Obstacles on the Path to AI

"There is no way in hell that you can learn billions of parameters with RL." I really love LeCun's provocative stances, but I get suspicious when people talk about impossibilities. RL is making huge strides. People adopted the same tone with neural nets years ago, and LeCun proved them wrong...

Everytime I read something like this, I get sad that I don&#x27;t understand most of it.<p>But then I get happy because at least I understand a little bit :)

The talk:<p><a href="http:&#x2F;&#x2F;techtalks.tv&#x2F;talks&#x2F;whats-wrong-with-deep-learning&#x2F;61639&#x2F;" rel="nofollow">http:&#x2F;&#x2F;techtalks.tv&#x2F;talks&#x2F;whats-wrong-with-deep-learning&#x2F;616...</a>

That&#x27;s only the slides. Is there a video of the talk?(assuming there is a talk, that is)

Although I&#x27;m working on deep neural nets, this material is too advanced to me. Looks like deep nets + bayesian reasoning is the next big thing.

The real obstacles to the path to AI is that we don&#x27;t even have the right questions to ask and people are looking for answers to the wrong questions and announcing the next big thing in AI.

I like the focus on vector space.