22:36Machine-translated from Azerbaijani by AIRead in Azerbaijani
Mathematician Alpöge says he found a proof of the Riemann hypothesis with Claude

Mathematician Levent Alpöge, who works at Anthropic, claims to have proved the Riemann hypothesis with Claude. The arguments were written in Lean, software designed to verify mathematical proofs on a computer. The work has so far been published only on Alpöge’s website, and the solution has not yet undergone independent review. Anthropic was not involved in preparing the announcement and has not disclosed an official position on the result.
Bernhard Riemann proposed the hypothesis in 1859. The problem has remained open for 167 years. The hypothesis connects the distribution of prime numbers to the zeros of a special mathematical function. Its main condition has been checked for the first 10 trillion solutions, but no proof valid for all cases has been found. If the proof is confirmed, it could clarify many questions in number theory. Alpöge believes the method he developed also solves the generalized Riemann hypothesis. According to the mathematician, with additional work, the method could also be applied to the extended Riemann hypothesis. He wrote that he sent the result to Anthropic CEO Dario Amodei on Sunday night. Alpöge published the work after the IPO held on Thursday, when the company offered its shares for sale on a stock exchange for the first time. Anthropic announced in August that it was working on this problem with Claude, but it did not expect the selected methods to result in a proof. The Riemann hypothesis is one of the 7 Millennium Prize Problems. The Clay Mathematics Institute, which administers the program, offers a $1 million (approximately 1.7 million manat) prize for each solution. A candidate solution must first be published, undergo review for at least 2 years, and be accepted by the global mathematical community. So far, only the Poincaré conjecture has been officially solved among these problems. OpenAI had previously claimed to have solved the Navier–Stokes problem, which describes the motion of fluids. The company had spent millions of dollars on the work and deployed 10,000 artificial intelligence agents. Although the result was considered credible, it has not yet been officially recognized. Allegations that the work used unpublished research also sparked controversy. OpenAI denied the allegations. Alpöge also said in July that he had solved the Jacobian conjecture with Claude.
Anthropic announced that same week that its Claude Fable 5.2 model had made progress on 2 theoretical problems involving black holes. The first problem concerns a phenomenon described by Stephen Hawking and Don Page in 1983. It examines the sudden formation of a black hole in a heated theoretical universe. Previous calculations had been performed for a large universe. Claude showed that the transition also occurs suddenly in a small universe. According to the result, black holes can also form within the framework of quantum theory. The other result concerns a 2-dimensional theoretical universe and examines the possible limit for the lightest black hole. By finding a pattern in the calculations, Claude established the exact upper bound for a simple bootstrapping method. In this method, a result is built on and verified against previous steps. The model mathematically demonstrated that the bound could not be lowered any further. Anthropic emphasized that the results do not apply to our universe and currently have no practical use. Nevertheless, they could help explain the relationship between quantum mechanics and the theory of relativity.
