From: Arthur T. Murray
Subject: Re: A JavaScript AI can read other Web pages.
Date:
Message-ID: <3a967a32@news.victoria.tc.ca>
Edward Hsu ·······@uclink4.berkeley.edu" wrote on Fri, 23 Feb 2001:
> "Arthur T. Murray" <·····@victoria.tc.ca> wrote in message
>> [...]
> HAHAHAHAHA
> Artificial intelligence is a joke. After 40 years of
> intense research, noone has created a system that can
> consistently fool human beings into thinking it is
> a human (the infamous Turing Test).
ATM:
["They laughed when I sat down to code."] No, seriously,
AI in the year of HAL is not a joke any longer. We are
not trying here to pass the Turing test, or even to win the
http://extrapink.com/botcontest/ among AI chatterbots.
We aim instead at the Omega Point of de Chardin, Tipler et al. via
http://www.ugcs.caltech.edu/~phoenix/vinge/vinge-sing.html (q.v.).
> [...]
> Besides, why would someone want to write an AI
> in JavaScript when LISP has so much more power anyway?
ATM:
JavaScript is turning out to be the perfect language for
communicating a basic mind-model to a universe of Netizens.
Embedded in the free, public-domain source code of the Mind are
"DIY AI" markers where do-it-yourselfers may modify such simple
aspects of the artificial intelligence as its front-end screen
presentation and the messages that pop up during a session.
Computer sales persons, booth representatives at shows,
anybody with a point to make or a product to present in public,
may use the free Artificial Mind in Transcript mode to
print out souvenir protocols of the interaction between a
human being and the primitive but watch-out! intelligence.
But mainly we expect "hackers" (in the good ESR sense) to
take the rudimentary beastie apart and port it to, say, LISP.
> Plus LISP is so good at recursive algorithms.
ATM:
Anyone who wishes to rewrite the Artificial Mind in LISP, please see
http://www.geocities.com/mentifex/lisp.html -- LISP for AI.
> All you have written is a simple parser that parses plaintext tokens.
> If you think that's cool, then you are really naive.
ATM:
The Artificial Mind simulates nerve fibers and their synapses
with elements in an array and flags on the elements, respectively.
Anyone with Internet Explorer may summon the Artificial Mind,
type in something like, "You teach me," watch the AI respond
"I teach you," and then see how the mind works by tracing out the
information-flow from concept to concept in Troubleshoot mode.
> Try looking at more sophisticated scripting languages
> like Network Query Language, WebL, Squeal,
> HEL, etc. I have links to a whole bunch of them at:
> http://thibs.menloschool.org/~hsu123/links2.htm
> You should read some classic literature like Oscar Wilde's
> The Importance of Being Earnest, or Charles Fourier's
> The Theory of the Four Movements. That will teach you to be more
> cautious about making extragavant proclamations on USENET.
ATM:
To say that here is the simplest possible Artificial Mind
for people to take apart and build upon as they please,
is not extravagant but a ho-hum Usenet phenomenon.
>
> ======================================================================
> Edward Hsu
> University of California, Berkeley
>
> Email: ······@uclink4.berkeley.edu
> Web: http://thibs.menloschool.org/~hsu123/
> [...]
It sounds like you have put a lot of work into this, how much exactly - I am
curious. Also, I have tried to use your web page and it seems thoroughly
unimpressive. I'm sure this is mostly because I don't know how to "access
it's intelligence". Please feel free to impress me with something it does.
Also, have you entered into that chatterbox competition thing? I looked
over some of the robots and they all look pretty poor (even the best ranked
ones - Alice, etc.). I have also played with Billy for a while and it is
very poor. Do you know where I can get access to some source or psudo code
for some of these bots, I'm curious to see where they make all their errors
when speaking - also to see how some of them work.
Thanks
"Arthur T. Murray" <·····@victoria.tc.ca> wrote in message
·············@news.victoria.tc.ca...
> Edward Hsu ·······@uclink4.berkeley.edu" wrote on Fri, 23 Feb 2001:
>
> > "Arthur T. Murray" <·····@victoria.tc.ca> wrote in message
> >> [...]
> > HAHAHAHAHA
> > Artificial intelligence is a joke. After 40 years of
> > intense research, noone has created a system that can
> > consistently fool human beings into thinking it is
> > a human (the infamous Turing Test).
> ATM:
> ["They laughed when I sat down to code."] No, seriously,
> AI in the year of HAL is not a joke any longer. We are
> not trying here to pass the Turing test, or even to win the
> http://extrapink.com/botcontest/ among AI chatterbots.
>
> We aim instead at the Omega Point of de Chardin, Tipler et al. via
> http://www.ugcs.caltech.edu/~phoenix/vinge/vinge-sing.html (q.v.).
>
> > [...]
> > Besides, why would someone want to write an AI
> > in JavaScript when LISP has so much more power anyway?
> ATM:
> JavaScript is turning out to be the perfect language for
> communicating a basic mind-model to a universe of Netizens.
>
> Embedded in the free, public-domain source code of the Mind are
> "DIY AI" markers where do-it-yourselfers may modify such simple
> aspects of the artificial intelligence as its front-end screen
> presentation and the messages that pop up during a session.
>
> Computer sales persons, booth representatives at shows,
> anybody with a point to make or a product to present in public,
> may use the free Artificial Mind in Transcript mode to
> print out souvenir protocols of the interaction between a
> human being and the primitive but watch-out! intelligence.
>
> But mainly we expect "hackers" (in the good ESR sense) to
> take the rudimentary beastie apart and port it to, say, LISP.
> > Plus LISP is so good at recursive algorithms.
> ATM:
> Anyone who wishes to rewrite the Artificial Mind in LISP, please see
> http://www.geocities.com/mentifex/lisp.html -- LISP for AI.
>
> > All you have written is a simple parser that parses plaintext tokens.
> > If you think that's cool, then you are really naive.
> ATM:
> The Artificial Mind simulates nerve fibers and their synapses
> with elements in an array and flags on the elements, respectively.
> Anyone with Internet Explorer may summon the Artificial Mind,
> type in something like, "You teach me," watch the AI respond
> "I teach you," and then see how the mind works by tracing out the
> information-flow from concept to concept in Troubleshoot mode.
>
> > Try looking at more sophisticated scripting languages
> > like Network Query Language, WebL, Squeal,
> > HEL, etc. I have links to a whole bunch of them at:
> > http://thibs.menloschool.org/~hsu123/links2.htm
> > You should read some classic literature like Oscar Wilde's
> > The Importance of Being Earnest, or Charles Fourier's
> > The Theory of the Four Movements. That will teach you to be more
> > cautious about making extragavant proclamations on USENET.
> ATM:
> To say that here is the simplest possible Artificial Mind
> for people to take apart and build upon as they please,
> is not extravagant but a ho-hum Usenet phenomenon.
>
> >
> > ======================================================================
> > Edward Hsu
> > University of California, Berkeley
> >
> > Email: ······@uclink4.berkeley.edu
> > Web: http://thibs.menloschool.org/~hsu123/
> > [...]