#258 – Yann LeCun: Dark Matter of Intelligence and Self-Supervised Learning
Lex Fridman Podcast
00:00
AI works largely functions by first principles
In classical optumal control, there is something called a moral predictive control. This was invented in the 19 sixties by obtimal control theorists. A lot of planning systems in robotics are based on this and er. And you can think of this as a form of reasoning. It's basically planning. We don't know how to do that with machines yet. Wharper: The big question of a i, i think, for the next ad is, how do we get machines to run predictay models to the world?
Play episode from 30:04
Transcript
