Trope Dodging by Hazard (LW 2.0) – “I think that it is useful to model people as having filters that exist somewhere in their decision making process, filters which ask the question, “Does what I’m about to do match a behavior trope that I’ve blacklisted?” If the answer is yes, then the potential action gets vetoed and the brain goes back to the drawing board. ”
What Evidence Is Alphago Zero WRT AGI Complexity by Robin Hanson (LW 2.0) – Throughout the history of CS we have developed many ‘superhuman’ general tools. AlphaGo:Zero is built from a new tool in this family. However averaged over all fields progress toward AGI has been slow. It is unlikely that deep learning is the last tool we will need.
Politics and Economics:
Maybe Reality Actually Does Have A Liberal Bias by Ozy – A priori there is no reason to assume that liberal and conservative ideology are equally correct. So the under-representation of conservatives in universities cannot be assumed to be the result of bias.
TransTech Conference Round Up by Aceso Under Glass – The TransTech conference is about speculative science. In practice this conference was mostly about vendor fair, with some vendors looking for buyers and some looking for funding. Aceso reviews the products he tried.