On 10/14/2012 11:36 AM, Craig Weinberg wrote:
On Sunday, October 14, 2012 2:19:14 PM UTC-4, Brent wrote:
On 10/14/2012 10:36 AM, John Clark wrote:
On Sat, Oct 13, 2012 Roger Clough <[email protected] <javascript:>> wrote:
> But if a computer beats you at an intelligent task, it would have to
be
programmed to do so.
And you would have to be educated to do so.
> which means that its intelligence would be that of the programmer.
Then how can the computer beat that very same programer?
I think this illustrates a problem with defining 'intelligent task'. Is
multiplying
to numbers together an intelligent task? It is to 2nd graders, but not to
the
programmer. Why? Because the programmer knows an algorithm for doing it
that is
simple enough to keep in his head. But to the programmer winning a chess
game is an
intelligent task - because he doesn't know exactly how to do it; he doesn't
know an
algorithm for doing it. He knows that he looks ahead and evaluates a few
moves and
countermoves and the further ahead he looks the more likely he is to win.
So he can
program a computer to do the same kind of thing, but look further ahead and
apply
the same evaluations, and then it can beat him. But it can't beat a really
good
chess player because the really good chess player has a better set of
evaluation rules.
Yes. So far so good.
Then the programmer can arrange for the the computer to look through and
enormous
number of chess games and infer its own rules
Not necessarily.
Not necessarily what?...the programmer can't arrange for the computer to learn from
previously played games? Why not? Did you actually read what I wrote, or are just
responding to the voices in your head?
Just because it looks to us that the computer is following rules doesn't mean
that it is.
Of course it might be following rules like: In this position P->N3 with probability 0.6
and Q->Q2 with probability 0.4. There's nothing that requires rules to be deterministic.
We should not assume that suddenly a disembodied conscious
Where did I write 'conscious'? Who are responding too?
agent appears somewhere just because we are impressed with the sophistication of a
particular reflex action. Reflexes can be as complicated as we want to make them, it
doesn't turn them into voluntary actions. The computer still has no choices.
It has lots of choices - that's what playing chess consists of, choosing moves.
It can't throw a match because it doesn't want to hurt someone's feelings.
Neither could Bobby Fisher.
which may then be so good as to beat even a really good chess player. But
now
neither the programmer nor anyone else know what rules the computer uses.
They
could look at a printout of the program and see what they are, but neither
they nor
anyone else, including the computer, would know why those particular rules
were the
good rules. To know that they would have to retrace the whole learning
process the
computer used to derive the rules. So NOW the computer is intelligent,
just like
the programmer, because it can win chess games but doesn't know exactly how.
It never knew how in the first place.
So you assert. But it could beat you.
What makes intelligence is the ability to step out of the system, to transcend the rules
entirely or understand them in a new context.
What's "the system"? The computer has the ability to step out an Indian defense and
create a line of play that next existed before. You're just trying to make up new
definitions of 'intelligent' to save your hypothesis. I very much doubt that you can
'step out of the system' of Newtonian spacetime and understand the warped spacetime of
general relativity - although you may be able to follow some rules (i.e. equations) to
make some inferences about it; but of course that didn't prove Einstein was intelligent.
You have an intuition that you're intelligent and computers can't be, and you're sure
you're right because you have no access to the the thought process that produced this
intuition. If you did you'd reject it as 'just an algorithm'.
Computers don't do that. They have no idea if they have invented the most brilliant new
way of winning at chess
But they would if they were chess history computers, or chess computers with a
history module.
or whether they have counted to 500 for the 500th time for no reason at all.
Is the computer conscious? I'd say it was a little bit conscious (of
chess).
Not even that. It wouldn't know a chess set
And you apparently don't know the difference between chess and a chess set.
if it was dropped on top of them. It's nothing but a sculpture called 'the machine that
greatly exceeds our expectations'. It has no point of view.
It has a point of view about chess - and it's a better, more comprehensive view
than you do.
Brent
--
You received this message because you are subscribed to the Google Groups
"Everything List" group.
To post to this group, send email to [email protected].
To unsubscribe from this group, send email to
[email protected].
For more options, visit this group at
http://groups.google.com/group/everything-list?hl=en.
