Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

If AI scaling is to be shut down, let it be for a coherent reason

scottaaronson.blog

Post reply on HN
Print view View original

11–20 of 475 posts

Re: If AI scaling is to be shut down, let it be for a coherent reason

#11

Post by nl »

Why are we even bothering talking about this?

It's Gary Marcus "neural networks don't really work" suddenly discovering they do, and literally trying to shut down research in that area while keeping his prefered research areas funded.

We know a bunch of the people whose names are on the letter didn't sign it (eg Yann Lecunn who said he disagreed with the premise and didn't sign it).

I'm so offended by the idea of this that I'll personally fund $10k of training runs myself in a jurisdiction where it isn't banned if this ever became law in the US.

Re: If AI scaling is to be shut down, let it be for a coherent reason

#12

Post by nemo44x »

So what part exactaly needs banning? Transformer technology? The amount of data used as input to the program? Number of GPUs allowed?

Otherwise you’re just trying to ban a company that has been more innovative than anyone else. Why should they stop?

Any talk of a ban or limiting action needs to be specific. Meanwhile instead of fear mongering why not work on developing AI resistant tech?

Re: If AI scaling is to be shut down, let it be for a coherent reason

#13

Post by matthewdgreen »

cs702 wrote: post #2

> ... I was deeply confused, until I heard a dear friend and colleague in academic AI, one who’s long been skeptical of AI-doom scenarios, explain why he signed the open letter. He said: look, we all started writing research papers about the safety issues with ChatGPT; then our work became obsolete when OpenAI released GPT-4 just a few months later. So now we’re writing papers about GPT-4. Will we again have to throw…

I don’t think Scott is serious about that (or if he is, he’s being uncharitable.) I think what the quoted speaker is saying is that nobody is able to keep up with what these models are doing internally. Even OpenAI (and Meta et al.) only seem to be making “so much progress” by pressing the accelerator to the floor and letting the steering take care of itself. And one of the major lessons of technological progress is that deep understanding (at least when humans are necessary for that, gulp) is much slower than engineering, largely because the latter can be parallelized and scaled.

Re: If AI scaling is to be shut down, let it be for a coherent reason

#14

Post by computerex »

I read the times article, and I also watched Lex Fridman's talk with Eliezer Yudkowsky. Frankly I don't think he is qualified, and I don't understand why anyone is even giving him any credence. His argument is literally:

> A mis-aligned super AGI will result in the death of humanity

I ask, how. He is making reductive logical leaps that don't make sense. It's FUD.

Re: If AI scaling is to be shut down, let it be for a coherent reason

#15

Post by 1827162 »

lwhi wrote: post #8
1827162 wrote: post #6

If the government tries to prohibit training these models, I think we should find a way to keep it going somehow. Yes, civil disobedience.

Why?

Because the folks at the NSA are not going to stop doing so... The government itself is going to continue with it. And we don't want to allow the state to have the monopoly on advanced AI.

Re: If AI scaling is to be shut down, let it be for a coherent reason

#16

Post by lezojeda »

LegionMammal978 wrote: post #5
rvz wrote: post #3

> On the other hand, I’m deeply confused by the people who signed the open letter, even though they continue to downplay or even ridicule GPT’s abilities, as well as the “sensationalist” predictions of an AI apocalypse. Says the quantum computing professor turned so-called 'AI safety employee' at O̶p̶e̶n̶AI.com who would rather watch an unregulated hallucination-laden language model run off the rails to be sold as th…

> Once again, he hasn't disclosed that he is working for O̶p̶e̶n̶AI.com again. From the article: > ... and while I’ve been spending the year at OpenAI to work on theoretical foundations of AI safety, I’m going to answer strictly for myself. (Not to say that OpenAI's name isn't dumb, or that there won't be issues from people directly plugging LLMs into important decisions.)

[dead]

Re: If AI scaling is to be shut down, let it be for a coherent reason

#17

Post by realce »

lwhi wrote: post #8
1827162 wrote: post #6

If the government tries to prohibit training these models, I think we should find a way to keep it going somehow. Yes, civil disobedience.

Why?

Because the NSA isn't going to stop, the CCP isn't going to stop. Anyone who doesn't stop is a threat to my personal freedom, so the only logical reaction to me is to empower yourself as well as possible and keep training.

Re: If AI scaling is to be shut down, let it be for a coherent reason

#18

Post by 1827162 »

1827162 wrote: post #15
lwhi wrote: post #8

Earlier quoted context omitted.

Why?

Because the folks at the NSA are not going to stop doing so... The government itself is going to continue with it. And we don't want to allow the state to have the monopoly on advanced AI.

Some kind of distributed training, BitTorrent style would be one way of getting around it, using thousands of GPUs worldwide? If we could somehow make the training process profitable, like a cryptocurrency, then that would be even better.

Re: If AI scaling is to be shut down, let it be for a coherent reason

#19

Post by realce »

1827162 wrote: post #15
lwhi wrote: post #8

Earlier quoted context omitted.

Why?

Because the folks at the NSA are not going to stop doing so... The government itself is going to continue with it. And we don't want to allow the state to have the monopoly on advanced AI.

Ha we replied the same thing at the same time - great minds!

Re: If AI scaling is to be shut down, let it be for a coherent reason

#20

Post by dandellion »

Havoc wrote: post #10

This whole thing is like leaving 20 kids alone in a room with a marshmallow each and telling them don't eat it. ...and then expecting all of them resist. The debate around whether we should tell the kids not to eat it and for what reasons is completely academic. Practically this just isn't happening.

If they're really so worried about it they should start raiding data centres yesterday, writing open letters is such an obvious waste of time that I have a hard time taking it seriously.
Post reply on HN