LeCun: There's 'no way in hell' scaling LLMs gets us to AGI

aran_nayebi · x · 2026-09-26

Quoting Asimov's line that a distinguished elderly scientist declaring something impossible is 'very probably wrong,' Ariel Kwiat highlighted Yann LeCun's blunt stance: scaling LLMs alone will never reach human-level intelligence or AGI — 'no way in hell.' The idea of 'a country of geniuses in a data center' is 'complete BS.' An LLM is a system with gigantic memory and retrieval ability: it may feel like a PhD beside you, but it can't invent solutions to genuinely new problems.

Related event: LeCun Insists Scaling LLMs Alone Will Never Reach AGI(2 posts)→

Original post →

More from AGI Musings

AGI Musings channel →