i tend to dislike such systems as well, but a correctly aligned superintelligence would surely be trustable with anything of the sort. if anything, it would at least know about the ways it could fail at this, and tell us about what it knows of those possibilities.
thanks for the egan story, it was pretty good!
i tend to dislike such systems as well, but a correctly aligned superintelligence would surely be trustable with anything of the sort. if anything, it would at least know about the ways it could fail at this, and tell us about what it knows of those possibilities.