Gary Marcus responds to 'you can't claim AI is both dangerous and useless' critique

GaryMarcus · x · 2026-09-08

A critic points out the apparent contradiction in Gary Marcus's posting: half his posts say "AI is dangerous, we must stop it" while the other half say "AI is bullshit, it will never be smart, these companies will crash." Marcus replies that he has "answered this hundreds of times," implying his positions are internally consistent (e.g., current AI hype vs. genuine risk are distinct claims), though he doesn't elaborate in the post. A recurring skirmish in the ongoing AI discourse around Marcus's stance.

Original post →

More from AGI Musings

AGI Musings channel →