Curtis Yarvin on A Theory of Pervasive Error — LessWrong