Artificial intelligence helped solve one of the most famous mathematical mysteries, for which a million dollars was promised
A mathematician from the American company Anthropic stated that, together with the artificial intelligence Claude, he found a proof of the Riemann hypothesis—one of the most famous mathematical mysteries, unsolved for 167 years. If the result is confirmed, it will mark a historic breakthrough in mathematics.

As The Wall Street Journal reports, on November 20, Anthropic researcher Levent Alpöge published on his website a purported proof of the hypothesis that mathematicians have been working on since 1859. According to him, the result was achieved after several months of collaboration with Claude.
The hypothesis was formulated in 1859 by the German mathematician Bernhard Riemann. It relates to the distribution of prime numbers—those that are divisible without a remainder only by one and themselves, such as 2, 3, 5, 7, and 11.
Mathematicians have long known how the number of primes approximately changes among increasingly large natural numbers. However, describing deviations from this pattern with precision is far more difficult. The Riemann hypothesis concerns the properties of a special mathematical function and, if proven, would significantly refine such calculations and confirm a number of other mathematical statements.
In 2000, the Clay Mathematics Institute included it in its list of seven "Millennium Prize Problems"—exceptionally difficult mathematical problems, each carrying a one-million-dollar reward for its solution. Until now, neither mathematicians nor the most sophisticated artificial intelligence models had managed to crack the hypothesis.
Alpöge presented the purported proof in an unusual format: as software code in the Lean programming language, which allows a computer to verify the logical correctness of mathematical reasoning step by step.
The researcher himself admits that the proof is based on a vast amount of complex mathematics, which he is still working to fully understand. When he attempted to generate a standard text exposition using AI, the result was an virtually unreadable document of over 250,000 pages. Alpöge now plans to prepare a more accessible explanation of his work.
According to the mathematician, the key result was obtained late in the evening of November 15, just days before Anthropic went public. The researcher even reached out to company chief Dario Amodei to coordinate the publication. Ultimately, he announced the discovery the day after Anthropic's stock market debut.
The announcement came amid rapid progress in artificial intelligence within mathematics. In September, OpenAI reported an alleged solution to another Millennium Prize Problem—the Navier–Stokes problem, which concerns the motion of fluids and gases. In October, the company presented hundreds of new mathematical proofs. Among them were findings from research into the quasi-Riemann hypothesis—a relaxed version of the famous problem that imposes less rigid mathematical constraints.
Renowned mathematician Terence Tao, a recipient of the Fields Medal—an award often referred to as the Nobel Prize of mathematics—hailed the progress achieved in this direction as the most significant in the last 50 years.
At the same time, Tao suggested that humans working alongside artificial intelligence might only manage to find a complete proof of the Riemann hypothesis by 2050. Yet less than a month later, an Anthropic employee claimed that such a proof had already been found.
However, Alpöge's claim has yet to be verified by independent mathematicians. Even if a computer confirms the logical validity of the proof, it is still necessary to ensure that it genuinely solves the problem posed. Under the Clay Institute's rules, to claim the million-dollar prize, the solution must be published and achieve widespread acceptance within the scientific community no sooner than two years after publication.
If Alpöge's proof withstands scrutiny, it will not only solve one of the greatest mathematical puzzles, but will also demonstrate just how fundamentally artificial intelligence can reshape the very process of scientific discovery.
Now reading
Why do Belarusians buy few books in Belarusian? Andrei Yanushkevich on thousands of books in storage, King’s flop, and the prospects of a Pazniak biography
Why do Belarusians buy few books in Belarusian? Andrei Yanushkevich on thousands of books in storage, King’s flop, and the prospects of a Pazniak biography
«I gave my word to both my father and her family that I wouldn't leave her.» A political prisoner and a Ukrainian woman met in an online game — and reunited in Poland
«I gave my word to both my father and her family that I wouldn't leave her.» A political prisoner and a Ukrainian woman met in an online game — and reunited in Poland
“They blackmailed me with ‘your wife's future and her mother's condition.’” Husband of Volha Kharlamava, sentenced to 10 years, spoke about KGB recruitment attempts
Comments
чего не заявишь, чтобы поддержать падающие акции