OpenAI claims it solved an 80-year-old math problem — for real this time

TechCrunch
OpenAI's new reasoning model has autonomously disproved a long-standing geometry conjecture, gaining verification from prominent mathematicians.

Summary

OpenAI has announced that its latest general-purpose reasoning model has successfully solved a significant mathematical conjecture, originally proposed by Paul Erdős in 1946. Unlike previous unsubstantiated claims by the company regarding solved math problems, this new proof has been supported by notable mathematicians including Noga Alon, Melanie Wood, and Thomas Bloom. The model discovered a new family of constructions that challenges the long-held belief that optimal solutions must resemble square grids. This achievement is considered a milestone, as it demonstrates the ability of AI to autonomously perform complex, multi-step reasoning, suggesting broad future applications in fields such as physics, biology, and engineering.

(Source:TechCrunch)