Re: Moravec's estimates?

From: James Rogers (jamesr@best.com)
Date: Mon Apr 09 2001 - 14:38:04 MDT


At 10:11 PM 4/7/2001 -0400, Brian Atkins wrote:

>Well the Playstation 2 did run into supercomputer export law restrictions,
>and is rumored to be used in Iraqi missle guidance systems...

Highly improbable, with respect to missile guidance systems. The most
advanced U.S. guidance systems today run on processors that would have been
considered decrepit five years ago. With very few exceptions most military
applications, advanced or not, require less than 50 DSP MIPS of real-time
horsepower. Even for applications like THAAD, which push the envelope of
real-time discrimination and intercept precision, you will find that the
military is using MIPS R3000 embedded cores and similar. Definitely not
anything to write home about.

>Actually if those teraflop CPUs from Sony/IBM pan out in 2006, it should
>be possible to build a "Blue Gene 2" that would have around 30 petaflops.
>(a Blue Gene CPU is only 32gflops)

Unfortunately, memory bandwidth has not followed this trend. These types
of processors are only really good for doing inordinate quantities of
computation on a relatively small sets of data.

>At this stage everyone has their
>own opinion. It won't be until we get near having a real working AI that
>we will be able to know for sure what it needs. I will be interested to
>see how intelligence scales with how much MIPS/memory is available.

The effect of CPU speed isn't terribly important in terms of pure
intelligence. Running at half-speed, you'll get half the number of results
in a given period of time but the intelligence of the results doesn't
change. OTOH, memory appears to have a scale-invariant logarithmic
relationship to any reasonable AI implementation, and is therefore arguably
a more critical resource limitation.

-James Rogers
  jamesr@best.com



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