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

Life is more than an engineering problem

lareviewofbooks.org

Post reply on HN
Print view View original

221–230 of 257 posts

Re: Life is more than an engineering problem

#221

Post by vouwfietsman »

mrcsd wrote: post #104
vouwfietsman wrote: post #97

Earlier quoted context omitted.

> But it is clear humans are capable of ... Its not though, nobody really knows what most of the words in that sentence mean in the technical or algorithmical sense, and hence you can't really say whether llms do or don't possess these skills.

Words are not reducible to technical statements or algorithms. But, even if they were, then by your suggestion there's not much point in talking about anything at all.

They absolutely are in the context of a technical, scientific or mathematical subject.

Like in the subject of LLMs everyone knows what a "token" or "context" means, even if they might mean different things in a different subject. Yet, nobody knows what "consciousness" means in almost any context, so it is impossible to make falsifiable statements about consciousness and LLMs.

Making falsifiable statements is the only way to have an argument, otherwise its just feelings and hunches with window dressing.

Re: Life is more than an engineering problem

#222

Post by vouwfietsman »

southernplaces7 wrote: post #116
vouwfietsman wrote: post #97

Earlier quoted context omitted.

> But it is clear humans are capable of ... Its not though, nobody really knows what most of the words in that sentence mean in the technical or algorithmical sense, and hence you can't really say whether llms do or don't possess these skills.

>nobody really knows what most of the words in that sentence mean in the technical or algorithmical sense And nobody really knows what consciousness is, but we all experience it in a distinct, internal way that lets us navigate the world and express ourselves to others, yet apparently some comments seem to dismiss this elephant of sensation in the room by pretending it's no different than some cut and dried computati…

> by pretending it's no different than some cut and dried computational system

This is not really what is going on, what is going on is a mix-up in interpreting the meaning of words, because the meaning of words is not transitive between subject matter unless we arrive at a scientific definition which is leading, and we have not (yet).

When approaching the word consciousness from a spiritual POV, it is clear that LLMs may not possess it. When approaching consciousness from a technical point of view, it is clear that LLMs may possess it in the future. This is because the spiritual POV is anthropologically reductive (consciousness is human), and the technical POV is technically reductive (consciousness is when we can't tell it apart).

Neither statements help us clarify opposing positions because neither definitions are falsifiable, and so not scientific.

Re: Life is more than an engineering problem

#224

Post by SkiFire13 »

bell-cot wrote: post #201
SkiFire13 wrote: post #148

Earlier quoted context omitted.

In theory you can prove a theorem just by enumerating all the possible proofs until you find the one for the theorem you want. This is extremely slow, but do you think there's any reasoning in doing this? Of course we don't know whether an LLM is doing something like this or actually reasoning. But this is also the point, we don't know. If you ask a question to a person you can be confident to some degree that they d…

> In theory you can prove a theorem just by enumerating all the possible proofs until ... An interesting hypothesis! I'm neither a mathematical logician, nor decently up to date in that field - is the possibility of this, at least in the abstract, currently accepted as fact? (Yes, there's the perhaps-separate issue of only enumerating correct proofs.)

It depends on what theory you're working in (at which point deciding whether to use one theory or another becomes more like a phisolophical question).

I'm mostly familiar with type theory, of which there are many variants, but the most common ones all share the most important characteristics. In particular they identify theorems with types, and proofs with terms, where correct proofs are well-typed terms. The nice thing is that terms are recursively enumerable, so you can list all proofs. Moreover most type theories have decidable type checking, so you can automatically check whether a terms if well-typed (and hence the corresponding proof is correct).

This is not just theory, there exist already a bunch of tools that are being used in practice for mechanically checking mathematical proofs, like Coq, Lean, Agda and more.

When I said "in theory" however it's because in practice enumerating all proof terms will be very very slow and will take forever to reach proofs for theorems that we might find interesting.

Since we're in the LLM topic, there are efforts to use LLMs to speed up this search, though this is more similar to using them as search heuristics though. It does help though that you can have automatic feedback thanks to the aforementioned proof checking tools, meaning you don't need costly human supervision to train them. The hope would be getting something like what Stockfish/Alphazero is for chess.

Re: Life is more than an engineering problem

#225

Post by inglor_cz »

dexwiz wrote: post #159
inglor_cz wrote: post #105

Earlier quoted context omitted.

OK, so I asked ChatGPT how to plant potatoes on Mars... ================================================= Planting potatoes on Mars would be a pretty fascinating (and challenging) task! While Mars has conditions that make growing traditional crops difficult—like low temperatures, low atmospheric pressure, and a lack of oxygen—scientists have been experimenting with ways to grow plants in Martian-like conditions. Here…

What is this supposed to prove? The question isn’t particular novel. There are decades of speculative terraforming discussions on the internet that you can search via Google, especially after the Martian book and movie.

It was supposed to prove nothing. I am not an expert on Mars and on a forum full of engineers, I expected someone more knowledgeable to start taking the answer apart systematically: this is truth, this is hallucination, this may be truth or not, but it is stated too optimistically.

I can see one thing. Mars is indeed further from the Sun than the Earth, but the light on the surface is probably still comparable to a slightly cloudy day on Earth, and plants grow just fine in slightly cloudy climates.

Re: Life is more than an engineering problem

#226

Post by mantas »

teleforce wrote: post #195
mantas wrote: post #32

Earlier quoted context omitted.

According to my tradition, my language is the best of the best. Get off my lawn with your language.

I don't speak Arabic but I wish I can perhaps later when I can get around and learning it. If you say that your language is the best or perfect then present your case with examples and proof. Don't run from good arguments and justifications.

I'm just making fun of claiming that some language is the best, because native speakers themselves say so. Especially talking about such chauvinist culture like muslim arabs.

Re: Life is more than an engineering problem

#227

Post by throwaway2037 »

I didn't know Ted Chiang before seeing this post. If anyone needs to know about him, he has a separate Wiki page: https://en.wikipedia.org/wiki/Ted_Chiang

    > Ted Chiang is an American science fiction writer. His work has won four Nebula awards, four Hugo awards, the John W. Campbell Award for Best New Writer, and six Locus awards.  Chiang is also a frequent nonfiction contributor to the New Yorker, most recently on topics related to computer technology, such as artificial intelligence.

Re: Life is more than an engineering problem

#228

Post by vicentwu »

"As an analogy, imagine that you could put your dog or cat into hibernate mode whenever you left on a trip. Your dog or cat might not notice, but even if they did, they might not mind. Now imagine that you could put your child into hibernate mode whenever you were too busy to spend time with them. Your child would absolutely notice, and even if you told them it was for their own good, they would make certain inferences about how much you valued them. That’s the situation the human characters in the story find themselves in." Fascinating.

Re: Life is more than an engineering problem

#229

Post by Byamarro »

hnisoss wrote: post #202
abdullahkhalids wrote: post #150

Earlier quoted context omitted.

> No one in neuroscience, psychology or any related field can point to reasoning or 'consciousness' or whatever you wish to call it and say it appeared from X. This is not a good argument. Natural systems, the subject of neuroscience/psychology, are much harder to analyze than artificial systems. For example, it's really difficult to study atmospheric gases and figure out Boyle's/Charles law. But put a gas in a close…

Yea but following your example/analogy you have gas-gas but brain-llm. So how can we then experiment? It's a simulation at best.

Both jets and birds fly but do it in a completely different way. Who said that there's only one way to achieve reasoning?

Re: Life is more than an engineering problem

#230

Post by eterps »

nuancebydefault wrote: post #165
eterps wrote: post #108

Earlier quoted context omitted.

> and wonder how an intelligent person can still think this, can be so absolute about it. What is "actual" reasoning here? Large language models excel at processing and generating text, but they fundamentally operate on existing knowledge. Their creativity appears limited to recombining known information in novel ways, rather than generating truly original insights. True reasoning capability would involve the ability…

> True reasoning capability would involve the ability to analyze complex situations and generate entirely new solutions, independent of existing patterns or combinations. You mean like alphago did in its 36th move?

Isn't that a non-generic 'reasoning-model' instead of something that is reminiscent of the large language model based AIs we use today?

The question is, is it possible to make reasoning models generic and can they be combined with large language models effectively.

Post reply on HN