Yann LeCun again challenged the dominant consensus in artificial intelligence with a $1 billion bet on JEPA, an architecture that seeks to replace text generation with models able to learn world representations and anticipate consequences. His thesis is frontal: LLMs are powerful to manipulate language, but insufficient to build reliable genetic systems. *** Yann LeCun argues that LLMs stand out in language, but lack real models of the world to …
This story is only covered by news sources that have yet to be evaluated by the independent media monitoring agencies we use to assess the quality and reliability of news outlets on our platform. Learn more here.
Yann LeCun again challenged the dominant consensus in artificial intelligence with a $1 billion bet on JEPA, an architecture that seeks to replace text generation with models able to learn world representations and anticipate consequences. His thesis is frontal: LLMs are powerful to manipulate language, but insufficient to build reliable genetic systems. *** Yann LeCun argues that LLMs stand out in language, but lack real models of the world to …