Re: Singularity and AIs (was: We're stuck with each other)

From: Eugene Leitl (Eugene.Leitl@lrz.uni-muenchen.de)
Date: Mon Jan 28 2002 - 05:40:36 MST


On Mon, 28 Jan 2002, Samantha Atkins wrote:

> Nope. But has there been much work on doing general-purpose

I know you know what a CA is, Samantha. I was addressing the strange being
known as the silicon love doll.

> computation on CA? What kind of languages, methodologies, algorithms

I haven't been following the field very closely, but there's material to
be found on http://www.stephenwolfram.com/publications/articles/
http://www.stephenwolfram.com/publications/books/ca-reprint/contents.html
Fredkin has also become active recently: http://digitalphilosophy.org/
Tommaso Toffoli is another from the triad worth reading (he's probably
switched chairs recently):

http://www.google.de/search?q=cache:nxVbsrdncFMC:pm1.bu.edu/~tt/publ.html+tommaso+toffoli&hl=en

Of course, I'm still hoping that Wolfram is going to deliver
http://wolframscience.com/ within the next few months.

> and so on most take advantage of such ultimate computer hardware?

You could design a rule to simulate signal propagation along wires and
define gates, then compiling your code into a 3d circuit by same means as
silicon compilers do, but it's hardly efficient. An argument can be made
that the ideal rule should be Edge of Chaos one, and that the pattern
should make no distinction between code and data. I.e., it would be
something which you couldn't design by hand.

> Shouldn't some small-scale virtual models be build and various
> software tools designed and evaluated on them today?

Yes, they should, but not many people are doing that, sadly.

> Please point me to any relevant literature on the software side you
> may have handy.

I think approaching the digitalphysics folks should give you plenty of
refs, if a session with Google is insufficient.



This archive was generated by hypermail 2.1.5 : Fri Nov 01 2002 - 13:37:36 MST