From: Robert J. Bradbury (bradbury@aeiveos.com)
Date: Sat Jun 30 2001 - 01:30:17 MDT
On Fri, 29 Jun 2001, Harvey Newstrom wrote:
> Nobody is saying that A.I. can't be predictable. A chess program is
> predictable, but it simulates chess playing. A calculator is predictable,
> but it can solve new problems.
But the chess program and calculator solve those problems according
to deterministic rules that can be determined if you have access
to the underlying code. The fact that you can "lay-out" the
underlying deterministic rules seems to contradict what many
would want to define as "intelligence". I would argue that
"intelligence" is having a large repeirtoire of pre-existing
patterns (that in some cases generate responses that manifest
in the physical reality) combined with a robust pattern matching
algorithm that is good at selecting the pattern likely to generate
the results you want in specific situations.
In so far as GAC doesn't have a need to have a result matching
its "desires", it is not "intelligent". In so far as it returns
an answer based on popular opinion (without an analysis of the
methods used to reach that opinion) it is also not "intelligent".
In so far as it returns the "common perspective" it is "intelligent".
It looks like "intelligence" may be another suit-case term.
> GAC seems to be only to answer questions with pre-recorded answers.
I would argue that much of what one encounters in human-to-human
interactions is "predictable" and cannot solve new problems.
When I ask you "How much is 2+2?" do you answer 4 based on
the fact that you understand the number theoretic aspects
of this question or because you were programmed to give this
response? People give "pre-programmed" responses based on a
"closest-fit" to previous responses that produced favorable results.
It would be far more interesting to watch someone who has had
no training in numbers or math answer "How much is 2+2?" E.g.
someone probably between the ages of 1-4. Humans are pre-programmed
with some concept of number/quantity -- where we get to the roots
of intelligence is in looking at how people who don't know the
pre-programmed answer parse and resolve the question. The other
point of reference is in the idiot savants who can perform phenomenal
mathematical calculations quite quickly and give you accurate answers.
I think I may adopt a perspective that "artificial intelligence"
is a suitcase term henceforth. So the use thereof on the list
is likely to draw flack from my perspective in that it lacks
specificity and is therefore counter-extropic.
### "When I use a word," Humpty Dumpty said, in rather a scornful tone,
### "it means just what I choose it to mean -- neither more nor less."
### "The question is," said Alice, "whether you CAN make words mean so
### many different things."
### "The question is," said Humpty Dumpty, "which is to be master --
### that's all."
Its unextropic (read unproductive) to slant words in favor of
your arguments. The precise definition of terms and what they
really mean determines whether or not productive debates may
be entertained.
Cross that line if you dare...
(Harvey, this is in no way a comment on your comment, its just
where your comment lead my thought patterns)
Robert
This archive was generated by hypermail 2.1.5 : Sat Nov 02 2002 - 08:08:22 MST