Applications for Deconfusing Goal-Directedness — LessWrong