From: Tommy McCabe (firstname.lastname@example.org)
Date: Sat Jan 03 2004 - 10:00:24 MST
--- Eugen Leitl <email@example.com> wrote:
> On Sat, Jan 03, 2004 at 05:26:17AM -0800, Tommy
> McCabe wrote:
> > Possibly true, but an AI would be smart enough to
> > recursive self-improvement, or intelligent
> > design-and-test, in place of blind mutation, for
> > improvement component.
> How do you know this? The only known instance of
> intelligence we know appears
> to use evolutionary selection as its driving force.
The only known existence of intelligence we know of,
us, only uses Darwinian evolution in cases like
programming, where we have no adaptations for it and
thus are stuck using abstract thought, and only then
because a computer can do ten billion dumb mutations
inthe time we can do ten intelligent improvements.
> I don't know what
> "recursive self-improvement" or "intelligent
> design-and-test" is, and either,
> I suspect, do you.
Intelligent design-and-test is what humans do- the
'mutations', instead of being random genetic bitflips,
are planned by some entity with general intelligence
(unlike evolution, which doesn't have general
intelligence). Recursive self-improvement, is where
you improve yourself, and the improved self has
capabilities improved enough to improve yourself
further, et cetera ad nauseam.
> As to "blind mutation", evolution is anything but
> blind -- or you wouldn't be
> there to write this.
You're confusing my words here. Natural mutation is
blind; natural selection isn't. I am not a creationist.
Do you Yahoo!?
Find out what made the Top Yahoo! Searches of 2003
This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:00:43 MDT