Harry Sue and The Methods of Rationality — LessWrong