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

Rodney Brooks on GPT-4

spectrum.ieee.org

Post reply on HN
Print view View original

81–90 of 412 posts

Re: Rodney Brooks on GPT-4

#81

Post by Gigachad »

jstx1 wrote: post #63
Gigachad wrote: post #25

Earlier quoted context omitted.

Feels like there have been a lot of previous technologies where the last 10% was far more complex than the first 90%. Self driving cars were pretty much solved a decade ago, and yet we still aren't there yet. VR was pretty much working and ready to change the world a decade ago, and we still aren't there yet. So it's hard to tell if this is an iphone moment where it just rockets off in to space and changes the world.…

The difference is that GPT-4 is already useful even in its current imperfect form. It’s not a 90% thing that needs to get to 100% for us to use it.

VR is useful in its current form. Self driving tech is being used for lots of assistance feature. They just haven’t changed the world yet.

Re: Rodney Brooks on GPT-4

#82

Post by p-e-w »

zone411 wrote: post #56

Rodney Brooks' deep learning predictions have not aged well. For example, in his 2018 blog post ( http://rodneybrooks.com/forai-steps-toward-super-intelligenc... ), he rates various approaches from 1-3 (with 3 being the best). Neural nets scored: Composition: 1 Grounding: 3 Spatial: 1 Sentience: 1 Ambiguity: 2 At that time, the potential of neural nets was already very clear. He also predicted that by 2020 we'll have…

I'll let you in on a secret: There aren't actually any "AI experts". There are machine learning experts, that is, people whose expertise lies in designing and analyzing systems that perform (semi-)automatic inference on data. But nobody can be an expert on "artificial intelligence", because we don't know what that word really means. We don't even know what intelligence really is. We have no idea how the human mind works. We don't understand emergence, at all, which is why we continue to be surprised when it happens.

So it shouldn't come as a shock when eminent figures commonly labeled "AI experts" make predictions that turn out to be fundamentally and embarrassingly wrong in a very short timeframe: They're just talking out of their behinds, like everyone else.

Re: Rodney Brooks on GPT-4

#83

Post by krainboltgreene »

ramraj07 wrote: post #47
fwlr wrote: post #33

> The large language models are a little surprising. I’ll give you that. I think this is the key point about LLMs that kind of explains the wide and polarized views on whether it understands or parrots, whether it can think or is the precursor to thinking or is a dead-end, whether it will catastrophically destroy the world, or “merely” make it steadily worse with bullshit, or just put a few industries out of a job. A…

The issue with all these experts is they still think it's human nature to be able to fully understand the world before they speak about it. On the contrary it's human nature (and all animal nature) to figure out how to navigate the world without fully understanding or having a complete model of it. All you need is a working model that affects the facets of the world you need to deal with. I still remember in the 90s…

> The issue with all these experts is they still think it's human nature to be able to fully understand the world before they speak about it

None of the experts think this.

Re: Rodney Brooks on GPT-4

#84

Post by steveBK123 »

thatguyknows wrote: post #72
p-e-w wrote: post #68

Earlier quoted context omitted.

> Making that ocean deeper is not a trivial problem that we can just throw more compute or data at. You can't possibly know that, given that we don't actually understand how LLMs work on a high level. > We've pretty much tapped out that depth with GPT4 GPT-4 is three months old and you're confident that its working principle cannot be extended further? Where do you get that confidence from?

Sam Altman said it himself. He seems like a reasonable source. If you're familiar with other fields of AI, adding more and more layers to ResNet was the hotness for awhile, but the trick stopped working after awhile.

Exactly, and OpenAI has been around nearly 8 years, consumed huge amount of data with tons of compute. They are just showing us the product now.

It is possible they've reached some 80/20 point and he is pretty honest about how much more extendable the current approach really is.

Would explain going to congress and asking for regulation (of their not-quite-there-yet competitors who they want a regulatory moat against).

Re: Rodney Brooks on GPT-4

#85

Post by chris-orgmenta »

softwaredoug wrote: post #3

Annoyed at all these N=1 articles from prominent thinkers about this stuff. Especially from scientists - can these sorts of folks please more carefully quantify, how often it’s “wrong” and then from that decide whether or not to “calm down”. Right now I suspect we hear from the outliers on both ends of the spectrum here. People who either see AGI happening tomorrow and the more dismissive crowd. But aside from what w…

I feel that 'we in the middle' are ignored. Maybe 'the middle' is too polysemous here. The person to the left and the person to the right are shouting at the person in the middle. They often have uncharitable arguments, and they often take arguments from the other side uncharitably. The person in the middle is ignored.

Tug-of-rope game theory means that no-one is going to start pulling from the middle. People join the (extreme) end that they want to slightly drag the conversation towards. Maybe that's part of it.

I wonder how many $XX,XXX I have historically spent on labour, for things that I now get in seconds. Data entry / manipulation? Sure. But also wisdom / knowledge. Entire industries HAVE collapsed over night. And will continue to collapse. And ignoring that - Why are so many assuming that AI is to be a human-replacement, rather than a limb/exoskeleton?

And why is the conversation 'This is magic' vs 'You are stupid, this is just pulling the wool over your eyes', rather than 'This is a valuable tool in our toolbelt, that will clearly create trillions in value - Just like search, just like the resistor, just like the lightbulb'.

I don't understand.... I'm not claiming LLMs to be magic. I'm not saying they are indications of AGI. I'm not saying that the world as we know it is over. But this is important. It's clearly important. It's clearly valuable. It has shown itself to be.

Yes, ChatGPT lies. OK? We know. We cater for that. We don't expect it not too.

It feels like talking to classic rock enthusiasts that dismiss electronica et al entirely. Fine, but god_damn_ you are missing out on some incredible sound design - My heart breaks daily from what I pump into my ears - Some from rock, some hip-hop, some electronica, and most excitingly, the >2000 merging, where the synergy of genres learning/borrowing from each other is.... just. great. Sit back and enjoy. It is not marvellous? ChatGPT is obnoxious, annoying, repetitive, avoidant, argumentative . But I'm still able to appreciate its value.

We could stop, right now. Freeze ChatGPT in its current state. It will still create trillions in value. I don't care if future improvements are incremental, at this point.

JUST the 'tip of my tongue' / synonymous words value from ChatGPT is useful. Not having to know exactly what term to plug into google.... This is the glue that binds the gray, while before we were stuck with the black and white.

At the very least, this is another 'Google it' revolution - In the 1990s, I remember idiots (inc. me) arguing in the pub for 8 hours, over facts that should have been verified within 10 seconds.

I, for one, am enjoying my new bionic arm.

Re: Rodney Brooks on GPT-4

#86

Post by klyrs »

wewtyflakes wrote: post #23
grugagag wrote: post #9

Earlier quoted context omitted.

> Remember when the Internet was new and no-one believed anything on it? Is this true?

This is true; it was a almost a meme before memes existed, as to say "I saw it on the internet so it must be true!", as a way to teasingly highlight how non-credible it was seen at the time.

"Kilroy was here" was a meme before "meme" was coined. I don't think the phrase "before memes existed" has meaning.

Re: Rodney Brooks on GPT-4

#87

Post by seanhunter »

jmyeet wrote: post #70

GPT-4 is pretty amazing but I, too, feel this is being overhyped. For me, a sobering example is how OpenAI does math (eg [1]). Specifically, the model clearly doesn't really understand multiplication and "learns" it from training data. This tends to get the first few and last few digits right for a simple multiplication with 6-7 digit numbers. Now you can solve that with plugins (eg training the model to recognize ma…

Most of the time when people find a maths problem that they can trick the model into getting wrong, it's also possible to get the model to give the correct answer with better prompting.

A trick that's worth knowing is just to ask the model to give each step in the solution and explain as it goes. This gives the model "time to think" and leads to better results.

Re: Rodney Brooks on GPT-4

#88

Post by krainboltgreene »

klipt wrote: post #79
fwlr wrote: post #33

> The large language models are a little surprising. I’ll give you that. I think this is the key point about LLMs that kind of explains the wide and polarized views on whether it understands or parrots, whether it can think or is the precursor to thinking or is a dead-end, whether it will catastrophically destroy the world, or “merely” make it steadily worse with bullshit, or just put a few industries out of a job. A…

> language encoded way more information about reality than we thought it did Language is roughly what separates humans from other apes ... so why would it surprise us that it encodes much of the information of civilization?

> so why would it surprise us that it encodes much of the information of civilization?

The Humanities are largely considered superfluous in the tech world, why are you surprised that they are surprised?

Re: Rodney Brooks on GPT-4

#89

Post by jstx1 »

Gigachad wrote: post #81
jstx1 wrote: post #63

Earlier quoted context omitted.

The difference is that GPT-4 is already useful even in its current imperfect form. It’s not a 90% thing that needs to get to 100% for us to use it.

VR is useful in its current form. Self driving tech is being used for lots of assistance feature. They just haven’t changed the world yet.

In their current state VR and self driving aren’t on the same level of usefulness as LLMs, not even close.

Re: Rodney Brooks on GPT-4

#90

Post by krainboltgreene »

consilient wrote: post #43
fwlr wrote: post #33

> The large language models are a little surprising. I’ll give you that. I think this is the key point about LLMs that kind of explains the wide and polarized views on whether it understands or parrots, whether it can think or is the precursor to thinking or is a dead-end, whether it will catastrophically destroy the world, or “merely” make it steadily worse with bullshit, or just put a few industries out of a job. A…

> My current (tentative) resolution of the surprise is that language encoded way more information about reality than we thought it did. (Enough information that you can fully derive reality from language seems improbable, but iirc it did derive Othello and partly derived chess and I would have thought there wasn’t enough information in language to derive those without playing the games as well, so I can’t rule it out…

> it's "read" a significant fraction of everything we've ever written about chess

I doubt heavily that a significant fraction of chess's writings are even available in digital format, much less inside of CommonCrawl and correctly trained on.

Post reply on HN