Conversations and talks that have shaped how I think about intelligence.
Really good MLST episode. Jeremy Howard is very clear here about where AI coding feels impressive and where people maybe start fooling themselves a bit.
Very good and important reminder that abstractions and concepts are still just models. They simplify things, but it helps to remember that the model is not the thing itself.
Very mathematical. I would be lying if I said I understand everything here, but I like that they are trying to build a deeper theory for deep learning instead of keeping research in that slightly alchemy-like mode forever.
More of a philosophical discussion about artificial intelligence and reasoning. There are a lot of very interesting takes in this one.
I am honestly a bit of an MLST fanboy. This page is mostly for keeping the episodes and talks I know I will want to come back to later.