[ExI] Unfrendly AI is a mistaken idea.

[Mike Dougherty]

On 5/24/07, Lee Corbin <lcorbin at rawbw.com> wrote:
> > Likewise, I think they will be just as interested in pouring their
> > consciousness back and forth between different vessels.  It will
> > probably be one of the main activities of both ourselves and these
> > superintelligences, to explore different sensoria, different powers of
> > thought, different types of consciousness.  Who wouldn't want to be a
> > cat for a day?
>
> I cannot imagine that I will at any time prefer a handicapped intelligence.
> Your idea strikes me as really weird. You want to waste a day being a
> cat?

I wouldn't want to exist ONLY as a cat, but in the interest of novel
runtime experience, I'd spawn a process to exist as a cat with the
expectation that it would merge again to increase my perspective on
various states of being.  Maybe the observation of the limited
resource/awareness state of being a cat were amusing to the main
process such that the cat were run on successive occasions - as an
object lesson involving the exchange of a large experiential context
"on being an animal" for example.

> Now remember that even anything as primitive as an orginal human
> being (uploaded, no doubt) will still have total formal control over
> his emotions. I guess you want to be a happy cat. Well, why not be
> an even happier advanced human and be able to appreciate it more?

My opinion is that "advanced human" and the AI of this discussion will
likely be so interconnected that it won't be realistic to discuss
terms like Us vs Them.  Humans currently have enough difficulty
drawing meaningful lines between groups - how much harder will it be
when our population has increased dramatically in number and
diversity?