Mathematician Alex Kontorovich: IMO is 'too easy' for AI now — labs will quit headline math next

AlexKontorovich · x · 2026-09-05

Rutgers math professor Alex Kontorovich recounts: two summers ago AlphaProof missed an IMO gold medal by one point; last summer half a dozen labs claimed gold-level performance. He predicted zero systems would chase gold this summer — because nobody cares anymore, and the IMO reverts to something 17-year-olds do.

He says he was largely right: the IMO has gone quiet because it's 'just too easy.' His next prediction: the number of major math problems labs publicize solving next summer will also be zero — not because they can't, but because such problems will return to being the work of professors, postdocs, and grad students.

Original post →

More from AGI Musings

AGI Musings channel →