"RL creates split personas" by Jan Betley

od LessWrong (Curated & Popular)