From: David Clark (clarkd@rccconsulting.com)
Date: Thu Oct 28 2004 - 15:45:16 MDT
Response to Ben Goertzel re Universal Ethics
> I thought it was obvious that "pattern-sympathy", in that essay,
> was a *defined term*, like energy or force in physics, or like the notion
of
> sympathetic resonance in physics.
What is the point of redefining the English language to hide the fact that
your
*theory* about patterns *magically* creating something useful is just bogus.
You say I "displays such a basic lack of understanding" but I think I
understood
your article quite well. At least well enough to know that you are trying
to describe
a concept "compassion" which has a well defined English meaning into
something that
fits your concept of programming an AI.
There is a huge difference to the method I described for creating an AI and
yours. You
try to fit a number of different methods into an "agent" context where some
agents
degenerate and others stick around. I said that there is some peripheral
usefulness to
this concept but you have made it the base of your implementation. My
design
makes a set of models the basis for the AI and all of the other useful
techniques can
update and refine that model as they go. The models can be arbitrarily
simple or complex.
With your agents coming and going, you have no stable place to try different
techniques or
even see if your teaching of the AI is progressing.
I don't appreciate your condescending attitude at all. Your paper didn't
sound too much
different than others I have read from Marc Geddes. Eliezer has him pegged
quite
right in his last few emails. Marc was castigated for comparing himself to
Einstein.
(and rightly so) but you say "one has to look at evolution and ethics as
intertwined
manifestations of the same deeper phenomena." You go on to argue your
"deeper
phenomena". Your document purports to be above the theory of Evolution and
ethics.
If I made such a grand statement, I think I would have backed it up with a
bit more
facts and rigor (if you will).
Your document purports to have *arguments* for your hypothesis. I made
points, IMHO
that showed your hypothesis to be wrong and your arguments to be nothing but
definitions or conclusions. I think your take on Natural Selection is also
quite *novel*
approach and my textbooks from University (2000) on *phenotypic selection*
show
absolutely nothing on "pattern-sympathy". (Human Evolution An Illustrated
Introduction
by Roger Lewin 1999 or The Origin of Moden Humans by Roger Lewin 1998)
I wasn't the person that said your document was "it just argues that some
degree of compassion
is natural and inevitable.", you were. As my proof that compassion is *not*
inevitable I
call to the stand all the people I see around me that don't help little old
ladies across the street.
I call all the people in the Western World that have a car and a fancy house
when thousands
die of starvation all over the world. If you want to make a hypothesis
about something
that we can all see around us all the time, don't you think you should at
least be somewhat
convincing?
> To write something to be potentially convincing to
> people with opposing points of view would require a lot more time, space
and
> didacticism.
Do you always just write for people who might already agree with you?
> Who's your daddy indeed!!!
I made this attempt at humor to show how silly I think your giving human
attributes
to a pattern to be. I used the analogy that a fork doesn't *want* to pickup
food and put it in my mouth! If you think a pattern is something so much
more
than what the dictionary says, then maybe you can point me to your own
personal
dictionary so I can figure it out.
David Clark
PS I took the time to respond to your paper and you obviously took offense.
I also
take offense to you saying I have a "basic lack of understanding". Are you
really so smart?
This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:00:49 MDT