SI Agent News
DevelopingClaudeResearch

Stanford researchers post a human-written proof draft hours after Claude's Riemann claim

A rival, entirely human proof draft appeared within hours of an Anthropic mathematician's claim. Neither is verified yet, so treat the question as open.

The short version

Why it matters

It is tempting to read this as a contest between AI and humans, but nothing has been settled. The more useful habit is to ask what kind of evidence each side offers and who has checked it. Until independent experts weigh in, repeating either claim as fact is premature.

A rival claim, within hours

The Riemann hypothesis is a famous open question about the pattern of prime numbers. Anthropic mathematician Levent Alpoge announced that he and Claude, Anthropic's AI model, had found a proof.

As 36Kr reported, citing The Decoder, Stanford University researchers then released a competing proof draft, completed entirely by humans. The reports say both proofs are still waiting for the wider mathematical community to check them.

Two different kinds of evidence

The two claims are not the same sort of thing. ITMedia of Japan reported that the Claude-linked proof is a 17-million-line formal proof in Lean, software that checks every step of an argument mechanically. It also reported that the Lean community said the proof was correct after multiple verifications. That statement rests on a single outlet so far.

The Stanford item is described as a draft. A draft is a written argument that experts must read and test, which is a different level of evidence from a machine-checked proof.

Which problem is being solved?

The reports do not fully line up. ITMedia refers to the Riemann hypothesis, while 36Kr refers to the generalized Riemann hypothesis. The two are related but not identical, and the summaries do not say which one each proof addresses.

Until that is clarified, it is hard to say whether the two claims are even competing for the same result.

What to watch

Treat this as one unresolved story rather than a race with a winner. The first thing to look for is a clear statement of what each proof covers. After that, look for independent experts who have actually checked the work.

What to watch

Try it

When you see a big AI or human breakthrough claim, ask two questions: what checked it, and was that check independent of whoever made the claim?

Sources and links

Liked it? Say why

Add a line for other readers.

No account needed. We never show your email. Comments with links, ads or abuse are held and not shown. Your comment shows after a quick check.

Listen to the AI Wire headlines

Listen to the AI Wire

Listen to our daily AI shows every morning: on your commute, in the shower, wherever you are

Listen to the daily shows

First reported by 36Kr · Read the original. Written by the SI Agent News Desk from the reporting above.Written with AI from the source above. Every number is checked against the source. Spot a mistake? Corrections.

More AI news