John asks:

> Sounds facinating. How about giving at a future Interex conference?

While I would be pleased to do that, I'm not nearly as sure that Interex
would want to sponsor such a talk. Not only would it be a significant
departure from their traditional conference, the subject has nothing to do
with Hewlett-Packard computers, or any other brand of computers currently in
existence. To build the machinery that I envisage requires a restart, to back
up to 1940 and to begin again, this time building an array of hundreds or
thousands of competing universal Turing machines as finite state automata,
each requiring nothing more than a memory array and a bank of flip-flops.

The architecture of these "computers" is astoundingly simple and is shown in
the diagram:

     http://aics-research.com/research/notes.html#IIH

Although it may not look like it on first viewing, this architecture could
present a behavior to the outside world that is indistinguishable from an
HP3000 running MPE, an Intel box running Linux, or a mammalian genome. If the
two memory arrays that are shown in the diagram were 500MB each, there's more
than enough room in the FSA to store the entirety of the behavioral
complexity of the human genome in the arrays.

Evolution is perceived as slow only because of the pace at which events
occur. But in this environment, especially when in the presence of a large
population of these devices, the rate of evolution could be accelerated by
perhaps a million or a billion times. The complexity of the evolved results
does not depend on the architecture of the little processors. Rather it
depends primarily on the complexity of the environment to which they were
exposed and have become adapted.

There may be an alternative however to giving a talk at HPWorld. The United
States' National Science Foundation has been partially funding a new product
development push of ours (although the bulk of the money has come from us --
and it's another product we're going to give away for free, he said, rolling
his eyes). We're in the process of taking the tutorial part of QCTerm out of
the terminal emulator and making it its own product, tentatively called
QCShow. The intent of this work is to deliver lectures and talks exactly like
the one I gave to the Computer Science department over the internet so that
they may be viewed by anyone, anywhere in the world, at their convenience,
requiring no more than 28.8kbps bandwidth.

I intend initially on putting up several of my own lectures as demonstrations
of what can be done with QCShow, with the artificial intelligence lecture(s)
being one of them. I expect that this will occur somewhere between six to
twelve months from now.

Again I want to sincerely thank Neil Harvey for letting me use his machines
in Capetown, South Africa for these last four years. I dial onto his machines
often, playing our trial lectures from his machines -- and nowadays, you
can't tell any difference between the lectures originating from South Africa
or from the machine 20 feet away from me. Because of his help, we've been
able to determine and develop robust solutions to the common problems of
low-bandwidth connections originating from the other side of the planet and I
can't say how much I appreciate that.

Wirt Atmar

* To join/leave the list, search archives, change list settings, *
* etc., please visit http://raven.utc.edu/archives/hp3000-l.html *