Postmodernism for rationalists — LessWrong