From: James Higgins (jameshiggins@earthlink.net)
Date: Wed Jul 03 2002 - 20:12:48 MDT
At 09:54 PM 7/2/2002 -0700, you wrote:
>The basic issue is that nothing is safe. Once the AI is significantly
>smart, we can't count on any system to contain the seed AI. There is no
>such thing as "safe information retrieval from the Internet", just *safer*
>information retrieval.
>
>If the goal of this thread was not to explore ways to contain a seed AI,
>then I apologize and would appreciate a clarification statement.
The goal was to explore ways to contain a seed AI, but not a trans-human
AI. Once the AI either starts getting beyond human level intelligence or
can run at extreme speed (like being able to process an OC-12 in real time)
then your pretty much screwed if you still need to contain it. So the (my)
intent was to discuss what methods to use initially when starting with a
baby / infra-human AI.
James
This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:00:40 MDT