Author Topic: Episode #714  (Read 1334 times)

0 Members and 1 Guest are viewing this topic.

Offline lucek

  • Not Enough Spare Time
  • **
  • Posts: 109
Re: Episode #714
« Reply #30 on: March 19, 2019, 05:36:24 AM »
The issue will not be primarily with desktop computers or even servers; it will be with embedded systems, many of which indeed use older, cheaper 32-but processors. Many routers, switches, etc. are of this type and may be impossible to upgrade.  And I don’t think Lucek is saying that the 2038 problem will cause the apocalypse; merely that it will require investment to remediate (indeed, many people have already spent many hours bringing 64-bit timecodes into common usage and even backporting them to 32-bit systems).  Like Y2K, it is a real issue that will cost a lot of money to fix, but which we have every reason to believe will, in fact, be fixed.

Also?  It’s not just a hardware issue; SQL, for instance uses 32-bit time codes and there is no way to remedy that without breaking compatibility with a great deal of existing code. As far as I’m aware, a solution to that has not yet been found.
Spot on. I was also drawing a parallel between the 2 because like in the early days of the 90s and before and now the idea that "it'll just get obsoleted and replaced so no worries right?" A lot of people are talking about it so we don't have a repeat of the mad rush that happened in 98 and 99. If we solve it over 19 years it's trivial. If we wait for 19 months it's difficult. If we wait for 19 days it's impossible. So the wise thing is to put the spotlight on it now and do the investment slowly.
You have the power, but. . .
Power is just energy over time and. . .
Energy is just the ability to do work.

Offline lucek

  • Not Enough Spare Time
  • **
  • Posts: 109
Re: Episode #714
« Reply #31 on: March 19, 2019, 05:39:11 AM »
Oh is no one on here interested in AI and redundant neural networks? I figured that would be the part of my post i'd be fielding questions about.
You have the power, but. . .
Power is just energy over time and. . .
Energy is just the ability to do work.

Offline brilligtove

  • Too Much Spare Time
  • ********
  • Posts: 6954
  • Ignorance can be cured. Stupidity, you deal with.
Re: Episode #714
« Reply #32 on: March 19, 2019, 08:13:01 AM »
Oh is no one on here interested in AI and redundant neural networks? I figured that would be the part of my post i'd be fielding questions about.

I thought your explanation was sensible and clear at that level of detail. Have you seen the 3blue1brown videos on neural nets?
evidence trumps experience | performance over perfection | responsibility – authority = scapegoat | emotions motivate; data doesn't

Offline gebobs

  • Seasoned Contributor
  • ****
  • Posts: 535
  • Me like hockey!
Re: Episode #714
« Reply #33 on: March 19, 2019, 09:24:08 AM »
I’m doing a little experiment with beta-carotene.  I love carrot juice.  I’m eating 1000 g of carrots a day, which is providing over 9 times the RDA.  No side effects so far (the only ones are caretenosis - an orangecolouration - and increased lung cancer in ex-smokers).

Thank you for your service, Mr. President!

Online Ron Obvious

  • Not Enough Spare Time
  • **
  • Posts: 189
Re: Episode #714
« Reply #34 on: March 19, 2019, 05:44:31 PM »
One last thing we are looking at Y2Ks big brother in just under 19 years. The unix timecode rollover is going to happen on January 19 2038.

Do you really think there will be a lot of legacy 32 bit systems with a 32bit time_t around at that time? 64 bit time_t will be more than sufficient for the whole of our civilisation even if we're extremely optimistic as to how long it will last.
Did you even listen to the interview on Y2K? Somebody will fix it was the mentality in the 70s and 80s.

Actually, I don't remember anybody talking about Y2K in the 70s and 80s, and I was in the industry at that time. Partially, that's because presumably only COBOL programmers were daft enough to use separate day, month, year (2 digits) fields for their dates, as those are a massive pain to operate on.  Most of us stored data as integer seconds even back then, and had already realised that 2038 was the real problem date, but since we could hardly imagine what the world of 2038 would look like (other than the flying cars, which were a given), nobody really worried about it.

For perspective, 19 years before Y2K was 1981.

Online daniel1948

  • Isn’t a
  • Reef Tank Owner
  • *********
  • Posts: 8078
  • I'd rather be paddling
Re: Episode #714
« Reply #35 on: March 19, 2019, 07:49:15 PM »
I got SoF right! That happens about one time out of ten. Yes, significantly worse than chance.

(click to show/hide)

The whole point of the item on golden rice was that the target populations eat almost nothing but rice. They can have a very low conversion rate and still get enough vitamin A to prevent blindness. And since there are no down-sides to golden rice, the only reason to oppose it is a quasi-religious fanatacism against GMOs, so powerful that saving a billion children from blindness seems insignificant. Even if you only save half, there are absolutely no down-sides to golden rice.

My prediction is that if it works, then we won’t hear anything from the anti-GMO ideologues.  If it doesn’t work, which hasn’t been excluded apparently since there’s been no studies of it in its target population (poor rural Bangladeshi), we’ll never hear the end of it.

I’m doing a little experiment with beta-carotene.  I love carrot juice.  I’m eating 1000 g of carrots a day, which is providing over 9 times the RDA.  No side effects so far (the only ones are caretenosis - an orangecolouration - and increased lung cancer in ex-smokers).

I had caretenosis once, when I was eating a huge amount of carrots. My doctor told me to cut down on the carrots and I did. I do love carrots.

My prediction: If golden rice significantly reduces blindness in kids in Bangladesh, the anti-GMO crowd will lie, and claim it had no effect on blindness but caused a drastic rise in autism.
Daniel
----------------
"Anyone who has ever looked into the glazed eyes of a soldier dying on the battlefield will think long and hard before starting a war."
-- Otto von Bismarck

Offline DevoutCatalyst

  • Well Established
  • *****
  • Posts: 1441
Re: Episode #714
« Reply #36 on: March 19, 2019, 08:25:23 PM »
I got SoF right! That happens about one time out of ten. Yes, significantly worse than chance.

(click to show/hide)

The whole point of the item on golden rice was that the target populations eat almost nothing but rice. They can have a very low conversion rate and still get enough vitamin A to prevent blindness. And since there are no down-sides to golden rice, the only reason to oppose it is a quasi-religious fanatacism against GMOs, so powerful that saving a billion children from blindness seems insignificant. Even if you only save half, there are absolutely no down-sides to golden rice.

My prediction is that if it works, then we won’t hear anything from the anti-GMO ideologues.  If it doesn’t work, which hasn’t been excluded apparently since there’s been no studies of it in its target population (poor rural Bangladeshi), we’ll never hear the end of it.

I’m doing a little experiment with beta-carotene.  I love carrot juice.  I’m eating 1000 g of carrots a day, which is providing over 9 times the RDA.  No side effects so far (the only ones are caretenosis - an orangecolouration - and increased lung cancer in ex-smokers).

I had caretenosis once, when I was eating a huge amount of carrots. My doctor told me to cut down on the carrots and I did. I do love carrots.

My prediction: If golden rice significantly reduces blindness in kids in Bangladesh, the anti-GMO crowd will lie, and claim it had no effect on blindness but caused a drastic rise in autism.

The sweet potato is my staple food. I changed to a yellow fleshed variety because the orange coloured ones gave my skin that weird orange hue. The carrot enthusiast may not have to cut back.


Offline Soldier of FORTRAN

  • Reef Tank Owner
  • *********
  • Posts: 8870
  • Cache rules everything around me.
Re: Episode #714
« Reply #37 on: March 19, 2019, 08:27:06 PM »
Ooooh
If global warming is real then how come I just felt this chill down my spine?

Offline Billzbub

  • Stopped Going Outside
  • *******
  • Posts: 4149
  • I know you know I know
Re: Episode #714
« Reply #38 on: March 20, 2019, 10:12:06 AM »
I got SoF right! That happens about one time out of ten. Yes, significantly worse than chance.

(click to show/hide)

The whole point of the item on golden rice was that the target populations eat almost nothing but rice. They can have a very low conversion rate and still get enough vitamin A to prevent blindness. And since there are no down-sides to golden rice, the only reason to oppose it is a quasi-religious fanatacism against GMOs, so powerful that saving a billion children from blindness seems insignificant. Even if you only save half, there are absolutely no down-sides to golden rice.

My prediction is that if it works, then we won’t hear anything from the anti-GMO ideologues.  If it doesn’t work, which hasn’t been excluded apparently since there’s been no studies of it in its target population (poor rural Bangladeshi), we’ll never hear the end of it.

I’m doing a little experiment with beta-carotene.  I love carrot juice.  I’m eating 1000 g of carrots a day, which is providing over 9 times the RDA.  No side effects so far (the only ones are caretenosis - an orangecolouration - and increased lung cancer in ex-smokers).

I had caretenosis once, when I was eating a huge amount of carrots. My doctor told me to cut down on the carrots and I did. I do love carrots.

My prediction: If golden rice significantly reduces blindness in kids in Bangladesh, the anti-GMO crowd will lie, and claim it had no effect on blindness but caused a drastic rise in autism.

Yeah, once the rice if widely available, they will just gather all the data they can on Bangladesh and blame whatever negative trends happen on the rice.  Pornography views went up?  Rice!  Crime rate went up?  Rice!  Death rate dropped 200%?  Not rice!
Quote from: Steven Novella
gleefully altering one’s beliefs to accommodate new information should be a badge of honor

Online 2397

  • Frequent Poster
  • ******
  • Posts: 2363
Re: Episode #714
« Reply #39 on: March 20, 2019, 10:15:33 AM »
Death rate dropped 200%?  Not rice!

Turns out it was the pornography.

Offline lucek

  • Not Enough Spare Time
  • **
  • Posts: 109
Re: Episode #714
« Reply #40 on: March 20, 2019, 10:49:46 AM »
Oh is no one on here interested in AI and redundant neural networks? I figured that would be the part of my post i'd be fielding questions about.

I thought your explanation was sensible and clear at that level of detail. Have you seen the 3blue1brown videos on neural nets?
Yeah it is a good summation. Or I remember it being so. It's been a while.
You have the power, but. . .
Power is just energy over time and. . .
Energy is just the ability to do work.

Offline stochastibot

  • Brand New
  • Posts: 7
Re: Episode #714
« Reply #41 on: March 21, 2019, 12:08:58 PM »
I'm a postdoc at the Robotics Institute at Carnegie Mellon, and some of our lab's work involves explainability in AI, so it's especially awesome to listen to an episode with a story relevant to my research area (How Intelligent is AI).

There's been a handful of publications over the last few years on which features are salient for decision making.  What's pretty interesting about this paper is that the authors introduce a method for automatically figuring out when a classifier has an anomalous decision strategy, when looking at all decisions made in a dataset.  It'll be interesting to see where this approach goes, and if it might be incorporated as a learning heuristic. 

Anyway, glad the podcast was able to point me to a paper of interest!

Offline fuzzyMarmot

  • Seasoned Contributor
  • ****
  • Posts: 604
Re: Episode #714
« Reply #42 on: March 21, 2019, 03:41:34 PM »
Welcome to the forums! I love your username :) "Stochastic" was a recent word of the week on the show, so you are in the right place! The Rogues love AI, so perhaps you would be a great guest for the show.

 

personate-rain