Story perspectives
Yann LeCun Challenges AI Path: Visual Data Over LLMs
11/18/2025
24 4
1 of 1
Story summary
- Yann LeCun, a prominent AI researcher at Meta, said large language models do not lead to human-level intelligence.
- Speaking at an event in Brooklyn, he urged a shift toward "world models" using visual data.
- His remarks come amid speculation about Meta's future, as the company continues heavy investment in LLMs.
- LeCun's comments reflect a broader AI industry debate about the direction of the technology and persistent uncertainty.
