From: Dale Johnstone (DaleJohnstone@email.com)
Date: Sat Nov 18 2000 - 09:00:47 MST
----- Original Message -----
From: "Ben Goertzel" <ben@intelligenesis.net>
To: <sl4@sysopmind.com>
Sent: Wednesday, November 22, 2000 1:06 PM
Subject: Evolving minds
>
>
> > I'm currently kicking around a multi-AI scenario whereby there's
> > a community
> > of similar minds and improvements are made only to one of them.
> > If there's a
> > problem it can be dealt with by the others. Even if the modified AI has
an
> > edge, it looses by weight of numbers. Any thoughts?... (perhaps on a new
> > thread).
>
> Modifying one member of a community of AI's at a time seems an
unnecessarily
> slow way of
> doing evolution.
I wouldn't call it evolution because it's not guided by natural selection
(in the mindless sense). It's deliberate, careful, intelligent modification.
The idea being that it doesn't degrade into an uncontrollable replicator
explosion through some loophole or quirk we didn't foresee.
You'll have to forgive me for not having time to look properly into your
work, but basically I find myself in broad agreement with what I read on the
webmind philosophy page about AI & wish you luck. :)
However, if it works, what's to stop it becoming too successful & turning
all the matter of Earth into webmind nodes?
For instance if your fitness function rewards successful algorithms, what
happens if one of them comes up with the bright idea of tricking a human
into giving it access to nanotech, then proceeds to build more computer
power for itself?
I'm not one for disasterbation but that stakes are high. We can't afford to
get it wrong.
> In an evolutionary system, typically, there is a continual influx of new
> forms throughout
> the population, which one can show through mathematics or simulations is a
> more efficient
> learning strategy.
Have you compared random mutation to intelligent design?
We have aircraft that fly faster, further, higher and for longer than any
bird. It didn't take us millions of years either.
Cheers,
Dale.
This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:00:35 MDT