Yann LeCun: Scaling LLMs alone will 'absolutely never' reach AGI
MickeySteamboat · x · 2026-09-28
Meta chief AI scientist Yann LeCun reiterated his skepticism about the LLM scaling paradigm, saying there is "absolutely no way in hell" scaling LLMs alone reaches human-level intelligence or AGI.
Key points:
- The idea of "a data center full of geniuses" is, in his words, "complete BS."
- LLMs may feel like a PhD sitting beside you, but they are fundamentally systems with enormous memory and retrieval capabilities — not systems that can genuinely solve problems they have never seen before.
The remarks are consistent with LeCun's long-standing position that current autoregressive LLMs lack true reasoning and world understanding, and that a different architecture (such as his world-model approach) is needed for higher-level intelligence.
Related event: LeCun: Scaling LLMs Alone Will Never Reach AGI(2 posts)→
More from AGI Musings
- Fermi estimate: brain may pack 500k-5M molecular switching units per 'parameter' — JosephJacks_ · 2026-09-28
- Pundit mocks 'a million digital employees' fantasy: you don't manage what outsmarts you — danfaggella · 2026-09-28
- Gen-Z is starting to admit they enjoy AI-generated content — LaillaElgohary · 2026-09-28
- Austin car theft rose 24% vs trends after switching off Flock license plate cameras — garrytan · 2026-09-28
- "Alignment is easy, just look at Opus 3"? Researcher pushes back: capable systems will break it — JacquesThibs · 2026-09-28
- Investor: Musk keeps winning the few irreversible decisions that actually matter — gabriel1 · 2026-09-28