Levent Alpöge, a mathematician at Anthropic, claims he has found a Lean-compilable proof of the Riemann hypothesis, a conjecture dating to 1859. Several participants in an expert chat say they have compiled his solution in Lean, but humans still need to understand and explain the argument.

Levent Alpöge is having a big week. His employer, Anthropic, became the first publicly traded frontier AI lab. Alpöge, a mathematician on the Anthropic side, also says he has solved one of the most famous problems in mathematics.

In an essay announcing the result, Alpöge wrote that “the precise timing is horrible.” He described messaging Anthropic CEO Dario Amodei to ask how he should announce that he had found a Lean-compilable solution to the Riemann hypothesis. The conjecture dates to 1859 and is among the Millennium Prize math problems, a set of challenges with a $1 million reward for each solution.

The hypothesis proposes that a hidden structure lies beneath the appearance of prime numbers. Proving it would be a landmark result, and the reaction from the mathematics community has been intense. Much of that reaction is playing out in a chat on Zulip, where experts are working through the argument in real time. Chris Birkbeck, a number theorist at the University of East Anglia, has been among those posting. The proof leans on a dense set of tools from modern number theory, including Galois representations, base change, Clozel’s purity and the main theorem of complex multiplication.

Has the Riemann hypothesis proof been verified?

It will take a while before mathematicians fully grasp the proof. Still, most participants in the chat appear confident that the Lean version will hold up. Lean is an open-source coding language used to express mathematical ideas as code, so that proofs can be formally verified by a computer. Several participants, including MIT mathematician Andrew Sutherland, say they have successfully compiled Alpöge’s solution.

Alpöge admitted in his essay that he felt a bit guilty about releasing the proof without a human-readable explanation. He said his lifelong fascination with the problem left him feeling he “could not live” if he waited to share the finding. He described the process as a “stream-of-consciousness jaunt through the experience of finding all the pieces of the argument through incredible confusion with Claude.” He wrote that the model “flew, and danced, and sparkled near that finish line.”

What does a Lean-verified proof mean?

A computer-verified proof is not the same as an understood one. For the result to mean much to the field, humans need to understand it and explain it to one another. Alpöge has committed to pursuing that work.

Alpöge and NYU mathematician Tristan Buckmaster also released another significant proof, this one related to the Navier-Stokes equations. Those equations have been another centerpiece of the push by frontier AI labs and mathematicians to crack long-standing open problems this summer.

Did a Stanford team also post a Riemann hypothesis proof?

Meanwhile, Stanford professor Ravi Vakil posted a proof of a generalized version of the Riemann hypothesis written by Stanford postdoc Alessandro Masullo. Vakil stressed that the work is entirely human. “It is all human, with no AI,” he wrote. “The hope was to bring Alessandro to Stanford for a year to polish the writing and then announce it after running it by enough people. Now it seems best to put out the current draft.”

Without a Lean version, it is too early to say whether Masullo’s proof holds. Masullo said the work reflects 12 years of effort. In a short introduction, he addressed mathematicians who feel discouraged by the current climate. “If any message at all this work is meant to send to the mathematician who feels discouraged by the pace and tone of the current climate, it is this: we are still here,” he wrote. “We are very much alive. We are not going anywhere.”

Alpöge’s Lean proof has now been checked by independent sources, which adds weight to his claim. The full mathematical picture will depend on how many experts can work through the arguments and explain them to their peers.

Este artigo foi escrito com a assistência de IA.
News Factory APP - notícias agênticas para impulsionar seu SEO e AEO.