave of someone tragically killed is what's disgusting, all to prove a point.

From Tesla:

"It is important to note that Tesla disables Autopilot by default and requires explicit acknowledgement that the system is new technology and still in a public beta phase before it can be enabled. When drivers activate Autopilot, the acknowledgment box explains, among other things, that Autopilot “is an assist feature that requires you to keep your hands on the steering wheel at all times," and that "you need to maintain control and responsibility for your vehicle” while using it. Additionally, every time that Autopilot is engaged, the car reminds the driver to “Always keep your hands on the wheel. Be prepared to take over at any time.” The system also makes frequent checks to ensure that the driver's hands remain on the wheel and provides visual and audible alerts if hands-on is not detected. It then gradually slows down the car until hands-on is detected again."

Source: https://www.teslamotors.com/blog/tragic-loss (https://www.teslamotors.com/blog/tragic-loss)

Worth a read in its entirety.
Title: Re: Self-Driving Cars
Post by: Jeremy's Sea on July 01, 2016, 06:16:58 PM
Its proving that if you have a system called autopilot, and you tell people it can be used, there is no way they are going to be paying as close attention to the situation, as they would if they were driving the car all themselves.

That is the point.  If the system requires you to put in the same amount of mental energy as you would without it, then it is pointless.  In fact, its worse than pointless, its extra dangerous, because now you have two voices, the computers and your own telling you how to react in any situation.  Imagine if you had the passenger always telling you what maneuver to make during a speed through traffic.  That would be insane.  Driving a car, with your hands hovering over the steering wheel, waiting to see if you are needed?  That is no way to pilot a car.
lol
99.999% of people manage to follow the directions on blenders and not remove their fingers violently every day. Hopefully your family has removed all sharp objects from your home, you seem the type to need the help.
Title: Re: Self-Driving Cars
Post by: kvuo75 on July 01, 2016, 08:09:49 PM

That is the point.  If the system requires you to put in the same amount of mental energy as you would without it, then it is pointless.

do you think airliner autopilots are pointless also?

you know, you can be using that mental energy on something else rather than basic control of the vehicle... like looking for other traffic. etc.
Title: Re: Self-Driving Cars
Post by: daniel1948 on July 01, 2016, 08:44:52 PM
Daniel,

I will tell you what I think is disgusting.  That people have been touting this technology way too soon.   

He was killed because they hadn't considered the effects of sunshine yet?  Are you kidding me.  They are going to get sued, and rightfully so, because if you haven't considered what sunhlight will do to your cameras detection system, you have no business installing this in a car and telling people to use it.  If the system requires you to be sitting there with your hands hovering over the steering wheel, and your foot in the air above the brake, what is the point of it?  How much time do you need to give the system to know if it is going to do the right thing?  What kind of way is that to drive a car?

Now, for those of you who are going to say, well, that's just the cost of innovation, well, how many more of these law suits do you think Tesla would be able to stand, before they would decide, hey this isn't what we were bargaining for?  3? 5? 10?

Sunshine, are you fucking kidding me?  And you are going to call me disgusting?  For having sense? 

Phooey:

The driver was blinded by the sun and didn't see the truck. In ANY OTHER CAR he would have been killed. The Tesla system WHICH IS NOT A SELF-DRIVING SYSTEM also didn't see the truck for the same reason.

You shamelessly bring up this tragic death to make a point which has NOTHING to do with this case. Tesla makes it abundantly clear that the car is not self driving, it merely detects SOME hazards that the driver might not. It does not claim to detect all hazards and does not claim to be able to prevent all accidents. That would be fantasy.

The Tesla system makes the car safer, but does not make it perfect. I don't expect you to understand this, because you are so obsessed with your tirade against self-driving cars that you are happy to exploit a tragic death even when it has nothing whatsoever to do with the object of your obsession, but accidents will happen with any technology. If you are going to use a car you will be safer with the latest and best technology, but perfect does not exist. This was a case where NEITHER the driver nor the car were able to avoid the accident.

You seem to think that self-driving cars and driver-assist technologies should not be permitted unless they can prevent 100% of all accidents. You'd rather have thirty-two thousand deaths a year than allow technologies that might cut that number in half, because for you, one death is too many if the car has technology, but thirty-two thousand deaths are acceptable as long as it's human drivers causing the deaths.
Title: Re: Self-Driving Cars
Post by: brilligtove on July 01, 2016, 08:48:11 PM
Daniel,

I will tell you what I think is disgusting.  That people have been touting this technology way too soon.   

He was killed because they hadn't considered the effects of sunshine yet?  Are you kidding me.  They are going to get sued, and rightfully so, because if you haven't considered what sunhlight will do to your cameras detection system, you have no business installing this in a car and telling people to use it.  If the system requires you to be sitting there with your hands hovering over the steering wheel, and your foot in the air above the brake, what is the point of it?  How much time do you need to give the system to know if it is going to do the right thing?  What kind of way is that to drive a car?

Now, for those of you who are going to say, well, that's just the cost of innovation, well, how many more of these law suits do you think Tesla would be able to stand, before they would decide, hey this isn't what we were bargaining for?  3? 5? 10?

Sunshine, are you fucking kidding me?  And you are going to call me disgusting?  For having sense? 

Phooey:

The driver was blinded by the sun and didn't see the truck. In ANY OTHER CAR he would have been killed. The Tesla system WHICH IS NOT A SELF-DRIVING SYSTEM also didn't see the truck for the same reason.

You shamelessly bring up this tragic death to make a point which has NOTHING to do with this case. Tesla makes it abundantly clear that the car is not self driving, it merely detects SOME hazards that the driver might not. It does not claim to detect all hazards and does not claim to be able to prevent all accidents. That would be fantasy.

The Tesla system makes the car safer, but does not make it perfect. I don't expect you to understand this, because you are so obsessed with your tirade against self-driving cars that you are happy to exploit a tragic death even when it has nothing whatsoever to do with the object of your obsession, but accidents will happen with any technology. If you are going to use a car you will be safer with the latest and best technology, but perfect does not exist. This was a case where NEITHER the driver nor the car were able to avoid the accident.

You seem to think that self-driving cars and driver-assist technologies should not be permitted unless they can prevent 100% of all accidents. You'd rather have thirty-two thousand deaths a year than allow technologies that might cut that number in half, because for you, one death is too many if the car has technology, but thirty-two thousand deaths are acceptable as long as it's human drivers causing the deaths.
Clearly. Because reasons. Duh.

Title: Re: Self-Driving Cars
Post by: phooey on July 01, 2016, 10:54:12 PM
The company that makes the cameras for Tesla released a statement, now saying that the cameras are only able to avoid rear end collisions, not lateral ones.

Gee, I didn't read all the press stories about this BEFORE the accident. 
Title: Re: Self-Driving Cars
Post by: SkeptiQueer on July 02, 2016, 12:18:32 AM
Nuclear bombs shouldn't have safeties, it'll just make people complacent.
Title: Re: Self-Driving Cars
Post by: Andrew Clunn on July 02, 2016, 12:28:35 AM
So long as these AIs are closed source, each car company will need to have its own tragedies to learn about edge cases.
Title: Re: Self-Driving Cars
Post by: phooey on July 02, 2016, 01:32:40 AM
I suppose it depends on how much those tragedies cost them. 

I can't imagine a company like BMW will be in too rush to put a self-driving car into production, knowing that a half dozen accidents could probably put them out of business. 
Title: Re: Self-Driving Cars
Post by: SkeptiQueer on July 02, 2016, 02:41:37 AM
That's surprising, you've never lacked imagination for anything else.
Title: Re: Self-Driving Cars
Post by: bimble on July 02, 2016, 09:34:52 AM

I can't imagine a company like BMW will be in too rush to put a self-driving car into production, knowing that a half dozen accidents could probably put them out of business.

2021 they're hoping to have a fully self-driving car... https://www.wired.com/2016/07/bmws-bold-plan-make-fully-self-driving-car-2021/ (https://www.wired.com/2016/07/bmws-bold-plan-make-fully-self-driving-car-2021/)
Title: Re: Self-Driving Cars
Post by: kvuo75 on July 02, 2016, 05:51:08 PM
https://www.youtube.com/watch?v=WsnKzK6dX8Q
Title: Re: Self-Driving Cars
Post by: phooey on July 02, 2016, 08:08:17 PM
"In addition to its cameras, the Model S has radar sensors that could have spotted the trailer. But Tesla Motors CEO Elon Musk wrote on Twitter that the radar “tunes out what looks like an overhead road sign to avoid false braking events.”


It keeps getting worse. 
Title: Re: Self-Driving Cars
Post by: phooey on July 06, 2016, 05:46:55 AM
Another Tesla crash.

What do you reckon is the number they can afford?


http://www.huffingtonpost.com/entry/tesla-crash-autopilot-mode_us_577c2b55e4b0416464110cbc?section=
Title: Re: Self-Driving Cars
Post by: Alex Simmons on July 06, 2016, 06:02:25 AM
This thread needs a bit of super pursuit mode....

https://www.youtube.com/watch?v=v8dWW9KxqlQ
Title: Re: Self-Driving Cars
Post by: arthwollipot on July 06, 2016, 06:17:52 AM
He was killed because they hadn't considered the effects of sunshine yet?

No, he was killed because he wasn't following simple instructions.
Title: Re: Self-Driving Cars
Post by: SkeptiQueer on July 06, 2016, 06:31:40 AM
Another Tesla crash.

What do you reckon is the number they can afford?


http://www.huffingtonpost.com/entry/tesla-crash-autopilot-mode_us_577c2b55e4b0416464110cbc?section=
Considering they likely can't be sued for the first one, and probably this one will be the same, I'll wager this isn't hurting them at all financially.
Title: Re: Self-Driving Cars
Post by: Fast Eddie B on July 06, 2016, 07:09:00 AM
Considering they likely can't be sued for the first one, and probably this one will be the same, I'll wager this isn't hurting them at all financially.

Perhaps they can't be successfully sued, but what would stop someone from bringing a suit?

Even suits without merit can be a huge financial burden for a company.
Title: Re: Self-Driving Cars
Post by: SkeptiQueer on July 06, 2016, 07:17:28 AM
Considering they likely can't be sued for the first one, and probably this one will be the same, I'll wager this isn't hurting them at all financially.

Perhaps they can't be successfully sued, but what would stop someone from bringing a suit?

Even suits without merit can be a huge financial burden for a company.
Nothing, but they almost certainly have enough money set aside to cover a few cases. You could try to sue Makita because you hurt yourself after deactivating the safety on your circular saw, but your case will end pretty quickly when the judge sees that you intentionally and willfully deactivated a safety feature and then didn't follow the instructions and did a dangerous thing.

I don't believe anyone could sue a manufacturer because they took their hands off the wheel and crashed.
Title: Re: Self-Driving Cars
Post by: daniel1948 on July 06, 2016, 09:09:36 AM
Another Tesla crash.

What do you reckon is the number they can afford?

http://www.huffingtonpost.com/entry/tesla-crash-autopilot-mode_us_577c2b55e4b0416464110cbc?section=

Do you post online every time a car crashes? No, I don't suppose so. Nobody ever said that autopilot would prevent any crashes from ever happening. It's an aid, not a magical charm, and it's an imperfect aid. But we already understand that you oppose any technology to reduce accidents. I'll bet you really hate air bags because people can still die in a car with air bags. Oh, and you must believe that brakes on cars are the absolute worst thing ever, since sometimes brakes don't stop a car in time to prevent a crash.
Title: Re: Self-Driving Cars
Post by: phooey on July 06, 2016, 09:46:04 AM
He was killed because they hadn't considered the effects of sunshine yet?

No, he was killed because he wasn't following simple instructions.

The simple instructions?  You mean like "When using the autopilot function, hover your foot directly above either the gas pedal or the brake (you decide) and keep your hands hovering over the steering wheel, ready to turn aggressively left or right (you decide) in an instant, if you feel the computer suddenly doesn't know what the fuck it is doing. Make that decision quickly whenever you are sure or not sure that the computer doesn't know what its doing, because we don't know when that instant will be, its not our problem." ?

That instruction? 
Title: Re: Self-Driving Cars
Post by: Jeremy's Sea on July 06, 2016, 01:52:37 PM
Another Tesla crash.

What do you reckon is the number they can afford?

http://www.huffingtonpost.com/entry/tesla-crash-autopilot-mode_us_577c2b55e4b0416464110cbc?section=

Do you post online every time a car crashes? No, I don't suppose so. Nobody ever said that autopilot would prevent any crashes from ever happening. It's an aid, not a magical charm, and it's an imperfect aid. But we already understand that you oppose any technology to reduce accidents. I'll bet you really hate air bags because people can still die in a car with air bags. Oh, and you must believe that brakes on cars are the absolute worst thing ever, since sometimes brakes don't stop a car in time to prevent a crash.

It's also worth noting-
Quote
...and the motorist told police he had activated the car’s autopilot feature...
I would rather blame the autopilot than my own idiocy too. Let's see if they can show he was actually using the feature, let alone using it properly.
But the goal posts have moved anyway in the conversation, it began that we will never have driverless cars, and now it's about how Tesla's autopilot system isn't able to replace human drivers, when (despite how stupid you have to be to ignore the actual instructions because it says autopilot which you may or may not assume means you can take your attention away from the road) it doesn't claim to.
Title: Re: Self-Driving Cars
Post by: phooey on July 06, 2016, 06:41:35 PM
Computer says: "Pay attention, pay attention, but I got this don't worry.  I got this, don't freak out, I see that car, I am moving over to avoid it, ..I got this, I am passing, I got this, chill out, I am just going to ..WAIT, Fuck I don't got this! You take it!" 

Some people just don't follow directions well I guess.  Tesla has nothing to worry about. 
Title: Re: Self-Driving Cars
Post by: TheIrreverend on July 06, 2016, 07:13:17 PM
(http://i.imgur.com/QM2CJDn.png)
Title: Re: Self-Driving Cars
Post by: daniel1948 on July 06, 2016, 09:41:04 PM
Computer says: "Pay attention, pay attention, but I got this don't worry.  I got this, don't freak out, I see that car, I am moving over to avoid it, ..I got this, I am passing, I got this, chill out, I am just going to ..WAIT, Fuck I don't got this! You take it!" 

Some people just don't follow directions well I guess.  Tesla has nothing to worry about. 

You are simply being dishonest. The computer does NOT say "I got this don't worry." The computer says, "keep your eyes on the road and your hands on the wheel, and together, with BOTH OF US paying attention, we will avoid a few accidents that either of us alone would fail to avoid."

When people refuse to understand that the so-called "auto-pilot" is an assist, not a self-driving system, the outcomes will be sub-optimal.

And the Model S is still the safest car on the road.
Title: Re: Self-Driving Cars
Post by: phooey on July 07, 2016, 03:13:07 AM
In almost no way at all is that better than simply driving the car yourself. 

In fact it is clearly much worse. 
Title: Re: Self-Driving Cars
Post by: Jeremy's Sea on July 07, 2016, 03:25:04 AM
In almost no way at all is that better than simply driving the car yourself. 

In fact it is clearly much worse.
Clearly, because in two documented cases something went wrong (when in one, it was complete negligence, and the other we don't have details yet). Troll.
Title: Re: Self-Driving Cars
Post by: arthwollipot on July 07, 2016, 07:21:44 AM
He was killed because they hadn't considered the effects of sunshine yet?

No, he was killed because he wasn't following simple instructions.

The simple instructions?  You mean like "When using the autopilot function, hover your foot directly above either the gas pedal or the brake (you decide) and keep your hands hovering over the steering wheel, ready to turn aggressively left or right (you decide) in an instant, if you feel the computer suddenly doesn't know what the fuck it is doing. Make that decision quickly whenever you are sure or not sure that the computer doesn't know what its doing, because we don't know when that instant will be, its not our problem." ?

That instruction? 

Would you fix that quote, please?
Title: Re: Self-Driving Cars
Post by: daniel1948 on July 07, 2016, 09:00:31 AM
In almost no way at all is that better than simply driving the car yourself. 

In fact it is clearly much worse. 

No. It is clearly much better, because it adds a level of protection. If you ignore all common sense and treat auto-pilot as if it were an autonomous driver, then it is worse. But if you use it according to the instructions, driving as you normally would and allowing the system to react to hazards that you fail to see, then it is clearly safer.

Note: "Better" does not mean perfect. "Better" does not mean that there will never be accidents. "Better" means that there will be fewer accidents.
Title: Re: Self-Driving Cars
Post by: phooey on July 07, 2016, 09:35:50 AM
I am sure no one here thinks that if you were the system designer, you wouldn't have forgotten to take the affects of sunshine into account. 

But Musk is perhaps a bit more busy than you. 
Title: Re: Self-Driving Cars
Post by: phooey on July 07, 2016, 09:43:58 AM
Daniel,

What evidence is there that it is better?

Now, if of course you are talking about warning systems, or even an emergency braking system, and not something that takes over any control of the vehicle, then that is totally different right? 
Title: Re: Self-Driving Cars
Post by: phooey on July 07, 2016, 09:45:06 AM
Perhaps I should get a job writing for Forbes.  Then I can write about significant things BEFORE they happen, like I did here, rather than wait until they do. 


http://www.forbes.com/sites/jimgorzelany/2016/07/05/maybe-its-time-to-curb-our-unbridled-enthusiasm-for-self-driving-cars/?utm_source=yahoo&utm_medium=partner&utm_campaign=yahootix&partner=yahootix&yptr=yahoo#215ed13b295d
Title: Re: Self-Driving Cars
Post by: Jeremy's Sea on July 07, 2016, 11:40:59 AM
Perhaps I should get a job writing for Forbes.  Then I can write about significant things BEFORE they happen, like I did here, rather than wait until they do. 


http://www.forbes.com/sites/jimgorzelany/2016/07/05/maybe-its-time-to-curb-our-unbridled-enthusiasm-for-self-driving-cars/?utm_source=yahoo&utm_medium=partner&utm_campaign=yahootix&partner=yahootix&yptr=yahoo#215ed13b295d
hahaha you literally predicted nothing here. You made a cynical blanket statement and waited for something to happen to confirm your cynicism.
It's like predicting the next mass shooting will happen in the next 6 months. Put people in large vehicles traveling at high velocities and someone will die and a computer can't 100% stop that. Viola. I'm a Forbes writing genius!
Title: Re: Self-Driving Cars
Post by: Fast Eddie B on July 07, 2016, 11:48:57 AM
In almost no way at all is that better than simply driving the car yourself. 

In fact it is clearly much worse.

I suggest you look up what typical human reaction time is.

Then compare it to typical machine reaction time.

Then tell us if you wish to reconsider your position.
Title: Re: Self-Driving Cars
Post by: PANTS! on July 07, 2016, 12:12:11 PM
What evidence is there that it is better?

A significant decrease in accidents per mile driven.
Title: Re: Self-Driving Cars
Post by: teethering on July 07, 2016, 01:05:56 PM
Alright, so setting aside the idiotic, ignorant spew that's coming from phooey, there is, underneath all the blathering, a kernel of a point that requires serious consideration.

Autopilot as Tesla currently has it is asking for trouble.  I'll re-emphasize, as they currently have it.  There's a problem, even if you put all the disclaimers you want, enabling it at your own risk etc etc, in allowing someone to enable a mode in their car that allows them to not pay attention most of the time and then expect them to be alert for a few critical moments.  Half-way autopilot is a bad idea, because humans don't work that way.  Fully automated vehicles are undoubtedly safer than human drivers, and computer assisted driving is clearly a win, but at some point in between on that spectrum you're in trouble.  I think Tesla should not ship with autopilot until it's able to fully take over the driving, which I think is probably only a few years away.
Title: Re: Self-Driving Cars
Post by: Johnny Slick on July 07, 2016, 01:08:15 PM
Is it *that* much different from cruise control, though? I think I can get behind the idea of having some kind of autopilot for driving on the freeway for long stretches with the caveat that you flat out don't use it while not on the interstate.
Title: Re: Self-Driving Cars
Post by: teethering on July 07, 2016, 01:45:45 PM
Is it *that* much different from cruise control, though? I think I can get behind the idea of having some kind of autopilot for driving on the freeway for long stretches with the caveat that you flat out don't use it while not on the interstate.

On a freeway with controlled access it's fine, but otherwise it's dangerous.  If the car can handle all but very uncommon contingencies, it's fine, but if you need the driver to take over in an emergency that's dangerous.  For what it's worth, this is what NHTSA also says.
Title: Re: Self-Driving Cars
Post by: daniel1948 on July 07, 2016, 03:23:24 PM
If Tesla-style auto-pilot prevents 100 accidents, but lulls a few people into complacency and causes 3 accidents, then it's a big win. But those 100 prevented accidents get no attention, while the three accidents get covered in every news outlet for a week straight, until something more sensational happens, and all the ranting neo-Luddites have a field day.

Tesla has done it exactly right. And just as you cannot prevent some people from driving drunk, so also you cannot prevent a few people from driving recklessly. I knew a guy who died after he and his pals decided to let some air out of their car's tires and drive on the railroad tracks. You will never convince everybody to use auto-pilot responsibly, but it still reduces the total number of accidents.

Some people, like Phooey, don't care how many people die as the result of human carelessness; they hate and/or fear technology so much that if it does not eliminate ALL accidents it's unacceptable to them. They think it's better for 100 people to die by human action than for one to die because of the failure of a machine. And Phooey's personal attacks on Elon Musk demonstrate how unhinged he has become over this issue.
Title: Re: Self-Driving Cars
Post by: PANTS! on July 07, 2016, 03:24:27 PM
Is it *that* much different from cruise control, though? I think I can get behind the idea of having some kind of autopilot for driving on the freeway for long stretches with the caveat that you flat out don't use it while not on the interstate.

On a freeway with controlled access it's fine, but otherwise it's dangerous.  If the car can handle all but very uncommon contingencies, it's fine, but if you need the driver to take over in an emergency that's dangerous.  For what it's worth, this is what NHTSA also says.

Yes, this is exactly the issue planes are having with auto pilot right now.
Title: Re: Self-Driving Cars
Post by: brilligtove on August 15, 2016, 01:44:42 PM
(http://imgs.xkcd.com/comics/horses.png) (http://xkcd.com/1720/)
Title: Re: Self-Driving Cars
Post by: Desert Fox on August 16, 2016, 07:27:22 AM
I would like to have a self driving car built that looks like an 80's Trans-Am with a red light in the front.  >:D

There was a morning radio show asking what car that people would most want from TV or movies, expecting I think for everybody to say the General Lee but KIT was by far the most popular. The idea of something being able to drive you to work and home is very compelling even though I do sometimes enjoy driving.
Title: Re: Self-Driving Cars
Post by: Alex Simmons on August 16, 2016, 08:05:07 AM
I would like to have a self driving car built that looks like an 80's Trans-Am with a red light in the front.  >:D

There was a morning radio show asking what car that people would most want from TV or movies, expecting I think for everybody to say the General Lee but KIT was by far the most popular. The idea of something being able to drive you to work and home is very compelling even though I do sometimes enjoy driving.
http://sguforums.com/index.php/topic,44070.msg9423094.html#msg9423094
Title: Re: Self-Driving Cars
Post by: seamas on August 16, 2016, 11:04:17 AM
I would like to have a self driving car built that looks like an 80's Trans-Am with a red light in the front.  >:D

There was a morning radio show asking what car that people would most want from TV or movies, expecting I think for everybody to say the General Lee but KIT was by far the most popular. The idea of something being able to drive you to work and home is very compelling even though I do sometimes enjoy driving.

Did people really forget the Batmobile ('60s one)????


For shame.



The Jag/ Hearse from Harold & Maude was also a favorite.

I can think of at least a dozen cars from TV and movies I would want more than the General Lee or KITT .

Title: Re: Self-Driving Cars
Post by: Friendly Angel on August 16, 2016, 11:57:18 AM
I can think of at least a dozen cars from TV and movies I would want more than the General Lee or KITT .

word


(http://www.oscarmayer.com/~/media/OscarMayer-New/images/wienermobile/wienermobile_billboard_lg.jpg)
Title: Re: Self-Driving Cars
Post by: stonesean on August 16, 2016, 12:52:16 PM
I can think of at least a dozen cars from TV and movies I would want more than the General Lee or KITT .


Agreed.

(http://media-cache-ak0.pinimg.com/736x/27/5a/f2/275af2707f319e8ebf852e2a2606e49e.jpg)
Title: Re: Self-Driving Cars
Post by: Johnny Slick on August 16, 2016, 01:45:52 PM
Um, lest we forget, KITT has Principal Belding inside of it.
Title: Re: Self-Driving Cars
Post by: seamas on August 16, 2016, 03:08:19 PM
Um, lest we forget, KITT has Principal Belding inside of it.

Hmm, I always thought it was either Dr. Mark Craig or Mr. Braddock.
Title: Re: Self-Driving Cars
Post by: Shibboleth on August 16, 2016, 03:24:24 PM


I can think of at least a dozen cars from TV and movies I would want more than the General Lee or KITT .

(http://harrisonboyce.com/blog/wp-content/uploads/2009/08/le_mans_movie_stills_01.JPG)
Title: Re: Self-Driving Cars
Post by: daniel1948 on August 16, 2016, 10:35:04 PM
I can think of at least a dozen cars from TV and movies I would want more than the General Lee or KITT .

word


(http://www.oscarmayer.com/~/media/OscarMayer-New/images/wienermobile/wienermobile_billboard_lg.jpg)

I actually saw that thing once, as a kid in L.A. Or, perhaps there are many of them and I saw one. I don't remember the car body. I have no memory of what it was mounted on. But the big wiener I remember. What a disgusting monstrosity!
Title: Re: Self-Driving Cars
Post by: Alex Simmons on August 17, 2016, 02:51:08 AM
There was always Herbie...
Title: Re: Self-Driving Cars
Post by: Shibboleth on August 17, 2016, 10:43:08 AM
There was always Herbie...

No there wasn't liar. There was a vast period of time before Herbie. I am sick of the Herbie truthers coming here and claiming he always existed.
Title: Re: Self-Driving Cars
Post by: The Latinist on August 17, 2016, 12:53:40 PM
Ford says they're aiming to have a fully autonomous car on the market in 2021.
Title: Re: Self-Driving Cars
Post by: HighPockets on August 17, 2016, 01:22:05 PM


I can think of at least a dozen cars from TV and movies I would want more than the General Lee or KITT .