Optimizing the Twelve Virtues of Rationality — LessWrong