Newsgroups: comp.ai.philosophy
Path: cantaloupe.srv.cs.cmu.edu!das-news.harvard.edu!news2.near.net!MathWorks.Com!zombie.ncsc.mil!admii!ovation!rscanlon
From: rscanlon@pica.army.mil (Raymond D. Scanlon (CCB))
Subject: Re: Strong AI
Message-ID: <CwEJ4x.441@pica.army.mil>
Sender: usenet@pica.army.mil (USENET Special Account <usenet>)
Nntp-Posting-Host: grant
Organization: U.S Army ARDEC, Picatinny Arsenal, NJ
Date: Tue, 20 Sep 1994 00:10:09 GMT
Lines: 53

In article <push-130994094940@mind.mit.edu> Pushpinder Singh wrote:

>In article <352i6s$7d5@portal.gmu.edu>, herwin@mason1.gmu.edu
>(HARRY R. ERWIN) wrote:

>> <stuff deleted>
>>
>> I suspect the argument from chaos is the most direct attack on
>> Strong AI. We already know that any digital simulation of a
>> chaotic process will eventually drop into a limit cycle.
>>
>> <more stuff deleted>
>>
>> Comments?

>This chaos attack isn't a problem at all.  I can't imagine
>detecting and compensating for limit cycles to be terribly
>difficult.  And if I'm wrong, it wouldn't be hard to hook up what
>you would consider a "true" random number generator to a digital
>computer (say, using the thermal noise from a resistor).  Still,
>good pseudorandom number generators take long enough to loop that
>I don't see why they wouldn't be good enough (you could always
>reseed once in a while off some presumably "true" random data
>point in the "real" world.)  But then again, if your machine has
>a sensory connection to the "real" world, the randomness of the
>sensory input itself ought to keep the machine from going into
>pointless loops.

>Finally, whose to say that people aren't subject to these "limit
>cycles" either?  Consider someone in a sensory deprivation tank,
>or a tenured professor :)

>-push

IMHO this point needs to be made at least once every three months
on this net. Some take digital to be synonymous with deterministic
and analog with non-deterministic. This is simplistic. It is an
abstract digital computer with no input from outside that is
deterministic. A real digital computer is easily interfaced with an
analog computer, with thermal noise, or with sensory input.

Anyone who would speak of AI should concern himself with real
machines, not abstractions.

Ray



"What is thought except a movement that is not connected to a motor
neuron."
          Attributed to Walle Nauta


