Question 1: Tell us about yourself. What is your background, and what is your current situation?
This much is known: Mentifex did much of his thinking in English,
running on at the mind with only brief branchings into German or
Russian or Latin -- ubi caritas et amor -- oftentimes slammed
(like a long-distance telephone subscriber) into an abruptly
foreign syntax and vocabulary by the mere vision or audition
of an alien presence, a Kundun as it were, such as "Kindergarten"
or "Gesundheit" und so ging es einige Minuten weiter.
Mentifex/ATM (Automatic Thinking Machine?) even designed this
lingo-switch into his Forthmind and other children (for ATM
reproduced only his mind and not his body), so that you may
slam a mentifex-class AI from one natural language into another
by suddenly moving into linguistic terra incognita, where you
will be joined by the AI made in the image and likeness of the
Mentifex made in the image and likeness of I AM WHO AM.
All else is speculation: The Bachelor of Arts degree in Greek
and Latin from the University of Washington in Seattle in 1968;
the brief start of U Cal Berkeley graduate school in classics
before a stint in the U.S. Army as a nuclear weapons electronics
specialist; the series of civilian day and night jobs financing
the life-bet on alife AI.
Current situation? Working totiexistentially to create an
artificial mind at http://sourceforge.net/projects/mind in
a context of several hundred Open Source AI projects
racing along different paths into the future.
Question 2: What is Mentifex? What is the Forthmind project?
Mentifex is a brand name or user ID that I adopted after seeing
the word "mentefact" as a play upon "artefact" in a book about
technology. I reasoned that mentifex is to artifex as mentefact
is to artefact. The resulting word "Mentifex" is Latin for
"mindmaker."
The Forthmind project is the Forth-language part of the overall
Mentifex AI project to create an artificial intelligence coded
initially in JavaScript for Web migration and in Forth for robots,
evolving towards full civil rights on a par with human beings and
towards superintelligence beyond any human IQ -- the Singularity.
Question 3: What is your opinion of Molecular Nanotechnology?
Have you done any work in this field?
Having little or no knowledge of the field, and never having
worked there, I have no opinion to offer.
Question 4: Do you think that breakthroughs in the field of
nanotechnology,
such as molecular electronics, will bring about general artificial
intelligence?
No, I suspect that the reverse will be true: Breakthroughs in AI --
such as the http://mind.sourceforge.net that I am here to talk about --
will migrate into nanotechnology.
Question 5: How would you characterize the investment climate for
emerging
technologies such as artificial intelligence and molecular
nanotechnology?
Gee, I dunno, the investment climate is infra dig to an otaku zoku like
myself. Leave it to the bean counters who concern themselves with
money.
Question 6: Is AI development stifled more from limited software or
from
underpowered hardware?
Neither the one nor the other. We don't worry about "limited" software
when our very goal is to create "unlimited" software, a
superintelligence
as predicted by Vernor Vinge in his paper on Technological Singularity
at
http://www.ugcs.caltech.edu/~phoenix/vinge/vinge-sing.html and as
discussed at
http://sysopmind.com/archive-sl4/current -- a forum of the Singularity
Institute.
As for the question of "underpowered hardware" -- computer hardware is
actually advancing far more rapidly than AI is. Once 64-bit CPU chips
become common and prevalent, then our baby AI Minds will be able to
address practically unlimited amounts of memory. Hardware is still
underpowered with respect to massively parallel processing (MPP),
but AI and MPP will probably converge just in time for each other.
Question 7: Will breakthroughs in computers and nanotechnology come
primarily from corporations or from the Academic community?
Under the Fifth Amendment to the U.S. Constitution, I decline to
testify on a subject about which I have not the faintest clue.
Question 8: What is your view of the concept of a technological
singularity?
Vernor Vinge, in my hero-worshipping opinion, is right on target. His
1993
paper on "Technological Singularity" is my all-time favorite document
on
the topic of artificial intelligence. I first read Vinge's paper when
it came out in the Whole Earth Review back around 1993, and since then
I have referred to Vinge's paper countless times on Usenet and on
webpages.
Although the whole point of Technological Singularity is that we can
not
see beyond the arrival of AI superintelligence, recently I have begun
to collect and read all available AI science fiction books which
attempt
to describe the emergence of AI. The following list of AI SF books --
resulting from a recent Usenet discussion in
news:rec.arts.sf.science --
is in alphabetical order by author, and so your readers may easily
search for these books in any book store or library.
Orson Scott Card -- Speaker for the Dead (1986)
Joseph H. Delaney, and Stiegler -- Valentina: A Soul in Sapphire
David Gerrold -- When Harlie Was One
Robert Heinlein -- The Moon Is a Harsh Mistress
Frank Herbert -- Destination: Void (1966)
James Patrick Hogan -- The Two Faces of Tomorrow (1979)
Victor W. Milan -- The Cybernetic Samurai (1985)
Rudy Rucker -- Wetware (1988)
Thomas Ryan -- The Adolescence of P1
Astro Teller -- Exegesis
Thomas T. Thomas -- ME: A Novel of Self-Discovery (1991)
Question 9: Are there any major corporations, Universities, or
organizations
working on the problem of general machine intelligence?
There are some minor universities and moribund corporations working on
general AI, but the most exciting progress is happening right now at
http://sourceforge.net in the Science/Engineering category.
Several hundred Open Source AI projects are racing into the future
at SourceForge. My own http://sourceforge.net/projects/mind began
on 18 July 2001 and I would like to tell the inside story right here.
Being only a low-tech philosopher-type with meager computer skills, I
watched enviously for more than a year as all the high-tech hotshots
launched ambitious projects on SourceForge. I did inaugurate a "PD AI"
project at http://freshmeat.net in 1999, but I was unceremoniously
booted off in June of 2000 when the freshmeat Linux maniacs discovered
that the Forthmind project had migrated from the politically neutral
Amiga platform to the untouchable Microsoft Windows environment.
When I decied to apply for an AI Mind base on SourceForge in July of
2001,
I went about it with methodical monomania. For instance, the "Software
Map"
paragraph about each project makes the all-important first impression,
so I carefully drew up a mission statement -- which I threw out and
rewrote spontaneously during the registration process.
It was not easy for the Latin and Greek classics major Arthur T. Murray
to establish a high-tech, theory-edge, brain-mind project on
SourceForge.
I had to go out and buy Karl Fogel's excellent book, "Open Source
Development with CVS" (Concurrent Versions System). Then I had to
obtain security software and CVS software to run on the Windows 98
machine which I had received free of charge from the now defunct
Free-PC.com
that gave away 40,000 computers in 1999 at the height of the dot.com
boom.
Being at the bottom of the computer science IQ barrel, where my
98th percentile test-ranking means that everybody else is smarter
than I am because they are at or above the 99th percentile, I had
to learn the ins and outs of moving files onto SourceForge by
bonehead trial and error. Every time I finally got something
right, I made sure to write down all the steps involved so that
I could repeat them.
Opportunist that I am, I grabbed the "mind" designation among the
SourceForge AI projects, but not totally without justification,
because after all I had indeed written a "Mind" program first in
Amiga ARexx (last released on 26 November 1995), then in Forth (1998)
and in JavaScript (2001). But impudence reeked from the SourceForge
Trove description of "an artificial intelligence coded
initially in JavaScript for Web migration and in Forth for robots,
evolving towards full civil rights on a par with human beings and
towards superintelligence beyond any human IQ."
These bold claims about "full civil rights" and "superintelligence"
were calculated to serve notice on all AI shops, friend or foe alike,
that Mentifex will pay any price, bear any burden, and meet any
hardship to assure the emergence and the success of the AI Mind.
Question 10: What advances do you foresee occurring in the AI field
during
the next decade?
"Wolf! Wolf! The AI wolf is coming." But this time it's true.
Question 11: Is Government funding of emerging technologies such as
AI
and
nanotechnology adequate?
No, because too much money is going to Big Physics. AI is much more
important and valuable than physics, and so we should re-channel
the enormous expenditures for physics into an AI Manhattan project.
Question 12: What are your plans for the future?
To keep working on the code of the AI mind at
http://mind.sourceforge.net and to try to
create a genuine example of classic AI software --
a simple artificial Mind so elegantly crafted
that it may serve as a starting point for
AI Minds in many different programmming languages.
