Newsgroups: comp.ai.genetic
From: Dan@gulch.demon.co.uk (Daniel Sumption)
Path: cantaloupe.srv.cs.cmu.edu!rochester!udel!gatech!howland.reston.ans.net!news.sprintlink.net!peernews.demon.co.uk!gulch.demon.co.uk!Dan
Subject: Re: New Penrose Book - Poll
References: <3kc574$n3j@newsbf02.news.aol.com> <D5LynH.1G8@Corp.Megatest.COM> <D5stzE.E9v@aplcenmp.apl.jhu.edu>
Organization: Bromide Gulch
Reply-To: Dan@gulch.demon.co.uk
X-Newsreader: Demon Internet Simple News v1.29
Lines: 41
X-Posting-Host: gulch.demon.co.uk
Date: Fri, 24 Mar 1995 21:18:22 +0000
Message-ID: <796079902snz@gulch.demon.co.uk>
Sender: usenet@demon.co.uk

In article <D5stzE.E9v@aplcenmp.apl.jhu.edu>
           krobair@aplcenmp.apl.jhu.edu "Robair kevin a. 301-925-0694" writes:

> Eventually, human intelligence will be duplicated
> artificially, whether using a digital computer (doubtful) or a device based 
> on some as yet undicovered physical phenomena. We may not be around to see it,
> but it sure would be interesting. Imagine such an intelligence coupled with
> the raw computing power of a digital computer. Awesome, but dangerous.
 
This seems to be an old argument repeated with monotonous regularity (and
maybe more suited to comp.ai.philosophy) but if, at the most fundamental
level, brain activity can be seen to arise from the firing or non-firing of
synapses (a digital process) then there seems to be some hope that we will
eventually be able to make a reasonable attempt at replicating this digital
process in silicon machines.

Most of my knowledge of this area I gained from Douglas Hofstadter's
brilliant book "Godel, Escher Bach" (plus the few Psychology lectured I 
managed to attend) so I may be a little biased in my views, but I see a lot
of sense in Hofstadter's view that human intelligence & consciousness must
be seen not just in terms of the 'base level' of digital occurences, but
in the increasingly complex self-referential hierarchies which arise from 
the digital phenomena, and the combination of digital phenomena, and the 
combination of combinations of digital phenomena ....

Hofstadter also has something to say about the idea of "human intelligence
coupled with the raw computing power of a digital computer", which is that
the two concepts are incompatible.  Humans do not forget things or get 
confused because our brains have evolved in some sort of sloppy organic
way which makes them less than ideally suited to data processing.  We do it
because we must accept input on a much higher level than a modern computer,
discarding any sensory input which seems irrelevant to our current aims
and making continuous reasoned judgements in the face of ubiquitous 
ambiguities.  If a silicon machine had to make real-time decisions in an
uncertain universe, using a similar amount of processing power to ourselves,
we might expect it to make similar errors.

Food for thought?

-- 
Daniel Sumption
