Yes, by default self-improving AI goes very poorly, but this is a plausible case where would could have aligned AGI, if not ASI.
Yes, by default self-improving AI goes very poorly, but this is a plausible case where would could have aligned AGI, if not ASI.