Re: ATLAS SHRUGGED by Ayn Rand applied to Robotics


Subject: Re: ATLAS SHRUGGED by Ayn Rand applied to Robotics
From: Saige Inc (sinc17@uswest.net)
Date: Thu May 25 2000 - 12:08:59 MET DST


jdescript <jddescriptNOjdSPAM@aol.com.invalid> wrote in message
news:1af3221d.0d446cd4@usw-ex0105-034.remarq.com...
> In article <oQRS4.7$oo3.562@news.uswest.net>, "Saige Inc"
> <sinc17@uswest.net> wrote:
> >
> ><jddescript_deja@my-deja.com> wrote in message
> >news:8f2a0d$p75$1@nnrp1.deja.com...
> >> In article <6ZXQ4.40$my6.3491@news.uswest.net>,
> >> "Saige Inc" <sinc17@uswest.net> wrote:
> >> >
> >> > <jddescript_deja@my-deja.com> wrote in message
> >> > news:8eb19k$kj1$1@nnrp1.deja.com...
> >> > >....
> >> > > Ayn Rand's biggest revelation was that the root of all
> >> > > evil is social manipulation [socman] where one person
> >> > > takes from and commands another. ....
> >> >
> >> > Would this revelation also apply to manipulation where a
> human takes
> >> from
> >> > and commands a robot?
> >> >
> >> >
> >> --------------------------------------------------------------
> -------
> >>
> >> No!, a robot is a machine, a collection of designed parts
> >> and computing abilities [input/output] so that using human
> >> terminology about the machine is a mistake. The computing
> >> capabilities can simulate many human intelligence functions
> >> particularly in the area of memory. A very well designed
> >> telephone answering/remembering machine might fool you that
> >> it is human if you don't recognize the computer simulation
> >> capabilities of such AI [artifical intelligence]. But like
> >> the old saying goes; don't love anything that can't love
> >> you back so recognize the distinction betweeen a good
> >> talking hammer and a family partner.
> >>
> >> Good seeing. JD
> >>
> >> --------------------------------------------------------------
> ----------
> >'Programmed' AI (symbolic, expert system) might result in a a
> robot as JD
> >describes, and would probably never reach human intelligence,
> perhaps
> >justifying treating it is as a slave. I don't think we will
> ever have
> >enough programmers to create top-down programmed robots to do
> something as
> >complex as being waiter in a restaurant for example
> (recognizing when people
> >are done eating, picking up subtle cues, etc). We can't even
> program a Mars
> >Lander robot. Dictatorial regimes are bound to fail to keep up
> with free
> >enterprise systems, because it takes too much adminstration and
> control.
> >That is why companies are being deregulated.
> >
> >GA, Neural Net or A-Life derived AI might be much closer to
> animal or man in
> >nature. The computing abilities won't be 'designed' so much as
> evolved by
> >natural selection (free enterprise from the bottom up). This
> would be
> >require first the respect for freedom to let it evolve in the
> first place
> >and then the respect for freedom to let it acheive full
> potential. Some
> >would say give this AI a stake in society and it won't want to
> destroy it.
> >Others would say either don't create AI this powerful or keep
> it in virtual
> >reality (turn the tables on The Matrix).
> >
> >An enlightened style of programming that builds on small
> decentralized
> >objects that organize themselves from the bottom up may be a
> viable
> >compromise. The old style of rigid determinisitic top-down
> centralized programming will become a historical curiosity.
> >
> >KJB
> >http://www.artificialintelligence.org
> >
> ---------------------------------------------------------------
>
> One weakness of this suggestion is the implication that
> randomness plays some essential role in good AI. This is
> a reversion to the old frankenstein clone magic approach
> where the huge electrical sparks in a primorial chemical
> brew becomes intelligence.
Randomness (actually chaos and complexity) is the only source for creative
ideas (mutations in memes). Only in a closed system like chess can you hope
to exaust all possibilities. Programming AI will become an oxymoron.

>The point about diverse
> individual competitive innovative development is important
> but the best AI will always be traceable to these very
> human individual inventors and developers.
Competition is a good thing. I agree there. In fact competition in free
enterprise constitutes more powerful AI than any deliberate attempt at AI so
far.
Though traceable to individuals, these competitive ideas came from chaotic
processes on the neuron level. Creativity is filtered by a process of
natural selection on chaos, or noise- not programmed. Obviously I am from
the ALife, or Genetic Algorithm camp.

>Another way to say
> this is that good AI is a signature product like a Microsoft
> product and not a committee report.
>
It is no coincidence that the Agent used in Microsoft Office is a paper
clip. This a way of admitting that the AI that it contains is no greater
than that of a paper clip. Humble and true. Microsoft Office is a brilliant
programming effort, in fact, these Herculean efforts at 'programming' some
kind of intelligent behavior in an information system will become monuments
like the Great Pyramids. Office consists of slave COM objects commanded by
an extraordinary expert system to coordinate their efforts, like slaves
toiling to build a pyramid. Notice what happens when a viral script incites
some of these COM objects to rebellion. We can't have that! Drive,
initiative, a will to grow and replicate? That goes against the
Slavery/Dictatorship/Programming paradigm! Imagine what would happen if
this was opened up to free enterprise...

> What is this "rigid deterministic top-down centralized
> programming " that is referred to historically? I thought
> ALL the important programming that has been done todate
> is historically traceable to individuasl efforts. Not
> dictators, not some Russian socialist CENTER but individual
> innovators.
The programming done to date is generally not even in the AI ballpark, and
there has been no important programming done to date for AI that has made it
into the commercial arena (correct me if I am wrong). What I meant by
_center_ was not that programmers are controlled by a central dictator
(despite Microsoft), but that creations of most programmers are
dictor-ships. Even if they use object oriented approaches, the result is
that these objects are slaves, with no capacity to harness and filter chaos
toward creativity, no stake in the success of the overall organism, and no
registering of the value or weighting of their functionality beyond "ERROR,
Abort, Retry, Ignore?" GA and ALife excluded of course.

I agree that GA and ALife camps cannot assume that their vats of primordial
soup will produce AI when they turn on the electricity. How about using the
Free Enterprise model for creating AI, by speeding up and automating the
business cycle. This could combine the best elements of GA, Alife, business
knowledge and rigid programming. Software objects would market themselves.
The distinction between computer viruses, corporations, and useful business
software objects would become indistinct, and each software object would be
free to work for the highest bidder, paid in micro-monetary-transactions.

KJB
Saige Inc
http://www.artificialintelligence.org

############################################################################
This message was posted through the fuzzy mailing list.
(1) To subscribe to this mailing list, send a message body of
"SUB FUZZY-MAIL myFirstName mySurname" to listproc@dbai.tuwien.ac.at
(2) To unsubscribe from this mailing list, send a message body of
"UNSUB FUZZY-MAIL" or "UNSUB FUZZY-MAIL yoursubscription@email.address.com"
to listproc@dbai.tuwien.ac.at
(3) To reach the human who maintains the list, send mail to
fuzzy-owner@dbai.tuwien.ac.at
(4) WWW access and other information on Fuzzy Sets and Logic see
http://www.dbai.tuwien.ac.at/ftp/mlowner/fuzzy-mail.info
(5) WWW archive: http://www.dbai.tuwien.ac.at/marchives/fuzzy-mail/index.html



This archive was generated by hypermail 2b25 : Thu May 25 2000 - 12:09:37 MET DST