From: Gordon Worley (redbird@rbisland.cx)
Date: Wed Aug 01 2001 - 10:50:20 MDT
At 5:12 PM +0200 8/1/01, Joaquim Almgren Gāndara wrote:
>"Improve"? I can see how you could use GA to optimize an algorithm,
>but to make it more intelligent? (You can't possibly mean an AI tool
>that opmitizes code, since that already exists and would imply that
>the Singularity should already have happened.)
Well, we have to make something that does more than optimization, but
makes code more intelligent (see below). This is the brute force
method of reaching the Singularity that would take until 2040 as a
recent prediction stated.
Here's an analogy for you. Back a couple years ago, when Mac icons
maxed out at 256 colors per pixel and were 32 *32 pixels in size
(1024 pixels^2), this meant that there were 1024^256 (2^2560)
possible icons (I'm ignoring small icons and icons with fewer
colors). Now, suppose we had ran a program that created ever icon
possible. Eventually, we get something that looks good. Better yet,
if we have an AI tool that knows what looks good, it could
genetically generate icons that looked better and better. Given
enough time, we'd come up with super-looks-.good icons. Now, this
problem space is tiny compared to that of AI, but eventually, with
enough processing power, even if we don't have an algorithm to test
for intelligence, we'll eventually hit upon the code that works.
>How would the AI tool know whether it's stumbled upon
>something intelligent?
Not sure on this right now. We probably have to have more experience
writing intelligent code first.
-- Gordon Worley `When I use a word,' Humpty Dumpty http://www.rbisland.cx/ said, `it means just what I choose redbird@rbisland.cx it to mean--neither more nor less.' PGP: 0xBBD3B003 --Lewis Carroll
This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:00:37 MDT