Re: What best evidence for fast AI?.

From: John K Clark (
Date: Mon Nov 12 2007 - 14:33:27 MST

On Sat, 10 Nov 2007 14:48:05 -0800, "Eliezer S. Yudkowsky" Wrote:

> I would begin by asking if there was ever, in the whole
> history of technology, a single case where someone *first*
> duplicated a desirable effect by emulating biology at a
> lower level of organization, without understanding the
> principles of that effect's production from that low
> level of organization.

I know this is not exactly what you were asking for, but itís close.
Just yesterday it was announced that for first time photosynthesis has
been simulated on a computer.

They then used a genetic algorithm to randomly change the thousands of
different variables that effect efficiency and than ran it through the
computer to see if it improved things. After running on a supercomputer
for a while they came up with a version of photosynthesis that was
nearly twice as efficient as the version nature came up with. The point
is the scientistís high level goal was clear, improve the efficiency of
photosynthesis, and they reached their goal without understanding what
all those low level changes they were making were doing.

Something similar happened in the area of antenna design.

The antenna the genetic algorithm came up with looked strange, itís not
something a human would likely dream up, but it worked very well, and
nobody knows why. A genetic algorithm has also been used for the
software interface between nerves and the fingers of a prosthetic arm
and in cochlear implants.

The unique thing about genetic algorithms is that this software improves
at least as fast as Moore's Law.

John K Clark

  John K Clark
-- - One of many happy users:

This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:01:00 MDT