Re: Effective(?) AI Jail

From: Samantha Atkins (samantha@objectent.com)
Date: Sun Jun 17 2001 - 01:59:34 MDT


Gordon Worley wrote:
>
> At 10:18 AM -0400 6/13/01, Gordon Worley wrote:
> >If our FAI can't handle an UFAI when ve runs into one.
>
> Oops, left something out. If our FAI can't handle an UFAI when ve
> runs into one, we're toast, so trying to teach ver on some weakly
> UFAI we might be able to program and then kill off wouldn't amount to
> much. The time could have better been spent.

Start a battle between AIs of this capacity in the immediate
neighborhood and we are very likely to be toast. There is an
excluded middle assumption above though. An un-Friendly AI
might simply be not-Friendly as in relatively indifferent unless
provoked.

- samantha



This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:00:36 MDT