Re: Uploads and AIs (was: Deliver us from...)

From: Spike Jones (spike66@attglobal.net)
Date: Sun Apr 08 2001 - 10:31:14 MDT


> Spike Jones wrote:
> >
> >I havent a clue what kind of future is waiting for
> > us post-AI, post singularity, and post nanotech. Hell, I cannot
> > even define exactly what I mean by the term "us".

Samantha Atkins wrote:

> That is simply honest. But we don't need to see into the future but see
> how to create a future that is viable. - samantha

Thanks Samantha. Since the formation of the sl4 group, I
have been giving the post-singularity world a lot of thought,
but I don't seem to have any real insights. That future seems
hidden behind a foggy veil of chaos.

For instance, regarding an earlier comment on chess software,
we didn't really know that human-like chess was a mechanical
task until it was demonstrated. Likewise, most chess playing
programmers might now argue that writing software is a task
that can be accomplished by software in an overgrown
calculator. When that occurs, then the process goes open
loop, and what happens after that is a mystery to me.

The first example of this might be already upon us: again
in the chess world. A version of chess software has adjustable
parameters. A uber-program selects the parameters, plays
the software against itself a bunch of times, adjusts the
parameters, plays again, etc, thus discovering the most optimal
configuration. Rumor has it that the software discovered a truth
millions of humans found in 500 years of play: that knights are
better than bishops in a closed board with lots of pieces, but
as the game progresses and the board clears out a little, the
bishops are better than the knights. There is a computer-
generated algorithm now that adjusts the relative value of
bishop vs. knight, making for a slightly better program.

Coooooool. {8^D spike



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