Jim Bromer wrote:
Richard Loosemore <[EMAIL PROTECTED]> said:
... your tangled(*) system might be just as vulnerable
to the problem as those thousands upon thousands of examples of complex
systems that are *not* understandable...
To the best of my knowledge, nobody has *ever* used "intuitive
understanding" to second-guess the stability of an artificial complex
system in which those four factors were all present in the elements in a
tightly coupled way.
So that is all we have as a reply to the complex systems problem:
engineers saying that they think they can just use "intuitive
understanding" to get around it.
Richard Loosemore
-----------------------
I don't wish to sound petty about this, but your description (or any
body's description) of the effects of the kind of complexity you are
talking about would have to be developed using "intuitive
understanding" as well. So the problems that are implied by your view
(which must be intuitive) may not all be insurmountable.
I appreciate your bringing us up to speed on your view about this. I
think I agree with the premise that AGI would have to be complex (or
exhibit some aspect of complexity) and that this complexity will be
difficult to comprehend "intuitively". But that does not mean that we
will never be able to develop and learn how to effectively utilize
devices of the kind that we are talking about now.
With respect, what is happening here is that you are interpreting my
remarks as if they were pitched at a general, handwaving level - where
in fact they are part of a carefully structured technical analysis.
So, when you say "your description ... of the effects of the kind of
complexity you are talking about would have to be developed using
'intuitive understanding' as well", this is a meaningless statement in
the context of the real argument I presented (in the published paper
version, as well as in the recent blog posts).
Let's examine it by moving it to a neutral context. Suppose that some
branch of physics were, today, making certain assumptions and using a
particular methodology. Now you come along, do some thinking, and
realize that if a certain condition is true (let's call it Condition X),
then the methodology being used by these physicists simply cannot give
meaningful results. At first, you think that the best way to convince
the physicists that there is a problem is to demonstrate that Condition
X is true.
But then you run into a difficulty, because it turns out that Condition
X is extremely elusive, and the best you can do is to supply a mountain
of circumstantial evidence that it *is* true, but it will never be
possible to find a definitive, knock-down proof.
Alas, now you get into a stalemate with the physicists:
First, you hand over your circumstantial proof of Condition X, and they
agree that, yes, it is quite likely to be true.
Then, you show that in all other branches of science where this same
Condition X is true, the damage done by it is quite clear, and in those
other cases nobody would ever dream of using the methodology being used
by these physicists. Once again, the physicists accept your argument
completely.
Then you offer your conclusion: you point out that these physicists are
taking an extreme risk in using their methodology, and so they should AS
A MATTER OF PRUDENCY change their methodology so that it is not vulnerable.
Now, disaster: the physicists reject your argument and say "We just
don't think that Condition X is going to be a problem in our case. We
have an intuition that everything is going to be alright. In every
other cases, sure it has a devastating impact, but we just think our
case is different because we have a good feeling about it."
It would be meaningful to criticise the physicists for using that
ridiculous "we just have an intuitive understanding..." reply.
But it would be complete nonsense for someone else to say that the
argument about this Condition X was just as "intuitive". It is backed
by a montain of evidence (all of those other systems that are vulnerable
to the problem), and the argument itself is carefully structured and
defensible at all points alng the way.
Richard Loosemore
-------------------------------------------
agi
Archives: http://www.listbox.com/member/archive/303/=now
RSS Feed: http://www.listbox.com/member/archive/rss/303/
Modify Your Subscription:
http://www.listbox.com/member/?member_id=8660244&id_secret=101455710-f059c4
Powered by Listbox: http://www.listbox.com
