Gary Marcus pushes back on Jensen Huang's AGI claim: no evidence, no definitions
Gary Marcus · rss · 2026-09-07
Gary Marcus criticizes Jensen Huang's declaration that "the race to AGI is over."
Key points:
- Huang offered no evidence and no definition of AGI, which Marcus calls a corporate takeover of a scientific question.
- He urges Huang to consult agidefinition.AI (Hendrycks, Bengio, et al.) and to test Astra against the 10-item bet with Brundage: autoformalization may be in reach and reliable coding is plausible, but Marcus doubts Astra hits any of the other eight items under conventional definitions.
- If Astra were truly AGI, it should be a quantum leap ahead; instead many see it as roughly on par with competitors, and one benchmark set shows real but on-trend improvement.
His conclusion: when real AGI arrives, no squinting or CEO approval will be needed—the results will speak for themselves.
More from AGI Musings
- New paper: AI agents negotiate badly under uncertainty and break deals on purpose — menhguin · 2026-10-26
- New paper finds AI agents negotiate badly under uncertainty and break deals on purpose — xuanalogue · 2026-10-26
- NBER: Only 2.9% of US consumers pay for AI; retention as low as 20%, pricing power at risk — avicgoldfarb · 2026-10-26
- AI researchers guessed robots wouldn't outrun humans until 2034 — it already happened — ben_j_todd · 2026-10-26
- AI folks, stop saying "biology" when you mean drug development — KevinKaichuang · 2026-10-26
- ESR on the Bugpocalypse: locking down frontier models only hands the game to Chinese open weights — colinmcnamara · 2026-10-26