Almost exactly a year ago, @sama gave a speech at YC where he said, almost in passing, that AI would solve a Millennium Prize Problem 'almost certainly' by the end of 2026. To give an idea of how hard that is, the median age of a Millennium Problem (from it being first posed) is 76 years. Only one has been solved since the list was formalized in 2000, and the others look as intractable as ever. The mathematician in me scoffed at this. Sam and OpenAI had clearly misunderstood the difficulty of these problems. Today, I saw a video featuring @wtgowers, Fields Medallist and one of our undergrad lecturers. “There is no doubt that the solution to the unit-distance problem is a milestone in AI mathematics: if a human had written the paper and submitted it to the Annals of Mathematics and I had been asked for a quick opinion, I would have recommended acceptance without any hesitation. No previous AI-generated proof has come close to that.” Time to eat my words. AGI is coming.