Artificial intelligence is now solving some of the hardest coding problems, is being used in medical research, and is even cracking problems that have resisted solution for years. The latest example comes from a Turkish mathematician working at Anthropic.
Levent Alpöge, a mathematician at Anthropic, the developer of the Claude AI model, has said he believes he solved the Riemann hypothesis, which has been unproven since 1859, with the help of Claude. Alpöge published the proof on his personal website in a programming language that computers can verify. The proof has not yet been confirmed.
He claims to have solved a 167-year-old problem
According to a report by The Wall Street Journal, Alpöge was working with AI to find a solution to one of the most famous unsolved problems in mathematics. The hypothesis predicts how prime numbers are distributed, and it has been among the questions mathematicians have pursued for 167 years.
On his site, Alpöge wrote: "In this age when mathematics is undergoing a revolution, I hope this argument offers a glimpse of the wonders that lie ahead and shows us the joy and wonder of discovering things that are about to be found in the simplest of things."
If confirmed, it would be the second such problem
The Riemann hypothesis was chosen in 2000 as one of seven Millennium Prize Problems, which are known to be extremely difficult to solve. If Alpöge's solution is verified, it would be the second Millennium Prize Problem solved by AI. According to the WSJ, humans have solved only one of these problems in 26 years, while AI models would have solved two within three months.
Anthropic declined to comment on the matter.
EDİTÖR NOTU: Kaynakta talimat benzeri metin vardı.