From: Thomas McCabe (pphysics141@gmail.com)
Date: Tue Nov 20 2007 - 15:33:48 MST
> > what are you actually trying to say?
>
> Both you and Mr. AI will have one goal in common and it's the most
> important one of all, both of you will want to continue to exist. If Mr.
> AI didn't have this goal he wouldn't be of any use to himself or to
> anyone else.
This is not always true; see http://www.acceleratingfuture.com/tom/?p=68.
> >"The AI will do it the same way you do" is not
> > a useful way of thinking about how to build an AI.
>
> I never said it was, I was asked how the AI tells a good goal from a bad
> goal and I said it will use the same sort of mental process that we do,
> and that process is called "Judgment", and I see no reason that sort of
> data crunching can only be produced by meat. I mean, what other answer
> could I give, did you expect me to give detailed blueprints for an AI?
Calling it "judgment" invites anthropomorphicism. Utility functions
are much cleaner mathematically anyway.
>
> > And what's up with "his"?
>
> What pronoun would you prefer? I hope you're not one of those people who
> insist I say "he or she" all the time.
I've long since given up attempting to keep pronouns straight. This
has been a problem for hundreds of years and it's not likely to go
away anytime soon. Use singular they, or ve/ver/vis.
>
> John K Clark
>
>
> --
> John K Clark
> johnkclark@fastmail.fm
>
> --
> http://www.fastmail.fm - A fast, anti-spam email service.
>
>
>
- Tom
This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:01:00 MDT