X-Message-Number: 5986
Date: Sat, 23 Mar 1996 20:34:09 -0800
From: John K Clark <>
Subject: A responsible agent?
-----BEGIN PGP SIGNED MESSAGE-----
In #5951 (Thomas Donaldson) On Mon, 18 Mar 1996 Wrote:
>You talk as if Deep Blue were a responsible agent. Deep Blue
>comes nowhere near to being a responsible agent
It depends on what you mean by "responsible". When Deep Blue
wins a game it is the cause of the victory, it's not the
ultimate cause but I can't tell you what the ultimate cause of
anything is. Perhaps your question is if it's accountable for
it's actions, if it should be punished for it's failings.
I think the only logical or moral reason to punish anything for doing
bad stuff is to prevent more bad things from happening in the future.
I agree that punishing Deep Blue for failing would not help bring more
victories and at any rate I have no idea how to go about it. Deep Blue is
far too primitive for that, it can't come close to passing The Turing Test.
>The fact that Deep Blue could beat them [Deep Blue's designers]
>in chess means nothing
According to you when Deep Blue's designer plays the machine he ALWAYS gets
the credit for a victory regardless of how the game turns out. That's a
pretty sweet deal, but it doesn't seem quite fair.
>my Apple IIGS can calculate faster than I can,
Well that certainly means something, something profound, it
means it can do some things better than you can, and machines
get better every year, people do not, at least their hardware does not.
The fact that machines can already do some things better than we can has
led to huge changes to our society. I expect the list of things that
machines are better at to get longer, much longer, and lead to
greater and greater changes in our culture.
>but remains a tool, not an agent.
I think the distinction between a tool and an agent will become
increasingly blurry until we become our tools.
>As for intelligence, I don't believe it can be compared
>to "elegance". [...] anyone who can see that a proof is
>elegant must first understand a great deal about the
>mathematics involved.
And anyone who can see that a proof is intelligent must first
understand a great deal about the mathematics involved. We can't
define intelligence or elegance, but we can recognize it and
even on occasion produce it.
>Who is to decide and how are they to decide?
You are to decide, and by using the Turing Test.
>two twins in an institution who sat by themselves and told
>prime numbers to one another, nothing else, all day long.
>And those prime numbers were not obvious, but very large).
>Is such a person intelligent?
I don't know and I don't care. If it makes you happy to call
that intelligent then do so, if it doesn't then don't,
but I don't see what it has to do with the price of eggs.
>do you want some other brain module, as yet undefined, to do
>something we don't yet understand? And if so, what?
If I don't understand it and I can't define it, then I can't
tell you what it is but I can still recognize it when I see it,
and in fact do so every day.
John K Clark
-----BEGIN PGP SIGNATURE-----
Version: 2.6.i
iQCzAgUBMVTKCn03wfSpid95AQGGOATwxVe+toBtHO5R/GV3NoJeLj7N3a2tlnW0
Ppc4Aj8tBEAQNM/sz3s96nZ2YVPcFUIqMUszN1hDqMNuZKwSlxrM96SVhn3RtWbw
IHigWD3iVAUsjpRbPTBMOl1QhZPbj8/kCSaDgGFKM7/xZfWsKmW/a1fu7hGClV7O
PQEQXV6LXdKyrgnTe8+uNTfM+QDmYzAFXDsYo+KGh79N4jil38Y=
=oCNb
-----END PGP SIGNATURE-----
Rate This Message: http://www.cryonet.org/cgi-bin/rate.cgi?msg=5986