Newsgroups: comp.ai
Path: cantaloupe.srv.cs.cmu.edu!bb3.andrew.cmu.edu!newsfeed.pitt.edu!godot.cc.duq.edu!newsgate.duke.edu!news.mathworks.com!uunet!in2.uu.net!netnews.worldnet.att.net!cbgw2.att.com!nntphub.cb.att.com!bigtop!news
From: Carlos Casize <rune@dr.att.com>
Subject: Re: artificial intellige
Content-Type: text/plain; charset=us-ascii
Message-ID: <31A637DA.7983@dr.att.com>
Sender: news@bigtop.dr.att.com (Netnews Administration Login)
Nntp-Posting-Host: walnut
Content-Transfer-Encoding: 7bit
Organization: RMACS
References: <8C0D385.01CA00376E.uuout@almac.co.uk> <4nls29$mhn@newsbf02.news.aol.com>
Mime-Version: 1.0
Date: Fri, 24 May 1996 22:27:38 GMT
X-Mailer: Mozilla 2.0 (Win16; I)
Lines: 31

DrmWeaver2 wrote:

> Okay, so if a program sets goals for an android, it can never reach
> "intelligence"...What about us?  We are "programmed" throughout our lives
> - by our families, society in general, our own experience, etc. - to
> bahave in certain ways.  We conform or don't, accept certain goals or
> don't for numerous reasons (usually defined as "motivations") in order to
> gain something we want - friendship, money, recognition, etc.  That
> doesn't make us any less intelligent (tho' sometimes we act as tho' we
> HAVE lost all intelligence).

We are not programmed in the sense you imply above.  Social interactions 
are complex enough that I think basic chaos theory can be applied to them 
(you can't be certain what the outcome of interaction x with person y, 
because there are to many variables of unknown weight to be defined, much 
less put into the equation) so we make intuitive 'guesses' as to how we 
should behave to achieve certain goal.  This ability to intuit things 
like social interactions is part of what we define as intelligence.  The 
other part - storing and retrieving information - computers already do 
much better than us. 

However programming is a set of algorithmically defined responses, even 
randomness in the algorithm is not really random. So a program can 
algorithmically set goals for an android, but what will determine that 
androids 'intelligence' is how it proceeds to carry out those goals.  If 
x always leads to y because that is how it is hard wired, then it is not 
exhibiting intelligent behavior.  If x leads to y because that is the 
smart thing to do, then it is exhibiting intelligent behavior. 

Later
R.
