On Quirrell's theory of human nature:

It seems to me that in general, his theory yields pretty accurate predictions, and to see what's wrong with it you probably need to know some evolutionary psychology; Eliezer's Adaptation-Executers, not Fitness-Maximizers. Evolution is going to favor genes for convincingly acting like you care about your friends without incurring too many costs for their sake. Yeah, it's also going to favor genes for seeing through easily faked signals, but to the extent that people can get away with it, it's going to favor genes for seeming to care about your friends without incurring the costs that would go along with that. The twist is that evolution while evolution produces behaviors that tend in that direction, it doesn't produce creatures who follow the strategies exactly, or consciously. The result is people who actually care about their friends... even if they very often don't act like it.