The theory on the thread that

My guess is that this represents a shift in focus at OpenAI away from math: they had their big meaningless trophy prize with the Navier-Stokes solution, they appear to have distanced the competition (Anthropic) and the public bought the notion that “math is solved”

seems like a plausible explanation? I'll guess we'll if OpenAI moves on now that they've "solved" math or if they follow up with a few more drops of "proofs".