Uncertainty in Deep Learning

Link post

Here is a PhD thesis I just got linked at work. I’m about halfway through the first chapter and it’s starting a discussion of practical near term AI safety.

Edit: I think linkposts are mildly broken?