From: Philip Goetz (email@example.com)
Date: Wed Feb 22 2006 - 08:49:37 MST
David Alvarez wrote:
>I'd agree with you except for the qualifier "sometimes". I'd say knowledge
>is most often good, and when it conflicts with happiness it is still often
>preferable. I didn't choose that example as proof, just as illustration.
I think Mike's point was that knowledge will not ALWAYS make one
happier, and thus, an AI that was programmed to be Friendly by telling
everyone things that
- they were unaware of
- were important to them, and
- were not the confidential information of other people
could cause quite a ruckus, contrary to intent. Sounds like the basis
of a 1950s science fiction short story.
This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:00:55 MDT