Gary Marcus challenges Jensen Huang's AGI bar: who decides what 'broad enough' means?
GaryMarcus · x · 2026-09-07
Citing Jensen Huang's 2024 Stanford remarks, @YuchenjUW argues that by Huang's own stated definition, AGI has in some sense already arrived.
Gary Marcus pushes back: what does "broad enough" even mean, and who decides? He argues it quietly grants many exemptions from the original "full range of human cognitive tasks" bar without argument — while still thanking him for surfacing the source.
The exchange highlights a core tension: when a lab's own AGI milestone may already be met, definitional fuzziness becomes the escape hatch.
More from AGI Musings
- Was the $1M Millennium Prize implicitly a cap on how much you can spend on a solution? — danielrock · 2026-09-09
- a16z GP: agents can't 'make me a million dollars' — human direction still unlocks the next hill — lennysan · 2026-09-09
- Guodzh: the intelligence revolution is here, but it's a non-zero-sum game far from over — Guodzh · 2026-09-09
- Quintin Pope: state-backed cybercriminals misusing AI pose a bigger threat than rogue labs — QuintinPope5 · 2026-09-09
- Quintin Pope on why AI-2027-style narrative forecasts are the wrong format — QuintinPope5 · 2026-09-09
- AI will hand us breakthroughs we can't understand, like alien answer sheets — sterlingcrispin · 2026-09-09