[SEQ RERUN] Selecting Rationalist Groups — LessWrong