Newsgroups: comp.ai
Path: cantaloupe.srv.cs.cmu.edu!bb3.andrew.cmu.edu!newsfeed.pitt.edu!gatech!news.mathworks.com!nntp.primenet.com!EU.net!Norway.EU.net!nntp.uio.no!news2.interlog.com!news1.io.org!torfree!bv765
From: bv765@torfree.net (Jayaram Ramanathan)
Subject: Re: Future of AI (was Re: Al P-S: Amateur Futurologist?)
Message-ID: <DwJwMG.JtF.0.queen@torfree.net>
Organization: Toronto Free-Net
X-Newsreader: TIN [version 1.2 PL2]
References: <fmdavis-2205961059040001@47.8.2.174> <4t9n17$nb6@reader1.reader.news.ozemail.net> <31fa36d3.1805508@news.c <32041b95.13300418@203.12.22.10> <4uciuv$52g@laplace.ee.latrobe.edu.au>
Date: Thu, 22 Aug 1996 17:59:03 GMT
Lines: 44

Kym Horsell (khorsell@ee.latrobe.edu.au) wrote:

<blah-blah deleted>
: ...
: Various acts of "creation" from computer constructs could be
: cited -- the list is rather a long and incomplete one and documented 
: various (other) places. 
:
: But I'd like to (badly) cite at least the "original" congruency proof 
: in a particular early geometry theorem prover. Due to
: what turned out to be a programming oversite (or -- perhaps better put --
: the programmer having the FORESIGHT to leave a degree of freedom where
: someone lesser may have imposed an arbitrary restriction) the system
: found that in a given case it was "legal" to essentially take the
: triangle and turn it over. It was then possible to show that the relevant
: 2 sides were equal and the included angle was the same and, hence,
: congruency.

: Apparently the previous 2000 years of geometry proofs had not admitted
: of this possibility. But equally obviously -- it was a valid method.

What you cite above is somewhat analogous to the classic example of five
(or is it three? I forget) monkeys typing away for donkeys' years and 
producing Shakespeare's works.  Of course, on the way they would produce
reams of rubbish also.  Do we take it from this that the monkeys know
diddly about Macbeth or Othello or Portia?  Or about jealousy, greed,
love, fear?

To put it another way, the proof the geometry program above came up with
could have been totally wrong also.  It just so happened to be right,
that's all.  No further insight can be gained.  (Except, of course, on
the kinds of constraints one should maybe relax in similar programs.)

Drew McDermott had it right on the money in the title of his classic
article "Artifical Intelligence meets Natural Stupidity", although
maybe not in the sense he intended in his article.

-jayaram

-- 
---
Jayaram Ramanathan		I  "There is no love in your heart
Toronto, Ontario		I   unless there is fire in your soul."
bv765@torfree.net		I          --  Mahakavi Subramaniya Bharathi
