Thomas comments on Trapping AIs via utility indifference