LeCun: LLMs are mostly information retrieval systems, not thinkers

ylecun · x · 2026-09-27

Yann LeCun, now Executive Chairman of AMI Labs, argues most people misread LLMs by anthropomorphizing systems that merely reproduce human functions. "Except for a few domains, LLMs are mostly information retrieval systems — they compress a lot of factual knowledge previously produced by humans and give easy access to it." The knowledge comes from people; what the system adds is compression and retrieval. He situates LLMs in a long historical lineage of such systems.

Original post →

More from AGI Musings

AGI Musings channel →