News for hackers. Stuff that matters.
Re:AI in mathematics is forcing big questions by therobots927 on Friday June 26, 2026 @10:26PM (#48694561)
It’s a well known problem in higher mathematics that even if you’ve solved a problem, often the proofs are incredibly long and complex and require an extensive amount of time spent by peers to review it.

It would be great if someone could explain to me how AI improves this situation. Even if AI thinks it’s solved a problem, unless the proof is incredibly efficient and well explained, it will be difficult to verify the correctness. One hallucination in 300 steps of logic is enough to destroy the entire proof.

Re:AI in mathematics is forcing big questions by hilbertseries on Friday June 26, 2026 @11:12PM (#48694776)
In 2012 Mochizuki claimed to have proved the abc conjecture by developing a new branch of mathematics. He was a respected mathematician, but the theories he had developed were so complex no one could determine if he was correct. It took six years until two number theorists dissected the proof and found a fatal flaw in it.
Re:AI in mathematics is forcing big questions by scheme271 on Saturday June 27, 2026 @04:42AM (#48696366)
Mochizuki and a group of mathematicians still claim that Stix and Scholze didn't actually identify a flaw and his proof was published in a journal (where Mochizuki is the chief editor and the reviewers were people from Mochizuki's instituion). I think most of the math community don't believe his proof although Mochizuki and some others claim it's valid.
Re:AI in mathematics is forcing big questions by jonahx on Friday June 26, 2026 @11:01PM (#48694724)
> It would be great if someone could explain to me how AI improves this situation.

It's main utility is in the search step, not the verification step. The search is the bulk of the work and creativity. Separately, as the sibling commenter pointed out, it will likely get better at the verification step as well, with integrations of tools like Lean.

> One hallucination in 300 steps of logic is enough to destroy the entire proof.

The situation with human mathematicians is not much different. Eg, Wiles original proof of Fermat's Last Theorem contained errors found by reviewers, which he later repaired.

Re:AI in mathematics is forcing big questions by tacomonstrous on Saturday June 27, 2026 @12:41AM (#48695201)
>The situation with human mathematicians is not much different. Eg, Wiles original proof of Fermat's Last Theorem contained errors found by reviewers, which he later repaired.

In fact, it was Wiles himself who realized there was an error.

Re:AI in mathematics is forcing big questions by jonahx on Saturday June 27, 2026 @11:08AM (#48698934)
Yes, at the prodding of the reviewer (from wikipedia):

> After the announcement, Nick Katz was appointed as one of the referees to review Wiles's manuscript. In the course of his review, he asked Wiles a series of clarifying questions that led Wiles to recognise that the proof contained a gap.

It seems likely Katz played a crucial role here.

Wiles is just a prominent example. In practice, other humans often play this "verification role" for mathematicians, whether via a formal review process or informal discussion.

Re:AI in mathematics is forcing big questions by skipkey on Friday June 26, 2026 @10:49PM (#48694662)
I would imagine that in the future AI will be doing proofs in Lean or whatever the successor to it, which gives you a pretty good confidence it’s correct.
Re:AI in mathematics is forcing big questions by bckgrndrdtn on Saturday June 27, 2026 @09:04AM (#48697908)
The future is now. See the article under discussion for some examples.
Re:AI in mathematics is forcing big questions by cdetrio on Friday June 26, 2026 @11:30PM (#48694855)
The article addresses that, a formal verification layer provides a computational check that a proof is correct.

There's an interesting wrinkle though. The formal definitions and statements need to be correct, i.e. need to faithfully translate the human-readable (informal) definitions and statements. A few months ago, Kevin Buzzard created an auto-formalization challenge: https://gist.github.com/kbuzzard/778bc714030b3e974ab5f403878...

It's a handful of human-written lines, a skeleton that defines an "API surface". If you use AI to generate the rest of the code, the API surface (along with the Lean compiler) acts to guarantee that the whole generated solution (tens of thousands of lines) will be correct.

The discussion thread about the challenge, which was solved a couple weeks ago, is here: https://leanprover.zulipchat.com/#narrow/channel/583336-Auto...

Without the human-written or human-audited skeleton, you might get AI-generated slop Lean code that compiles, but contains nonsensical definitions/statements and thus vacuous proofs. Though from what I gather, recent models are much improved in their capability to generate correct statements and definitions and proofs, for both formal math and informal math.

A discussion thread about a different auto-formalization effort without human-written API surface, with some correct output and some slop: https://leanprover.zulipchat.com/#narrow/channel/583336-Auto...

Re:AI in mathematics is forcing big questions by jstanley on Saturday June 27, 2026 @01:21AM (#48695391)
This is what Lean is for: https://lean-lang.org/

If you have the LLM generate Lean code, and it compiles, then the proof is correct and you don't need to bother checking its working. (You still need to check that it is proving the theorem you asked it to prove).

Re:AI in mathematics is forcing big questions by AaronAPU on Saturday June 27, 2026 @07:52AM (#48697453)
I’m playing devil’s advocate here, so go easy on me. But how completely do we know lean is perfectly true in all cases?
Re:AI in mathematics is forcing big questions by bckgrndrdtn on Saturday June 27, 2026 @09:03AM (#48697903)
There's a lot to be said about that: https://lean-lang.org/doc/reference/latest/ValidatingProofs/

But let's assume Lean is perfect: - you have to trust your file system - you have to trust your OS - are you sure your monitor displays "proof is valid" correctly - maybe there's a bug in your CPU hardware? - or background radiation flipped a bit in your RAM - do you trust your brain? Maybe you went to bed last night firmly believing "that proof is wrong" and you woke up this morning with a different belief that "the proof is valid".

Re:AI in mathematics is forcing big questions by pfdietz on Saturday June 27, 2026 @11:59AM (#48699353)
If there's concern, I understand Lean can dump out a proof object, and this object can be checked for validity by a separate, smaller program. One could even write that smaller program multiple times.

It would be interesting to see how Lean is being tested. Pummeling it with inputs from AI proof search engines sounds like a good way to expose hidden bugs, just as pummeling a compiler with randomly generated programs is a good way to expose previously unknown compiler bugs.

Re:AI in mathematics is forcing big questions by throw310822 on Saturday June 27, 2026 @09:08AM (#48697935)
> even if you’ve solved a problem, often the proofs are incredibly long and complex and require an extensive amount of time spent by peers to review it.

I hope mathematicians have the self-respect of not establishing rules such as "if a proof is longer than X, break it up in smaller ones, otherwise it'll be outright rejected".