It seems he avoids the harder question by using a relatively weak view of what "AGI" can do (he says: "It will not have the biggest new ideas, it will require lots of human supervision and direction, and it will be great at some things but surprisingly bad at others.")
Which is exactly what you'll get from every frontier lab ceo. My point was that they have been asked this question many times before, because parent comment said they wonder why nobody asks them that.
But isn't that just what AI does now? They're basically saying That