Yann LeCun from Meta dismisses concerns about A.I. posing an existential threat, calling them "complete nonsense."

Artificial Intelligence pioneer Yann LeCun doesn't believe artificial intelligence is actually on the verge of becoming intelligent.
Yann LeCun from Meta dismisses concerns about A.I. posing an existential threat, calling them "complete nonsense."

Artificial Intelligence pioneer Yann LeCun doesn't believe artificial intelligence is actually on the verge of becoming intelligent.

A.M. Turning Award-winner and senior researcher at Meta, Professor at New York University LeCun has often expressed his skepticism. "We need to have the beginning of a hint of a design for a system smarter than a house cat" before we worry about controlling super-intelligent AI, he tweeted this week.

He elaborated on his opinion in an interview with the Wall Street Journal, where he reflected on the question of A.I. being so intelligent that it threatens to take over the world by saying, "You're going to have to pardon my French, but that's complete B.S.

To this, LeCun finally concludes that the real large language models of today are never at a level of some cat-level abilities such as persistence in memory, reasoning, planning and understanding of the physical world. For him, LLMs only prove that "you can manipulate language and not be smart" and will never lead to true artificial general intelligence or AGI.

He is not a full-on skeptic of AGI, though. He does believe that new approaches will be needed. As one example, he cited work that his Fundamental AI Research team at Meta has done in digesting real-world video.

 

Blog
|
2024-10-13 19:14:33