Riemann Hypothesis Proof: Anthropic Researcher Says Claude Solved Maths’ Greatest Riddle
A mathematician at Anthropic has claimed that he worked with the company’s AI model Claude to prove the Riemann Hypothesis, the most famous unsolved problem in mathematics, which has resisted every attempt at a solution since 1859. Levent Alpoge announced on Friday that he believes he has found the proof, in what would be a historic moment for mathematics if it withstands scrutiny.
“I believe I have, working with Claude, discovered a proof of the Riemann Hypothesis,” Alpoge wrote in a four-page note released on Friday. The claim has not yet been independently verified, and outside experts are still checking the work.
The hypothesis, first proposed by German mathematician Bernhard Riemann in 1859, concerns the distribution of prime numbers, the building blocks of all whole numbers. It is one of seven Millennium Prize Problems selected by the Clay Mathematics Institute, which offers 1 million dollars for a solution to each.
What Alpoge claims
Rather than publishing a conventional mathematical paper, Alpoge released his argument in Lean, a programming language and proof assistant that checks mathematical reasoning step by step. He released more than 17 million lines of computer code with what he claimed was the complete proof, and said the method also resolves a broader version of the problem known as the Generalized Riemann Hypothesis.
Alpoge acknowledged that the full written argument would be extremely long, and that he had not yet mastered all of the mathematics involved. He said he plans to write it up properly. He also said he had consulted Anthropic chief executive Dario Amodei before announcing the result.
The proposed proof builds on an older technique associated with the Russian mathematician Sergey Stepanov, but applying it to integers had proved difficult because one part of the technique had no obvious equivalent there. Alpoge has said Claude found a way forward by combining concepts from several branches of mathematics.
According to Alpoge, the process was collaborative. He identified promising approaches and directed the work, but as it progressed, Claude began making advances faster than he could make sense of them. “Claude far outpaced me near the finish line,” he said.
Why mathematicians are cautious
Independent mathematicians have yet to verify the result, but early checks are encouraging. Andrew Sutherland of the Massachusetts Institute of Technology has confirmed that the code Alpoge published passes several rigorous checks, and two separate proof-checking systems have also accepted the result.
Brian Conrey, a leading specialist on the Riemann Hypothesis, told AFP: “It’s really the biggest thing I think that’s ever happened in mathematics.” James Maynard, Regius Professor of Mathematics at the University of Oxford and a 2022 Fields Medal winner, said: “If this fully checks out, it is clearly a historic milestone in mathematics, establishing one of the great mysteries of the subject.”
Hours after Alpoge’s announcement, Stanford University mathematics professor Ravi Vakil said a postdoctoral researcher in the same department had independently submitted a proof of the hypothesis. Anthropic declined to comment on Friday as experts around the world rushed to check the work.
Solving the hypothesis would have no immediate practical application, but mathematicians describe it as a compass that has pointed the way to new findings for more than a century. Earlier work had established that at least 67.2 per cent of the relevant values follow the pattern the hypothesis predicts. The new claim, if correct, would establish that every one of them follows the pattern, without exception.
A year of AI maths breakthroughs
The claim crowns a remarkable year for using AI to solve open problems in mathematics. In July, Alpoge used an Anthropic model to disprove the Jacobian conjecture, a problem that had stood for 87 years. In August, Anthropic reported that an unreleased version of Claude had improved the longstanding bound on the share of zeros on the critical line from 41.6 per cent to 67.2 per cent.
In September, OpenAI said an internal AI system had produced a proposed solution to the Navier-Stokes problem, another Millennium Prize Problem about the movement of liquids and gases. In October, the company released hundreds of major proofs at once.
The timing of Alpoge’s announcement has drawn attention: his work on the problem came together late on Sunday night, just days before Anthropic staged the largest initial public offering in history on Thursday. Alpoge acknowledged that the timing could be viewed cynically, without elaborating.
Frequently Asked Questions
What is the Riemann Hypothesis?
The Riemann Hypothesis is a mathematical problem first proposed by Bernhard Riemann in 1859. It predicts a precise pattern in how prime numbers are distributed. It is widely regarded as the most important open question in pure mathematics and one of seven Millennium Prize Problems, each carrying a 1 million dollar prize.
What exactly has Anthropic’s researcher claimed?
Levent Alpoge, a mathematician at Anthropic, says he worked with the company’s Claude AI model to discover a proof of the Riemann Hypothesis, and of the broader Generalized Riemann Hypothesis. He released the proof as computer code in Lean, a proof-checking language, rather than as a traditional paper.
Has the proof been verified?
Not yet. Independent mathematicians are still checking the work. MIT mathematician Andrew Sutherland has confirmed the released code passes several rigorous checks, and two proof-checking systems have accepted it, but full independent verification will take time. A Stanford postdoctoral researcher has also independently submitted a proof.