Pinned
I have shifted my view on AI safety from "humans should always remain in control" to "humans should give up control as soon as is safe and practical"
We should aim to build an orderly machine civilization with better institutions than humanity has and much stronger property





















