A mathematician forced the neural network to prove the Riemann hypothesis. . An anthropophage collaborator posted a four-page note stating that he and chatbot Claude had found a proof of the Riemann hypothesis. The problem goes back to 1859, concerns the distribution of prime numbers and is on the "Millennium Problems" list with a million dollar prize. The work lasted about 6 weeks and a non-public version of the model was used. The final result of Alpege assumed the Lean language, which allows you to test the mathematical line of thinking by line. The code is over 250,000 lines. The writer himself called the event a "race for science", and admitted that at the finish line Claude was faster than him. The independent review has not yet been completed. Rens the Tao mathematician noted that it would take time to understand, although the presence of a living scribe makes the task easier. If approved, this would be one of the biggest breakthroughs in the history of mathematics. Follow Max I Feedback.