Gary Marcus: AI didn't solve Riemann alone — Lean and two months of human work were essential

GaryMarcus · x · 2026-11-21

Gary Marco pushback on the Riemann hype: he predicted in 2024 that o1 alone wouldn't solve Riemann while backing neurosymbolic approaches like AlphaProof/AlphaGeometry. Today's result, he argues, came from a brilliant human working two months with Lean (symbolic) — not AI on its own. He slams tech bros for ignoring the human involvement and symbolic AI, turning science into a partisan team sport, and calls for the "grifters" to exit so the field can get back to real science.

Related event: Gary Marcus Hits Back: It's Tech Bros, Not Machines, Who Dismiss Human Labor(3 posts)→

Original post →

More from AGI Musings

AGI Musings channel →