Reply to Jebari and Lundborg on Artificial Superintelligence — LessWrong