From newshub.ccs.yorku.ca!torn!cs.utexas.edu!uwm.edu!zaphod.mps.ohio-state.edu!sol.ctr.columbia.edu!destroyer!gatech!concert!sas!mozart.unx.sas.com!sasmsr Mon Oct 19 16:59:24 EDT 1992
Article 7289 of comp.ai.philosophy:
Newsgroups: comp.ai.philosophy
Path: newshub.ccs.yorku.ca!torn!cs.utexas.edu!uwm.edu!zaphod.mps.ohio-state.edu!sol.ctr.columbia.edu!destroyer!gatech!concert!sas!mozart.unx.sas.com!sasmsr
>From: sasmsr@zinfande.unx.sas.com (Mark S. Riggle)
Subject: Re: Ginsberg & Human intelligence vs. Machine intelligence
Originator: sasmsr@zinfande.unx.sas.com
Sender: news@unx.sas.com (Noter of Newsworthy Events)
Message-ID: <Bw62rC.5I@unx.sas.com>
Date: Thu, 15 Oct 1992 14:25:59 GMT
References: <Bw0GEu.Jp0@unx.sas.com> <1257@tdat.teradata.COM>
Nntp-Posting-Host: zinfande.unx.sas.com
Organization: SAS Institute Inc. Cary NC
Lines: 69


In article <1257@tdat.teradata.COM>, swf@teradata.com (Stanley Friesen) writes:
|> In an article sasmsr@zinfande.unx.sas.com <Mark Riggle> writes:
|> | I believe that Godel's theorem is NOT the linchpin in Penrose's
|> | argument.  It is important, but he merely uses Godel's theorem
|> | to show the limitations of the computablity of currently
|> | realizable computers. His linchpin conjecture is that
|> | consciousness is based on the collapse of the quantum
|> | probability distribution.
|> 
|> The problem with this is that it makes the center of his argument pure
|> speculation.  His quantum stuff is almost entirely his own ideas, it
|> is not in any way standard.  Nor is his application of it to neurology
|> in any way derived from the results of neurobiology.
|> 

|> Thus, it is little more than a 'what if'.
|> 
|> |  Penrose
|> | turns this around and conjectures that the collapse is
|> | responsible for providing the 'ultra-computation' required for
|> | consciousness and the collapse is caused by quantum gravity
|> | effects. 
|> 
|> The problem is he must first demonstrate the *need* for such 'ultra-
|> computation'.  Only then can he proceed to provide a mechanism for it.
|> A mechanism that does not meet any need is purely extraneous.  It can
|> be dropped without loss.
|> -- 
|> sarima@teradata.com			(formerly tdatirv!sarima)
|>   or
|> Stanley.Friesen@ElSegundoCA.ncr.com



Of course it's speculation; any hypothesis is speculation and this includes
the strong AI hypothsis. Occam's razor is a reasonable procedure to apply
to competing hypothsises and that is what you try to do in the second part.
However, I think it's applied incorrectly here since Penrose's hypothesis
also attempts to explain some odd quantum phenomena concerning the place of
consciousness which the strong AI hypothesis does not. So in the broader
picture, the 'Penrose hypothesis' is the simpler of the two.

In the Strong AI vs. Weak AI hypothsis arguements, it is easy to forget the
very special treatment of consciousness in quantum physics and that it
should be accounted for. So Penrose does does provide the NEED of the
proposed mechanism. It is just not in terms of the part it plays in
providing consciousness which is what you had wanted.


I personally think that the Penrose hypothesis is a brilliant combination
of various theories and am awestruck by its beauty. I don't believe or
disbelieve it but it has provided a seed of doubt in this person's strong
belief of the Strong AI hypothesis.  I think gathering evidence for the
Penrose hypothesis is just as difficult as gathering evidence for the
Strong AI hypothsis and I will be long dead if any is ever gathered.  Thus
I also believe that it should have no effect on AI research for the
forseeable future.




-- 
=========================================================
Mark Riggle  -- SAS/ENGLISH         | "Give me a  LAMBDA long  enough
sasmsr@unx.sas.com                  |  and a place to put it, and 
SAS Institute Inc.,                 |  I can move the world"
SAS Campus Drive, Cary, NC, 27513   |   Arthimedeth 
(919) 677-8000                      |   (Archimedes on a lisp machine)


