>not AGI here (which is quite a ways out, realistically) You don't know how far away it is. >If anything - we should be accelerating development toward the go

>not AGI here (which is quite a ways out, realistically) You don't know how far away it is. >If anything - we should be accelerating development toward the goal of AGI, as the implications for humanity are profound. Given we don't know how far away it is but current models are matching Human performance on lots of tasks and we don't know any way to ensure their safety it's entirely reasonable to be scared.