It feels like publicity stunt. Of they really care about research, they would at least have given more reasonable citations in the Navier-Stokes paper.
> Otherwise you could just put this problem into an automated theorem prover (we've had those forever, they are actually just brute-forcing it).
There are many automatic theorem provers that do very clever stuff, just as the underlying theroy describes.
> I am shocked how people can deny that solving Navier Stokes requires some sort of intelligence.
It is absurd to waste time discussing whether it is inteligent or not. It is just an algorithm, we know how it works, and it does exactly what we expect it to do. LLMs are not magical things. The main difference is the scale: for Navier-Stokes they spent in 3 days more money that the whole mathematical community over the last 20 years easily.
By the way, I'm not saying that LLM's are useless, that I'm anti-AI or anything like that.
Intelligence does not seem to be magic either, as LLMs are proving now. It is indeed a waste of time to argue that LLMs are not intelligent in their own way, they obviously are. If Navier Stokes doesn't convince you, nothing will.
I just used a £89 Codex subscription to do very intelligent things with it, stuff that I would have had to sit down and ponder and work on for quite a while, and I have a PhD in that. I didn't need to do anything special except explaining the problem(s) to the AI, and my theory of it so far. It took it from there. If that is not intelligence, nothing is.
We know what calculations it does. We have some hazy idea of some bits of how those calculations lead to something that at least somewhat resembles intelligent behaviour. But that's a far cry from actually knowing how it works.
For instance, suppose you give one of today's frontier models some of those chain-of-cubes rotation puzzles (the sort that infamously men are about 1sd better at than women, statistically speaking). How well will it do? I have absolutely no idea and I'm quite sure that a more detailed understanding of the transformer architecture would not make my guesses any better. (Actually, I do kinda have some guesses but they're based on a vague notion about how the models might be partitioned between vision-y bits and language-y bits, and it's very possible that that notion is out of date.)
> it does exactly what we expect it to do
Were you, let's say 6 months ago, expecting it to resolve one of the Millennium Prize problems?
(I do agree that it is more productive to ask "what can and can't they do?" than "should we classify that as intelligent or not?".)
> for Navier-Stokes they spent in 3 days more money than the whole mathematical community over the last 20 years easily.
Are you sure?
(The numbers I've heard, which I admittedly have no very strong reason to trust, don't seem that way to me.)
> Were you, let's say 6 months ago, expecting it to resolve one of the Millennium Prize problems?
I didn't expect them to throw millions of dollars at each famous math problem. But one year ago we already had LLMs that solved IMO problems, no?
> Are you sure? (The numbers I've heard, which I admittedly have no very strong reason to trust, don't seem that way to me.)
Math has very little founding compared to other science domains. Also, if you filter mathematicians by specialization in PDE and that have worked on Navier-Stokes, then you end up with a very niche community.
> For instance, suppose you give one of today's frontier models some of those chain-of-cubes rotation puzzles. How well will it do?
I feel like this is not the correct way of thinking about it. We can also ask, for instance, how well a state-of-the-art algorithm for the salesman problem works on a particular graph topology. People do PhD thesis on topics like that, so the answer is not obvious at all. For LLMs we still don't have a curated theory that explains what they're good/bad at, and that you don't see how to extract an answer from the definitions is no surprise since this is obviously not an easy problem. But all this is normal because this is a rather new topic (models of this scale appeared when? 3 years ago? That's nothing for science).
Anthropomorphizing LLMs has added so much noise to this discussion.
Yes, one year ago we had LLM-based AI systems solving some IMO problems. My impression is that most observers at that time didn't expect them to be solving Millennium Prize problems within a year.
> Math has very little funding compared to other science domains.
True. But to whatever extent the numbers I've seen are correct, for the whole mathematical community to have spent less on Navier-Stokes than OpenAI did -- even if we value the tokens they spent at something like market rate rather than at what the compute actually costs them (which might be right since any capacity they use internally can't be sold to customers) -- the average number of mathematicians working on Navier-Stokes since 2000 would need to be somewhere around four (depending of course on how well paid they are), and that seems too low to me.
> I feel like this is not the correct way of thinking about it.
It seems to me that if you say "It is absurd to waste time discussing whether it is intelligent or not. It is just an algorithm, we know how it works, and it does exactly what we expect it to do." then this only makes any sense if your "knowing how it works" and "what we expect it to do" enable you to predict what it can and can't do.
(I repeat that I agree that what matters is what it can do, not whether we choose to apply the term "intelligent" to it. But unless I misunderstood you were saying somewhat more than that.)
> Anthropomorphizing LLMs has added so much noise to this discussion.
I think sometimes it helps, sometimes it hurts, and sometimes it's indifferent, because LLMs are like us in some ways and unlike us in some ways. (The same goes for many other things, but LLMs are much more like us in some important ways than any other human-made artefacts.)
For me, mathematicians seem to be still having an inner discussion rather a making these essays for the more general public. In any case, I think that statements of the type "there's actually no role for human Mathematicians" are completely non-serious, so it'd sad that the discussion concentrates on that.
I agree, but it's still something we shouldn't eliminate a priori. I think Mathematics as a profession will be greatly enhanced by AI, but I can't prove that.
> In the (extremely) short run, yes. In the long run, those jobs will also be done by AI.
Let's be honest: we don't know. Maybe you're right, but for the moment it's more likely that you're not. And countries cannot bet on that vague intuition at the cost of destroying their research communities and world leadership (which takes decades if not a century to achieve).
> 1) My way of thinking, and 2) what I already know and how well I recall it in this context.
I don't think we're discussing pedagogy. Good _research_ exposition is instead related to communicate your intuition and way of seeing things. The conceptualization of a given situation or problem is what is valuable, how you connect it with other stuff, etc. It is then up to you to memorize and interiorize it.
Seems to me to be two different sides of the same coin. Pedagogy is about finding a way to communicate to me an idea based on my intuition and seeing things. A research exposition is about presenting the idea in terms of your intuition and seeing things.
Good research exposition is the same as good "pedagogy" just for a different audience. Both have to consider didactics. When writing a paper, you teach something to the fellow researchers who know less about a thing than you.
A professor of mine had an anecdote of meeting Serre and complaining to him about Bourbaki style and how hard it is for students.
Serre's reply was "But we never wrote those books for students! We wrote them for researchers to have a handy reference for all proofs of basic results."
reply