Anthropic mathematician claims he used Claude to prove 19th-century Riemann hypothesis
Turkish American mathematician Levent Alpöge, who works at San Francisco-based AI company Anthropic, announced on his personal website on Friday that he used the company's software Claude to help prove the Riemann hypothesis. Alpöge said the solution took about six weeks and shared verification code written in Lean, which other mathematicians have begun reviewing. His note appeared one day after Anthropic went public on the Nasdaq stock market index, timing Alpöge called "horrible," adding he would have finished earlier if possible.
Proposed in 1859 by German mathematician Bernhard Riemann, the nearly 170-year-old problem examines how prime numbers appear and are distributed along the number line. It is one of seven Millennium prize problems selected by the Clay Mathematics Institute, which offers a $1m reward for each solved problem.
The claim also underscores growing discomfort over wealthy tech companies applying AI to major mathematical problems. On Friday, The New York Times reported that independent Italian mathematician Alessandro Masullo released an incomplete proof after almost 12 years of work without AI because he did not fully trust it. In September, OpenAI similarly drew scrutiny after spending $15m on 10,000 AI systems to solve part of another Millennium prize problem, which University of Warwick mathematician James Robinson criticized as costly boasting carrying potential environmental damage.
Continue in the app — vote & join in ➔