A mathematician at the artificial intelligence company Anthropic says he believes he has proved the Riemann hypothesis, the most famous unsolved problem in mathematics, with help from the company's AI model Claude. The claim, posted on Friday, November 20, 2026, has set off a surge of public interest, with searches for "Riemann hypothesis" spiking on Google Trends. But the proof has not been confirmed, and no independent mathematician has yet verified it.
"I believe I have, working with Claude, discovered a proof of the Riemann Hypothesis," Levent Alpöge wrote in a statement posted on his personal website, as reported by Scientific American. Alpöge is a number theorist who works at Anthropic. He posted the claim himself, on his own website, rather than through a company announcement, and Anthropic has not issued a company statement about it.
The Riemann hypothesis was first proposed in 1859 by the German mathematician Bernhard Riemann. It concerns the distribution of prime numbers, the numbers divisible only by one and themselves. Primes look scattered at random along the number line, but the hypothesis says they follow a hidden pattern. In mathematical language, it says that all the non-trivial zeros of the Riemann zeta function lie on a single "critical line," where their real part equals one half.
The problem has guided number theory for more than 165 years, and most leading experts considered it far out of reach. It is one of the seven Millennium Prize Problems named by the Clay Mathematics Institute in 2000, each carrying a one-million-dollar prize, which can only be claimed after publication in a journal and two years of peer scrutiny. Countless results assume the hypothesis is true, so a confirmed proof would settle them at once. Alpöge says his argument also confirms the Generalized Riemann Hypothesis, a much broader version of the claim.
What Alpöge has released is unusual: a four-page note on his website, plus the proof written as code in Lean, a proof assistant that checks whether every step of an argument follows from the definitions and results it is given. No formal research paper has been published yet. His note says the full written argument would run to more than 250,000 pages, which is why he released the Lean formalisation instead.
That format is both a strength and a reason for caution. A Lean check can confirm that each step follows from the last, but it cannot by itself show that the statement being proved is really the Riemann hypothesis. Experts must now check that the hypothesis was coded correctly and that no hidden assumption slipped in — the checklist set out by mathematician Kit Yates, as reported by Nacionale. As of Friday, no outside mathematician had reviewed the work.
The claim follows a run of AI-assisted mathematics from Alpöge in 2026, including a Claude-generated solution to the Jacobian conjecture in July. Alpöge earned his doctorate at Princeton under Manjul Bhargava in 2020 and was a Junior Fellow at Harvard before joining Anthropic.
For now, the Riemann hypothesis remains an unproven problem, not a settled breakthrough. Mathematicians will examine the Lean code, test whether it really proves the hypothesis as stated, and, if it survives that scrutiny, work toward a readable paper and formal peer review. Even if confirmed, the proof would have no immediate practical use. As Scientific American notes, it will not upend global banking or make the internet run faster. Alpöge struck a reflective note, writing that what he saw "so amazed me that I believe it is one dawn across science of something far greater."




No comments:
Post a Comment