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

A misalignment of AI in mathematics

mathandai.org

Post reply on HN
Print view View original

951–956 of 956 posts

Re: A misalignment of AI in mathematics

#951

Post by eru »

robotpepi wrote: post #947
eru wrote: post #880

Earlier quoted context omitted.

Why do you need AI to ask the questions? You give your AI the problem, like 'make this computer program faster and more reliable', and it'll go off and do the math necessary to make it so.

so you're 100% sure that AI will solve every intellectual problem in the future, since as long as that's not the case, it's us who need to ask the questions. I don't know man, I wouldn't bet on that. What if we end up being wrong and then there is not research community?

Huh? Where do I say that AI needs to solve every intellectual problem?

For all we know, P vs NP will never be resolved neither by human nor AI, and the world will just keep turning.

Re: A misalignment of AI in mathematics

#952

Post by eru »

robotpepi wrote: post #939
eru wrote: post #881

Earlier quoted context omitted.

Applications don't care whether the math was proven and understood by humans or computers. Your algorithm will get faster no matter where the insight came from.

to be honest it is difficult to discuss with someone who doesn't even try to understand the basics of basic science (and how it compares with _applied_ sicence), yet talks with so much confidence. even the solution to navier stokes won't have an immediate practical effect...

Huh? The resolution of Navier Stokes won't have much of an effect, yes.

There's lots of problems like that. Eg if we prove P != NP, that won't have much of an immediate effect either.

However, there's also plenty of problems whose solutions will have practical effects, some even immediate.

Re: A misalignment of AI in mathematics

#953

Post by robotpepi »

eru wrote: post #878
robotpepi wrote: post #847

Earlier quoted context omitted.

The problem with your opinions is that you tend to state certain very quesitonable ideas with 100% confidence. Right now, we don't know if AI will be able to solve bigger problems, or if they'll be so efficients that you no longer need a whole datacenter for running them. We don't even know how much of creative work they are able to do. We cannot make decisions that could destroy decades of progress just because of h…

> Right now, we don't know if AI will be able to solve bigger problems, or if they'll be so efficients that you no longer need a whole datacenter for running them. For the latter: I assume that having a whole data centre will always be an advantage. I am saying that for a fixed target, like proving the Rieman hypothesis from scratch, the required hardware will shrink. And, yes, the Rieman hypothesis hasn't been prove…

> I am saying that for a fixed target, like proving the Rieman hypothesis from scratch, the required hardware will shrink.

Yes, but to what point the hardware will shrink? There are several orders of magnitude of difference in what in your mind AI will become and what more conservative people believe. You take your view as granted...

Re: A misalignment of AI in mathematics

#954

Post by grey-area »

YeGoblynQueenne wrote: post #945
grey-area wrote: post #935

Earlier quoted context omitted.

I’m no big AI cheerleader, but why don’t you try these models to see what they offer - they might suggest some things you haven’t thought of, or save you some time in your research. Are you sure they’d one-shot your algorithm? Maybe they’d come up with worse, maybe better, but you are in a perfect position to actually judge the validity of any claims a LLM makes in this realm, which most people are not. As to why bot…

>> I’m no big AI cheerleader, but why don’t you try these models to see what they offer - they might suggest some things you haven’t thought of, or save you some time in your research. One reason not to try is to create an air gap between the OP's ideas and the data that future models can train on. We saw that the mathematicians who trusted OpenAI and Anthropic with their preliminary work found the rug pulled out of…

Unless you’re very weak-willed, I wouldn’t worry about independence. LLMs make mistakes all the time, they are nothing like an all-knowing oracle nor are they going to replace humans despite the absurd fantasies of LLM fans and those with a vested interest.

Plagiarism is an interesting point, though honestly I think it would be fairly easy to work out who had published similar research if the LLM gives you an idea - personally I see this as the weakest argument against using them, as long as you are strict about attribution - all work like this depends heavily on the research of others - the LLM is just a tool to aid that research IMO.

Opsec is a fair point, and it might be worth avoiding the completely amoral OpenAI at this point for that reason. There are open models though.

Re: A misalignment of AI in mathematics

#955

Post by grey-area »

reasonableklout wrote: post #795
redox99 wrote: post #754

Earlier quoted context omitted.

Of course you're not going to get rich with the kind of software that LLMs can one shot these days. But that kind of software like to-do lists or basic CRUD have been saturated for over a decade, way before LLMs. People overestimate how much you can one shot, yeah a good prompt can get you 90% there but that 10% remaining often takes months of extra work. Software has always progressed this way, lots of devs back the…

But what does the human researcher do in this future? If they are not needed to understand the result, then what's their role? Asking the right questions? But how will they know what questions to ask if they don't have a deep understanding of the domain earned by sweating the details themselves? And how long will they be needed to ask the right questions, how long until AI can do that too? > Software has always progr…

They are needed to guide the research in fruitful directions and understand the result.

LlMs don’t understand, they generate, though they have fooled a lot of people who should know better.

Re: A misalignment of AI in mathematics

#956

Post by pyentropy »

well_ackshually wrote: post #902
charcircuit wrote: post #888

Earlier quoted context omitted.

You don't have to read 32 million lines. You can just look at what got proved. The proof checker lets you trust those millions of lines you did not read.

No, you don't understand: that proof is useless. When you solve mathematical problems, you open up new ones in the process of doing so. You create new research. You create new theories, new notations, new thought. This is just ticking a checkbox. And even worse and more time wasting even: you have ZERO proof that there's no latent Lean bug. Especially in a proof this large.

Not a Lean expert but some of the proof tactics used to prove are probably novel? Or, you could prompt agents later to analyze which lemmas or parts of the proof are surprising or applicable to other problems?
Post reply on HN