upvote
The AI labs have out considerable effort in trying to find and patch lean exploits. They explicitly set agents and have them try to prove false.

> Daniel used OpenAI internal models to discover new soundness issues in the official Lean kernel and runtime

https://leodemoura.github.io/blog/2026-8-24-postmortem-for-t...

They found several bugs and they have patched them. Lots of work going into making sure lean is sound.

reply
This is a crucial point. There have been many bugs in Lean (and in other proof assistants for that matter). Proof assistants work well on human input, because it was created with a certain intent.

We simply don’t know what those 13M contain and whether it “makes sense” and doesn’t trigger Lean bugs. (There are “independent” lean verifiers, but historically they contained the same, or similar, bugs.)

reply
It is possible, although the post notes that the proof was also verified by the Comparator, which means any exploited bug has to also be present in that checker. Which is not unheard of, but is much less likely than merely an exploit in Lean 4.
reply
The comparator was only used to verify that the final statement indeed is a valid formalization of Fermat's Last Theorem, not that the proof leading up to it is correct.
reply
That must have slipped through Kevin Buzzard's review, which is not entirely unplausible with 29500 theorems to verify...

I think they should spend another few billion tokens and let agents try to disprove any of those statements or links between them. Then I'd be a lot more convinced.

reply
Nope! :(

Meaning, people and LLMs are finding 1=0 bugs in formal verification tools. I have no idea how likely this is in this case, though!

reply
Anthropic surely is well aware. Most likely they asked separate agents multiple times to code review the proof and look for exploits.
reply
Not just lean, but math foundation itself, I am not strong expert, but my understanding is that there is no fully recognized axiomatic foundation for modern math, all proposals could lead to some weird results.
reply
There is, or rather are, fully recognized axiomatic foundations. You are free to choose one you like. Of the most popular ones is ZFC or ZF, but there are others (some lead to the same results some not). The main criteria for popularity is how useful it is. You can even make your own axiomatic where 2+2=5, but it would be useless.

You probably heard about Goedel Incompleteness -- the proof that the the axiomatic itself cannot be proven, like using ZFC to prove ZFC, but that's another topic.

It would be fun to play with this Anthropic/Lean formalization under different axiomatics.

reply
Interestingly, in his ICM 2026 lecture, Terence Tao specifically mentioned that Lean is not based on ZFC.
reply
> Goedel Incompleteness -- the proof that the the axiomatic itself cannot be proven, like using ZFC to prove ZFC, but that's another topic.

Godel theorems are for systems with basic arithmetic, zfc doesn't include arithmetic, thus are not object of Godel theorems.

reply
If you start with "I'm not a strong expert" maybe you should stop continuing saying wrong stuff. What you just wrote is completely wrong.
reply
support your point with explanation or be ignored :-)
reply
Godel proved that any system expressive enough to produce an arithmetic is incomplete. He initially proved it for the peano axioms but then it got generalized. ZFC can produce an arithmetic. Also, before being arrogant and demanding explanations, you should give them first for your claims
reply
> expressive enough to produce

you understand that "expressive enough to produce" are not obvious elements of zfc, that's some average consumer napkin math and not strict formalization.

reply
why should they be obvious? they are derived and have been thoroughly proven.
reply
looks like we are in disagreement
reply
A quick google search shows different proof assistants have been used to obtain the Peano axioms from ZFC, such as Isabelle/ZF and Metamath. I think you're just wrong
reply
What are you nerds fighting about please explain
reply
deleted
reply
you are entitled to have your opinion :-)
reply
and you are entitled to talk about maths while rejecting maths
reply
coming back to your argument about peano being obtained from zfc, you obviously can't prove that it happened using purely zfc, and not some logical framework embedded into those proof assistants.

I said I am not expert, I am indeed not expert in zfc and godel theorems, but I am an expert (phd) in actual formalization theory. Formal theory is very simple concept: its alphabet, set of formulas on top of this alphabet, and function which translates one formula to another.

ZFC can't "obtain" peano, simply because it doesn't have say * operator defined. You need to do something on top of it. Additionally, zfc itself looks like loosely formalized say in wikipedia (and I am not sure if there is any strict formalization anywhere), we take it as common sense that it can utilize some simple logical rules (e.g. modus ponens), but what are exactly rules, which could be separate topic of research, this detail is skipped.

reply
That increases the likelihood that they are right.

> support your point with explanation or be ignored :-)

Anyone who says "Godel theorems are for systems with basic arithmetic, zfc doesn't include arithmetic, thus are not object of Godel theorems" and isn't joking warrants a permanent ignore.

https://math.stackexchange.com/questions/1366560/why-does-g%...

https://math.stackexchange.com/questions/1090437/how-to-prov...

reply
imo, those two links are example of rather low quality weird math discussions, but you can keep your opinion
reply
> Moreover, Robinson arithmetic can be interpreted in general set theory, a small fragment of ZFC.

https://en.wikipedia.org/wiki/Zermelo%E2%80%93Fraenkel_set_t...

reply
> interpreted

its hard to me to tell what this means formally(as I said I am not expert). There is no "interpret" operator in zfc. I believe what it says if you add some robinson axioms + some logical rules on top of zfc, you can carry your results.

reply
ZFC has greater consistency strength than PA.

If we take ZFC (or some other set theory) as our meta theory, we can easily see that the axiom of infinity (of ZFC) gives a set of natural numbers (using the von Neumann encoding), which, when equipped with the successor function, is a model of the natural numbers.

reply
zfc doesn't have functions, so you are building something new on top of it.

Also, I am not sure successor function is enough for PA.

reply
That is wildly wrong.
reply
Lean is based on Type Theory not ZFC.
reply
ZFC is probably the biggest foundation, and only Choice is apparently controversial. The results aren't that weird, they're just different and occasionally more useful than using !Choice.
reply
Reply to sibling - lean4 doesn't rest on ZF or ZFC. https://lean-lang.org/theorem_proving_in_lean4/Axioms-and-Co... However I believe an equivalence of power has been shown between the two.
reply
Roughly, yes. See B. Werner (1997) “Sets in types, types in sets”.
reply
do we know if claude's formalization is built on top of zfc and not zfc+extra?

zfc itself is not sufficient, you need some layers of extra concepts formalization to fit specific problem domain(e.g. zfc doesn't define even basic arithmetics), which also could have potential issues.

reply
Within a given inference system, one can define concepts. This doesn’t add any axioms. It is, in essence, just a way to abbreviate things.
reply
ok, you now added some unknown inference system in addition to zfc
reply
The proof system is relatively easy to verify.

I am not entirely sure about lean, but the core algebras for systems like lean are in the 100s of lines of code.

You can likely convince yourself it is correct in a weekend or less - especially with an Ai to help you understand it.

reply
Most systems i have seen are way beyond a 100 lines. And their GitHub repository contain many issues, often soundness bugs. (Granted, many get fixed very fast.)
reply
You need to understand the concept of the core algebra and 100s (with the s), then I think you'd be better positioned to understand my comment.

And granted, I don't know the exact details about Lean. It might be that they don't have an incredibly simple core - as has elsewise been the norm.

reply
the Nanoda type-checker for Lean is ~5,000 lines of Rust:

https://leodemoura.github.io/blog/2026-3-16-who-watches-the-...

...and for those who are looking to roll-their-own:

https://ammkrn.github.io/type_checking_in_lean4/title_page.h...

...and some thoughts on putting stuff in the kernel:

https://lawrencecpaulson.github.io/2026/07/30/Collatz.html

reply