RE: Hardware Progress

From: Peter Voss (peter@optimal.org)
Date: Sat Jun 23 2001 - 14:32:44 MDT


Hear, Hear

Apart from Kurzweil's own figures & arguments, a good case can be made that
we already have more than enough (total) hardware today - provided we don't
try copy the brain synapse by synapse.

A number of supporting arguments are in my notes for Extro
http://www.optimal.org/peter/hyperintelligence.htm

In addition, Kurzweil's graphs tend to focus $1000 machines. A credible
design for Real AI could mobilize a lot more resources.

Hardware is *not* the problem. Good theory & software design are.

www.optimal.org - Any and all feedback welcome: peter@optimal.org

 -----Original Message-----
> Kurzweil estimates that "supercomputers will achieve one human brain
capacity by 2010, and personal computers will do so by around 2020."
>

And yet he also points out in recent presentations that it may be possible
to implement human-equivalent functionality using 1000 times less "power" in
computers. He gives an example of the auditory
subsystem.

Do you ever get the feeling there is a bit of cognitive dissonance there
with him? He seems to want to have a human-centric future, even out past
2030, yet the technology trends don't seem to point that way
when you look closely... hopefully this dissonance will lead him to a more
SL4 kind of reasoning. Or perhaps he is simply presenting a more
conservative side in public?

--
Brian Atkins


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