Kaj_Sotala comments on AI risk bibliography (draft)