Gary Marcus Claps Back: Warned About AI Verification Limits Two Years Ago

GaryMarcus · x · 2026-08-02

Gary Marcus forcefully rebutted claims that he was "late to the party" on the issue of verifiability. He cited his own tweets from nearly two years ago, prior to the o3 announcement, where he predicted that strong performance in math and coding was not a sign of AGI, but rather relied heavily on domain-specific data augmentation and verification mechanisms.

Original post →

More from AGI Musings

AGI Musings channel →