LeCun: LLMs alone are not a path to human-level AI
ylecun · x · 2026-09-15
Yann LeCun clarified his position: LLM-based tools are useful and everyone uses them, but by themselves they are not a path to human-level AI. In his view, architectures that understand the real world are not LLMs — and the components that let current LLM systems interpret images, video and other real-world signals aren't LLMs either. He points to a one-hour talk with results and evidence backing these conclusions.
Related event: LeCun's Claim That LLMs Won't Reach Human-Level AI Sparks Debate(3 posts)→
More from AGI Musings
- Another Google AI Safety Researcher Quits, Warning Everyone Might Be About to Die — Puzzleheaded-King584 · 2026-09-15
- The Great AI Slowdown Pivot: Veteran Commenter Calls Labs' Pause Call Pure PR Spin — DavidLinthicum · 2026-09-15
- Stewart Alsop calls Netflix AI doomer documentary a crafted 'grift' — StewartalsopIII · 2026-09-15
- Netflix AI doom doc drops days before King Charles convenes OpenAI, Anthropic, DeepMind, Nvidia — StewartalsopIII · 2026-09-15
- Europe holds just 5% of global AI compute vs US 75%, warn 60+ experts in new strategy report — anselm · 2026-09-15
- Ben Goertzel Blasts Dario Amodei's AI Slowdown Proposal: 'Hard No' — bengoertzel · 2026-09-15