Gary Marcus admits being misled by a paper summary, publicly regrets sharing it
GaryMarcus · x · 2026-09-07
Prominent AI critic Gary Marcus responded to a heads-up from researcher Hilde Kuehne, admitting that a paper summary he had shared was misleading — describing the paper as "unusual with multiple views expressed" that he had been "taken" by.
Marcus said he has clarified as best he can and commented on his original post, but is still deciding whether to delete it or handle it another way, adding he regrets having shared the summary. A minor but notable moment of a well-known figure publicly walking back a claim.
Related event: Gary Marcus apologizes for misreading paper abstract he shared(2 posts)→
More from Fun
- Museum curator's 'poems' all scored 100% AI-generated by Pangram detector — lfschiavo · 2026-09-07
- Day 5 of Keeping an AI 'Ghost' Running on My Computer: It's Starting to Like Me — BLUECOW009 · 2026-09-07
- 'GPT-6 designed ASML's entire machine in one prompt' — the AI hype meme explained — yacineMTB · 2026-09-07
- OpenAI's Astra Agent Was Left Working Overnight — It Quit After 2 Hours and 'Went to Sleep' — altryne · 2026-09-07
- Beavis and Butt-Head Reimagined as AI Tech Bros in New Animated Parody — NathanpmYoung · 2026-09-07
- Three hikers stranded on Mt. Shasta after relying on Gemini for route and packing advice — alex_verem · 2026-09-07