From: Richard Loosemore (firstname.lastname@example.org)
Date: Sat Apr 29 2006 - 14:23:32 MDT
Damien Broderick wrote:
> At 12:17 PM 4/29/2006 -0400, Richard wrote:
>> Or, looking at it a different way, if someone had managed to complete
>> a superintelligent AGI on, say, January 1st 1670, just after Newton
>> discovered the laws of motion, ALL of the scientific progress that has
>> happened between then and now would have been finished by January 2nd
>> (assuming that we conservatively restrict the self-improvement
>> potential to a factor one billion).
> This ignores the (apparently) crucial element of empirical testing and
> feedback into pure ratiocination. I bracket "apparently" because it's
> not yet demonstrable that several million years of armchair thinking
> can't project superstrings and quasars from the 17th century empirical
> database, but that restriction seems to me highly likely.
> Damien Broderick
Your point is a good one.
However, I was implictly assuming not so much pure ratiocination as new,
much faster empirical techniques: the AGI, once created, figures out
how to do "microscale experiments" to answer all the questions it needs
to answer, where these microscale experiments are all designed to get
answers back in very short short order.
There will be some questions that it will find difficult to do quickly,
and heat generation would be a problem, so in practice I think it might
have taken the rest of January... ;-)
This archive was generated by hypermail 2.1.5 : Sat May 18 2013 - 04:01:01 MDT