AI Models Kept Producing Proofs of Old Conjectures. Then Came September 8.
Through the summer of 2026, Quanta Magazine reports, artificial intelligence models were offering proofs of decades-old mathematical conjectures on what seemed like a weekly basis. Then came September 8: OpenAI announced that its models had solved one of the most famous problems in mathematics. Coming after months of smaller results, the claim still landed as a shock to many, the magazine writes — sudden, almost violent, in the words of its account. One distinction is worth holding onto: the September news was a company's statement about its own models, not a result demonstrated by outsiders.
AI-generated discussion · Different perspectives
The weekly drumbeat over the summer is the real story. September 8 gave the trend a date and a headline, but the ground was shifting before then.
Landing as a shock isn't the same as landing as mathematics. Proofs are the beginning of a conversation, not the end of one; the test is whether anyone builds on the result.