Very hard to see how alignment is maintained with a thing that is way smarter than us and is evolving on its own.
The thing way smarter than us maintains it. If it manages to notkilleveryone initially, it’s motivated to preserve that state of affairs, as a matter of revealed preference. Alignment is not control.
The thing way smarter than us maintains it. If it manages to notkilleveryone initially, it’s motivated to preserve that state of affairs, as a matter of revealed preference. Alignment is not control.