A lot of the big existential fears about AI seem to boil down to basically “it will try to excel at tasks like maximizing profit or power even if that means people get hurt or exploited” and I keep thinking … don’t we already have that?