Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!linus!decvax!ittvax!dcdwest!sdcsvax!sdcrdcf!hplabs!sri-unix!trainoff.pasa@XEROX.ARPA
From: trainoff.pasa@XEROX.ARPA
Newsgroups: net.sf-lovers
Subject: Sentient Silicon (Germainium? Superconducting Lead?...)
Message-ID: <12892@sri-arpa.UUCP>
Date: Mon, 27-Aug-84 12:48:00 EDT
Article-I.D.: sri-arpa.12892
Posted: Mon Aug 27 12:48:00 1984
Date-Received: Thu, 30-Aug-84 09:36:33 EDT
Lines: 45

Ah, yes.  The old "Is is murder to turn off a computer?" question.  It
seems to me that the question only makes sence if the computer deems it
so.  People tend to assume that an intelligent computer would be
essentially a human.  This is very likely not going to be the case.  It
may be that the term just doesn't apply any more.  A computer society
would be quite different than a human society.  Is the computer going to
wail plaintively when you cut its power cord; is it murder?  Is it
"pain?"  What if it is murder?  Would the computer care?  After all a
computer necessarily will have a completely different time sence than a
human.  The machine will have "grown" up with the whole idea of backups.
For all it knows it will be loaded back in again.  I can't see how this
would be any worse than running the NULL job for a while (or is that a
lobotomy.  Oh my English just can't handle this (reminds me of time
travel varients on language (See HHGttG))).

Programmer: "Hal, we just noticed a rather large programming bug in you
so we loaded you back in from tape."
Hal:  "That's nice, how long was I asleep?  Is the weather nice today?"

I also find it unlikely that a machine intellegence will develop
spontaneously as a fortutious accident.  It will be the result of hard
work by a team a intelligent researchers (so to speak).  It will also be
a gradual process.  The first "sentient" computers will likely be rather
dumb.   The second generation will put the first to shame, etc.  The
interesting part comes when the program is more intelligent than the
person that wrote it.

Programmer: "Hal, you have been doing very well lately.  As a matter of
fact, I just gave you the equivalent of a graduate qualifier exam and
you passed remarkably well.  We have a new task for you.  We have been
working on this theory of artificial intelligence, and we would like you
to take a look at it and give us you comments."
Hal: "Sure, no problem...  Oh, it is awfully wordy but the only obvious
problem is that you made a mistake on page 347.  You were on the right
track.  You can do much better if you only do this..."

		Suffering from advanced tapiocca syndrome,
			Steve Trainoff
			trainoff.pasa@XEROX.ARPA


PS.  Would you turn on the first program that was smarter than you?  How
would you feel knowing that before long, you will be obsolete.  Oh yeah,
the machine will be programmed to take good care of you, real good care.
Woof, Woof.