LeCun: Scaling LLMs to AGI Is 'No Way in Hell'; Researcher Pushes Back

aran_nayebi · x · 2026-09-26

Yann LeCun doubled down on his skepticism: "There's absolutely no way in hell" scaling LLMs yields AGI, calling the "country of geniuses in a data center" idea complete BS. In his view, LLMs are systems with gigantic memory and retrieval ability, not systems that invent solutions to new problems — "it feels like a PhD next to you, but it's not a PhD."

Nathan Calvin pushed back, noting that after the current AI paradigm solved a Millennium Prize Problem, such claims should finally be put to rest.

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

Original post →

More from AGI Musings

AGI Musings channel →