Rationality: Appreciating Cognitive Algorithms — LessWrong