Joel Becker comments on Eli’s review of “Is power-seeking AI an existential risk?”