Konstantin Kakaes at Quanta Magazine:

The Riemann hypothesis has long stood as the most prominent open question in mathematics. On its surface it deals with when a particular function, called the zeta function, equals zero. More deeply, it concerns how prime numbers are distributed. It also has far-reaching consequences across numerous areas of mathematics.
On the morning of Friday, November 20th, Levent Alpöge, a mathematician at the artificial intelligence company Anthropic, uploaded a plaintive four-page paper to his personal website in which he wrote, “I believe I have, working with Claude, discovered a proof of the Riemann Hypothesis.” Because his attempts to get Claude (Anthropic’s AI model) to produce human-readable text amounted to over 250,000 pages, he has not shared such a text. He has, however, made available 17 million lines of code in the programming language Lean which he says contain a formal verification of the proof, which is of a stronger result called the generalized Riemann hypothesis.
Lean experts around the world are racing to independently verify that the code works as he claims.
