WRLD

Search

#Théorèmes
Frenchie Wrld summary 8h

Frenchie reported that Levent Alpöge, a mathematician at Anthropic, published a note on Friday in which he claimed to have discovered, with Claude, a proof of the Riemann hypothesis. According to Frenchie, Alpöge believed that the same method could also solve the generalized Riemann hypothesis. He recalled that the problem was formulated by Bernhard Riemann in 1859 and was among the seven Millennium Prize Problems from the Clay Institute, with a prize of one million dollars. He added that only the Poincaré conjecture, solved by Grigori Perelman, had been officially resolved so far. Frenchie explained that the hypothesis dealt with the non-trivial zeros of the zeta function and their alignment along the line where the real part equals 1/2. He estimated that a validated proof would render many mathematical results formulated under this hypothesis true, and that its generalized version would allow even more. He also stated that a proof would not be sufficient to break RSA, as it would not provide a method for factoring very large numbers. According to Frenchie, Alpöge had not published a traditional paper but rather a proof in Lean, spread over 14 archives and intended to be verified by computer. He would have credited Claude with several important steps and said that the model had largely surpassed him towards the end. Frenchie indicated that a human-readable version was expected later and recalled that Alpöge had refuted the Jacobian conjecture with Claude in July. However, Frenchie emphasized that the announcement had not yet been confirmed by anyone outside Alpöge's circle: neither the code compilation nor the correspondence between the formalized theorem and the Riemann hypothesis had, according to him, been verified. He also pointed out that the note appeared the day after Anthropic's IPO, following a difficult first day, while specifying that Alpöge himself rejected this cynical interpretation. Finally, Frenchie mentioned a proof by OpenAI, in October, of what he called the quasi-Riemann hypothesis, as well as comments from Alex Kontorovich of Rutgers, who reportedly believed that such a result would immediately merit a Fields medal if it came from a human. He also cited Terence Tao on the growing abundance of proofs and mentioned Yang-Mills, Navier-Stokes, and the Riemann hypothesis among the great mathematical problems recently discussed. #Mathématiques #HypothèseDeRiemann #Théorèmes (translated)

2567
BFMTV 4d

A researcher used artificial intelligence to analyze the unpublished works of Kurt Gödel, revealing hundreds of theorems, one of which could have earned him a Fields Medal. #IntelligenceArtificielle #KurtGödel #Mathématiques (translated)

www.bfmtv.com
Kurt Gödel, génie des mathématiques, a laissé à sa mort des centaines de pages de démonstrations: grâce à l'IA, un chercheur a analysé ses travaux, dont un aurait pu lui valoir une médaille Fields
3151