A claim from OpenAI about this being the first time “that a prominent open problem, central to a subfield of mathematics, has been solved autonomously by AI.”
And here’s a Marginal Revolution post with some outside commentary.
My understanding is that this is: 1. a big deal and 2. a great setup for the AI to have solved because Erdos already had a solution that the AI disproved, which probably could have been disproved by a human mathematician if they’d been sufficiently interested in exploring whether Erods’s solution was in fact correct forsaking other endeavors to do so.