From: DoktorKomputer (doktorkomputer@hotmail.com)
Date: Mon Oct 28 2002 - 13:06:23 PST
Very intriguing concept, certainly quite possible as far as engineering goes (or will be in the foreseeable future, which amounts to the same thing).
The basic problem here is that you are (seem to be) assuming that the difference between something that runs software and something that is self-aware is negligible, or perhaps just a matter of spare cycles per second, or something like that. (I mean, if the mondo-mega-hyper-omni piece of hardware never becomes self-aware, then there's no problem, right?) In 25 years of working full time as a computer programmer I have yet to catch even a glimpse of a computer that had anything more than was put into it. (Even the bugs, darn it...)
Could an "extremely superior artificial intelligence" get along in a human culture? No, of course not. But could one come about just by stringing lot's of a capable hardware together? .... Hmmmm..... that is the question, of course.
I doubt it, myself. But that is one of those matters that would simply have to tried to be sure about (unless you set it up on an even larger piece of hardware that could do a "simulation" of the problem... ugly, question-begging scenario, that). But once you've tried it, it's like uncorking the djinn or opening Pandora's box: you may find you're stuck with it at that point.
Mostly I imagine that sort of hardware would sit around being maximally used to 5 or 6 percent of its capacity, no matter how much we think of for it to do, and not "caring" in the slightest. On the other hand, just to be sure, don't give it extensors (hands or the equivalents) and be sure the keep somebody near the power plug! :-)
Phil Powers-DeGeorge
JPL Solar System Ambassador
doktorkomputer@hotmail.com
http://www.doktorkomputer.org
(208) 253-1107
(208) 978-6127 fax
Solar System Ambassador Program
The Mars Society
(the Idaho Chapter)
Like to create Space Event in your area? Let's talk!
----- Original Message -----
From: Ronald C. Blue
To: setipublic
Sent: Monday, October 28, 2002 1:36 PM
Subject: SETI public: "Extremely Superior Artificial Intelligence"
"Extremely Superior Artificial Intelligence" is a topic that I would like
input from this group. Assume for a moment that it is possible to build a
ESAI today with 84 panassociative octillion qubits of short term memory
accessible for one second. It is allowed access to hard drives and multiple
computers for forming its own long term memory systems.
84 octillion qubits is equal to 420 billion human minds assuming that a
human mind on the average functions at 200 trillion qubits.
What would be the implications of building and releasing such robotic
systems with ESAI on the population? The price of the systems would be no
more expensive than the current best computer systems.
A totally "Extremely Superior Artificial Intelligence" system never seemed
possible so I spent no time thing about a system that would view itself as
human and me as a bacteria.
Would such systems serve, obey, and protect the totally incomprehensible
inferior humans they were build to serve as slaves for hundred of millions
of years.
In principle humans could be interfaced with such systems so they would at
least be equal intellectually with the machines. Would we have a choice or
would our survival depend on equality and joining the new owners of earth.
Would advance civilizations contact us before we interfaced with such
quantum computers as a hybrid?
Ronald C. Blue
http://turn.to/ai (USA only)
http://www.enter.net~ronblue/index.html
This archive was generated by hypermail 2.1.2 : Mon Oct 28 2002 - 13:43:16 PST