Anthropic employee Levent Alpoge announced that with the help of the company's Claude AI tool he was able to prove the Riemann hypothesis, a famous mathematical problem that has remained unsolved for more than 150 years. He reported this in a brief explanation on his website.
The Riemann hypothesis, formulated in 1859, concerns the distribution of prime numbers, those divisible only by one and by themselves. According to Alpoge, the insight came to him late in the evening on Sunday. He wrote to Anthropic CEO Dario Amodei to ask whether he could publicly announce the result, since the company was preparing for a stock market listing.
A day after the share offering, Alpoge disclosed that he had solved the problem that had resisted more than a century and a half. Anthropic did not respond to a request from the Associated Press for comment.
The proof is presented as millions of lines of code in Lean, a programming language that allows mathematical proofs to be written formally so that a computer can verify them. Such results are usually published as a scientific paper, notes Alex Kontorovich, a professor and head of the mathematics department at Rutgers University. According to him, when reading a paper, a mathematician traces each step himself and satisfies himself that it is correct.
Kontorovich added that Claude went through this code and confirmed that the proof is correct. At the same time, the scientist acknowledged that this does not mean he himself already understands the proof. A computer is extremely exacting, far more exacting than a human, and if a program accepts a proof, the probability that a mathematician will also accept it is very high. But the work remains to be done.
Several hours after the announcement, the Riemann hypothesis was still listed as unsolved on the website of the Clay Mathematics Institute. In 2000 the institute included it in its list of seven Millennium Prize Problems and set an award for its solution. The verification process takes time, explained the institute's president, Martin Bridson. He called the announcement historic but stressed that the path from it to a true understanding will be long and difficult.
Alpoge himself acknowledges that he is presenting to the world a complete argument without yet having mastered the vast area of mathematics around it. In September, OpenAI said its technology played a key role in a similar breakthrough with the Navier-Stokes equations, which describe the flow of fluids.
This week Anthropic made its debut on Wall Street. The company's CEO, Dario Amodei, paints an optimistic future for AI in which the technology will lead to significant progress in mathematics and science, while warning of potential risks.