Alpöge claims AI-assisted Riemann Hypothesis proof in Lean
TL;DR
- Levent Alpöge, a mathematician at Anthropic, says he and Claude have found a proof of the Riemann Hypothesis, extending to the Generalized version.
- The argument is posted in Lean, a language that mechanically checks each step; MIT's Andrew Sutherland independently compiled it and called it historic.
- The Clay Mathematics Institute attaches a one million-dollar Millennium Prize to the problem, first posed by Bernhard Riemann in 1859.
Levent Alpöge, a mathematician on Anthropic's staff, says he has proven the Riemann Hypothesis working with the company's Claude model, in a claim that has not yet been independently verified. Scientific American reported the announcement on Friday.
"I believe I have, working with Claude, discovered a proof of the Riemann Hypothesis," Alpöge wrote. He posted the argument in Lean, a programming language that mechanically checks each logical step, and said the result extends to the Generalized Riemann Hypothesis, an expansion of Bernhard Riemann's 1859 conjecture about the distribution of prime numbers.
Early reactions from number theorists are cautiously positive. MIT's Andrew Sutherland independently compiled the Lean code. "This is indeed an historic moment," Sutherland said, also calling the work "a human triumph" and crediting the people who built Lean and large language models. Hector Pasten of Pontifical Catholic University of Chile and Carlo Pagano of Concordia University and Google DeepMind told the magazine the new ideas could "open up brand new avenues of thinking in number theory."
The Clay Mathematics Institute keeps a one million-dollar bounty on the problem as one of its Millennium Prize Problems. Scientific American wrote that "Issues can arise in Lean, but it shouldn't take long for experts to verify whether the hypothesis is properly coded." Three researchers we track in our Who's Who directory were circulating the piece within hours of publication.
Shared on Bluesky by 3 AI experts
-
It is time to freak out www.scientificamerican.com/article/anth...
View on Bluesky →
Originally reported by scientificamerican.com
Read the original article →Original headline: Anthropic mathematician claims to prove Riemann hypothesis, most famous problem in all of math