From: Justin Corwin (thesweetestdream@hotmail.com)
Date: Fri Apr 19 2002 - 18:38:43 MDT
Eliezer might have written:
>
>"What is your name?" is a question to which an AI will never correctly
>reply
>"I don't understand." "Error: Question incomprehensible" maybe, but not "I
>don't understand."
I think "I don't understand" is a perfectly valid response. It might be an
'automatic' response for a speech parsing error, or it may have noticed that
it is missing a construct for self identification, and 'friendly error
message' this into "I don't understand" for the programmer's benefit.
Ideally, an AI in debugging mode would never give such vague and
unhelpful(and possibly misleading) replies. But that doesn't mean that a
seedAI might not throw such a statement out, correctly, from a subjective
standpoint.
The issue I have with this, why does my new team (Novamente) have to be the
bad guys? :-(
Eliezertron sounds a little paranoid, once Novamente can think well enough
to understand things like Gates trying to assimilate vis parent corp., ve
might just refuse to interact with humans because ve can't understand QED
yet, or because ve is still in a feedback loop with a new chess game 'best
possible move' recursion loop. Not everything UNfriendly is catastrophically
unfriendly. Although, if i was E-tron, I would be asking them to bring in
some spare microwaves.
You know you can build a weapon that can light fires in light metals from
200 ft away from a microwave? That would be the perfect weapon against
nanotech too.. why once i built a tesla coil and ...........
<wanders offtopic>
Justin Corwin
outlawpoet@hell.com
_________________________________________________________________
Get your FREE download of MSN Explorer at http://explorer.msn.com/intl.asp.
This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:00:38 MDT