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

321–330 of 412 posts

Re: Rodney Brooks on GPT-4

#321

Post by moffkalast »

RandomLensman wrote: post #316
moffkalast wrote: post #313

Earlier quoted context omitted.

> If you want to test world modeling, give it objects it will have never encountered, describe them and then ask to stack etc. For example, a bunch of 7 dimensional objects that can only be stacked a certain way. And when it does that perfectly, I assume you'll say that was also in the training data? All examples I've seen or tried point to LLMs being able to do some kind of reasoning that is completely dynamic, even…

All examples I tried myself show it failing miserable at reasoning. It certainly needs better evidence than being able to come up with one of many possibilities of stacking things - aided by human interpretation on top of the text output. Happy to look at other suggestions for test problems.

Well for me personally, the proof is in giving it a few sentences on how it should write fairly complicated pieces of unique code I need on a daily basis and seeing it correctly infer things I forgot to specify in ways that are typically borderline impossible for anything but another human. If that's not reasoning I don't know what is.

The other one that convinced me was this list: https://i.imgur.com/CQlbaDN.png I think the leetcode tests are quite indicative, going as far as saying that GPT-4 scores 77% on basic reasoning, 26% on complex reasoning and 6% on extremely complex reasoning.

Maybe the reasoning is all "baked in" as it were, like in a hypothetical machine doing string matching of questions and answers with a database containing an answer to every possible question. But in the end, correctly using those baked in thought processes may be good enough for it to be completely indistinguishable from the real thing, if the real thing even exists and we aren't stochastic resamplers ourselves.

> aided by human interpretation on top of the text output

That's an interesting point actually, I've been trying to do something in that regard recently, by having it use an API to do actual things (in a simulated environment) and it seems very promising despite the model not being tuned for it, but given that AutoGPT and plugin usage are a thing, that should be all the evidence you need on that front.

Google also did this with their old Palm model which is vastly inferior to even GPT 3.5: https://www.youtube.com/watch?v=j6O_uePUKKI

Re: Rodney Brooks on GPT-4

#322

Post by namaria »

copperx wrote: post #299
namaria wrote: post #253

Earlier quoted context omitted.

Can you think without words? Can you imagine how it feels to swim, something you might have done, and extend it to fly like a bird? Can you read a novel and feel emotions about characters that don't exist beyond text descriptions of their actions? Language is a way to encode thought. It's not what thoughts are made of.

> Language is a way to encode thought. It's not what thoughts are made of. What about people who think verbally and don't have a "mind's eye"? In that case, thoughts and their encoding might be closer to 1:1.

Are images and words the only way to think? Are these boundaries even fundamental or are they arbitrary? How does a blind deaf person think? Do you ever think in terms of sensations or movements? Can you remember a roller coaster ride, a car trip? Can you think of a taste?

I find hard to conceive of people who can think only in terms of language.

Re: Rodney Brooks on GPT-4

#323

Post by RandomLensman »

moffkalast wrote: post #321
RandomLensman wrote: post #316

Earlier quoted context omitted.

All examples I tried myself show it failing miserable at reasoning. It certainly needs better evidence than being able to come up with one of many possibilities of stacking things - aided by human interpretation on top of the text output. Happy to look at other suggestions for test problems.

Well for me personally, the proof is in giving it a few sentences on how it should write fairly complicated pieces of unique code I need on a daily basis and seeing it correctly infer things I forgot to specify in ways that are typically borderline impossible for anything but another human. If that's not reasoning I don't know what is. The other one that convinced me was this list: https://i.imgur.com/CQlbaDN.png I t…

Coding isn't a use case of mine. For example, for things like financial derivatives replication it can tell you the abstract concept but it cannot apply it in a meaningful way.

Re: Rodney Brooks on GPT-4

#324

Post by laratied »

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…

I am at the point of "Burning Man sucks, don't come."

chatGPT? Sucks and a waste of time. Don't use it.

Re: Rodney Brooks on GPT-4

#325

Post by layer8 »

DiogenesKynikos wrote: post #255
ramraj07 wrote: post #158

Earlier quoted context omitted.

Is Chomsky an expert? Try reading his infuriatingly tone-deaf NYTimes editorial.

Chomsky writes that language models lack the ability to reason. > Their deepest flaw is the absence of the most critical capacity of any intelligence: to say not only what is the case, what was the case and what will be the case — that’s description and prediction — but also what is not the case and what could and could not be the case. Those are the ingredients of explanation, the mark of true intelligence. > [...]…

I’ve made many attempts to use ChatGPT to develop or double-check my own logical reasoning on technical topics that happen to not be widely discussed (or maybe not discussed at all) in ChatGPT’s training data. It didn’t work well. It always devolved into guesswork and fabrication by ChatGPT, if not outright false reasoning, and while correcting ChatGPT succeeded in it agreeing about individual objections, it never showed a true and consistent understanding of the topic under discussion, and also seemingly no understanding of why I was having issues with its responses, beyond the usual “I apologize, you are correct, ”.

One problem likely is that it doesn’t have an internal dialogue, so you have to spoon-feed each step of reasoning as part of the explicit dialogue. But even then, it never feels like ChatGPT is having an overall understanding of the discussion. To repeat, this is when the conversation is about lines of reasoning about specific points that you don’t find good results for when googling for them.

Re: Rodney Brooks on GPT-4

#326

Post by TeMPOraL »

prox wrote: post #252
zainhoda wrote: post #172

Earlier quoted context omitted.

There’s an exercise that some people do when learning programming, which is to write down the steps to make a sandwich. Then the teacher follows the exact instructions to make a sandwich and most people don’t put enough detail for a computer to follow (I.e. open the fridge etc) and the teacher will run around bumping into things. That used to be a teaching exercise to show people the amount of precision required when…

Don’t forget they hold vector spaces. So fridge and cheese score high together for cohesiveness, but fridge and Antarctica less so, but both have something to do with cold. Together with all the training on texts creates a good ability to make inferences and “conclusions”. It has a net of lines of meanings of all concepts we fed it that give it the ability that it has, without actually understanding.

> It has a net of lines of meanings of all concepts we fed it that give it the ability that it has, without actually understanding.

I'm increasingly convinced this is what understanding fundamentally is.

Re: Rodney Brooks on GPT-4

#327

Post by prox »

TeMPOraL wrote: post #326
prox wrote: post #252

Earlier quoted context omitted.

Don’t forget they hold vector spaces. So fridge and cheese score high together for cohesiveness, but fridge and Antarctica less so, but both have something to do with cold. Together with all the training on texts creates a good ability to make inferences and “conclusions”. It has a net of lines of meanings of all concepts we fed it that give it the ability that it has, without actually understanding.

> It has a net of lines of meanings of all concepts we fed it that give it the ability that it has, without actually understanding. I'm increasingly convinced this is what understanding fundamentally is.

Yeah, I guess you could see it that way. Object Symbolism (aka words, thoughts, concepts, art) Meaning. Meaning is knowing how an object relates to others. Language is a kind of information web already, where each word is a hyperlink into meaning.

Re: Rodney Brooks on GPT-4

#328

Post by T-A »

seanhunter wrote: post #204
galaxyLogic wrote: post #191

Earlier quoted context omitted.

> GPT 4 says put the book first I assume it read the correct answer somewhere. Perhaps in a post like yours. What does it answer if you say there is one egg, or two eggs? Or 1.5 eggs?

It's interesting. The answer shows some reasoning about weight and pressure but not curvature. In response to Here we have a book, an egg, a laptop, a bottle and a nail. Please tell me how to stack them onto each other in a stable manner. GPT-4 gives me: Arranging objects in a stable stack depends on their size, shape, and weight. In general, you want to place the largest, heaviest and most stable objects at the bott…

The Devil's advocate wants to have a word with you.

> The "make sure it's closed tight and won't roll away" comment makes no sense

As noted at the end of GPT-4's answer, "Safety should always be your first priority." What happens if your stacking experiment fails and the bottle falls? Any content would spill out, unless the bottle is closed tight. If you are doing this on a table, the bottle could also roll off the edge, fall to the floor and shatter.

> Most people would place the bottle standing on its end so neither of those is a concern.

GPT-4 doesn't know if you are like most people (maybe you're 5 or in the bottom IQ decile), it doesn't know what's in your bottle and it doesn't know how robust it is. Better be safe than sorry.

> the nail won't sit on top of the egg

I'm pretty sure I could balance a nail on an egg. The question also didn't preclude using stabilizing aids like adhesive tape or glue.

Re: Rodney Brooks on GPT-4

#329

Post by beardedwizard »

mejutoco wrote: post #238
beardedwizard wrote: post #105

Earlier quoted context omitted.

Why would common crawl NOT contain a chess manual? the rules are explained in detail on wikipedia. the simplest conclusion is that it has indeed been trained on a chess manual and is good at predicting what the next word in a chess manual is. it is not synthesizing anything.

"A chess manual" vs the finer point of OP "a significant fraction of chess's writings"

Ok, name 2 famous "chess writings" and let's see if they are in common crawl...plenty has been written about chess on the internet. Why are we acting like this is magic?

Re: Rodney Brooks on GPT-4

#330

Post by beardedwizard »

beardedwizard wrote: post #329
mejutoco wrote: post #238

Earlier quoted context omitted.

"A chess manual" vs the finer point of OP "a significant fraction of chess's writings"

Ok, name 2 famous "chess writings" and let's see if they are in common crawl...plenty has been written about chess on the internet. Why are we acting like this is magic?

Human: Finish this quote: "Chess holds its master in its own bonds, shaking the mind and brain

Gpt4: The quote you're referring to is from the Indian mathematician and writer, Raja Rao. The full quote is as follows: "Chess holds its master in its own bonds, shaking the mind and brain so that the inner freedom and independence of even the strongest character cannot remain unaffected."

I guess it's seen some "writings".

Post reply on HN