From: Eugene Leitl (Eugene.Leitl@lrz.uni-muenchen.de)
Date: Sun Apr 20 1997 - 23:44:13 MDT
---------- Forwarded message ----------
Date: Sun, 20 Apr 1997 22:55:41 -0400 (EDT)
From: Tom Easton <teaston@acadia.net>
To: psychoceramics@zikzak.net
Subject: psychoceramics: Moravec's Mind Age
Nikolai et al.:
A couple of years ago, Moravec sent me (as the Analog book columnist) the
Mind Age manuscript for timely review. I read it, wrote the review, and
when Bantam canceled the project shelved the review. Now that you're saying
the book is about to be published elsewhere, I can expect to use the review
after all. Meanwhile, FYI, here's what I had to say:
In his 1988 Mind Children (Harvard University Press), Hans Moravec
sketched the future development of robots, proposed that within about 50
years it would be possible to transfer human minds into computers, and
speculated on the replacement of biological intelligence by machine
intelligence. To him, robots would be the creations of our minds just as
our children are the creations of our bodies, and since it is our minds that
make us most distinctively human, our "mind children" would be our children
in the more important sense. At the same time, mind transfer would offer us
true immortality.
In Mind Age: Transcendence Through Robots, Moravec tells us that
"Among many nice reviews of [the earlier book] an angry few brandished words
like 'horrific,' 'nightmare' and 'immoral,' and at least one was too irate
to publish." I find that curious, for to a mind steeped in science and
science fiction since adolescence, Moravec's ideas are not terribly strange
even if they do push the envelope of what even a visionary scientist might say.
Yet Mind Children was very much a part of a discussion of the
potentials of artificial intelligence and robotics that began almost as soon
as the first digital computers began chewing their bytes. Norbert Wiener
invented cybernetics to "study brainlike activity in animals and machines."
And in 1950, Alan Turing proposed that if one could not tell the
intellectual difference between a computer and a human conversationalist,
there was no such difference. The computer would be intelligent in the only
way that matters.
Since then a great deal of effort has gone into meeting the demands
of the "Turing test." So far success has been only partial, but there has
been enough success to justify immense optimism in the minds of some, such
as Moravec. On the other hand, others still resist the notion that a
machine, a mere thing, could ever possibly really think. After all, it's too
deterministic, quantum uncertainty is obviously essential to true thought,
and it
couldn't possibly have a soul, a mind, emotions, a sense of beauty, a...
The list of deficits is
endless; it also tends to be revised every time clever programmers figure
out how to give a
computer a new capability.
To some extent, Mind Age: Transcendence Through Robots is a
recapitulation and updating of the earlier book, reviewing the development
of robots up to the point where they can take control of their own
continuing improvement or evolution. Moravec devotes considerable space to
what a civilization dominated by autonomous machines must
mean for human economies and does not shrink from the thought that most
humans will be
content to eat the lotus in idle bliss, the few adventurous souls will move
into robotic bodies or cyberspace and go forth into the universe, and
eventually even the robotic superminds will be made obsolete by what can
only be described as an intelligent universe, a universe converted to
computational apparatus and sheer mind in which all reality is virtual.
Science fiction readers, who have seen many of these ideas before,
find them enchanting. Others find them terrifying, which explains some of
the reactions Moravec receives, even from such scientific luminaries as
Roger Penrose, who has spent a good deal of effort arguing against the idea
of machine intelligence. Moravec is too polite to call Penrose out in
person, but in his last chapter, he does borrow the form of Turing's seminal
1950 paper to attempt--like Turing--to lay to rest the various objections to
the notion of machine intelligence. These objections he phrases like so:
machines have no souls, so cannot think; thinking machines cannot be
possible, because the consequences would be too dreadful; mechanical
reasoning is too limited; machines can have no inner experience to give
meaning to their "thoughts" or actions; machines will never be kind, moral,
joyous, perceptive, original, etc.; computers do only what we can program
them to do; computers don't have nerve cells; it is not possible to specify
for a machine what to do in every possible circumstance a human can
encounter; computers can't have ESP (Moravec considers this a genuine
nonissue but deals with it because Turing did).
Moravec is not always convincing, but I do not think it is possible
to be more so. The immense leaps in the development of robots and
artificial intelligence that he discusses cannot be predicted confidently
given the present infantile state of the art. However, it is just as
unreasonable for the opposition to say those leaps cannot be made. Both
sides display an almost religious adherence to their beliefs. Certainly
Moravec has faith in his dreams. His
opponents have faith in tradition.
I'm inclined to side with Moravec, and I recommend Mind Age as a
stimulating, provocative treat for your own mind.
Sincerely,
Tom Easton
http://www.thomas.edu/easton/profid.htm
http://www.sff.net/people/teaston/
This archive was generated by hypermail 2.1.5 : Fri Nov 01 2002 - 14:44:22 MST