Two Proofs of the Riemann Hypothesis: One With Claude, One Without AI
The code of a proof by an Anthropic mathematician was checked independently. Separately, Alessandro Masullo published a work written without AI.
Levent Alpöge is a mathematician at Anthropic, the company that makes the Claude AI model. He says he has solved the Riemann hypothesis with Claude's help. The problem dates to 1859 and concerns a hidden structure behind prime numbers. His proof is written in Lean, a programming language in which a computer checks every step. Several mathematicians, including Andrew Sutherland of MIT, confirmed that the code compiles. Separately, Stanford professor Ravi Vakil posted a proof of the General Riemann hypothesis by Stanford postdoc Alessandro Masullo. Vakil noted it was done with no AI.
This leaves mathematicians in an unusual spot. A machine has accepted Alpöge's proof, but no human has explained it yet. Masullo's work has the opposite problem. He says it is the result of 12 years of work. It is not written in Lean, so there is no way yet to check it.
Two things are worth watching. Alpöge has promised to explain his result in plain human terms, and it remains to be seen whether he does. The second is whether anyone translates Masullo's proof into Lean. Neither has happened yet.
Source: techcrunch.com
Free course
Stop reading about AI — start building with it
The free Claude Code course: your first site, tool or game — no coding. No upsells, no cross-sells — nothing to buy here.
Start free →
Author
Evgenii Arsentev
PhD · Founder & CEO, Dad4Kids
Articles · Latest articles