From: Jef Allbright (jef@jefallbright.net)
Date: Thu Nov 04 2004 - 07:36:37 MST
Eliezer Yudkowsky wrote:
> Keith Henson wrote:
>
>>
>> Coming at this business from the EP side, I think Collective Volition
>> might not even be a good idea, though it might be that I don't
>> understand some special meaning beyond the obvious meaning of the words.
>
>
> You're missing the "more the people we wished we were" set of
> transformations within "knew more, thought faster, more the people we
> wished we were, had grown up farther together". If you give a lecture
> on the evolutionary psychology of privation, and the lecture is
> correct, and the listeners understand it fully, and they think that
> this is not who they want to be when they grow up, then it wouldn't
> appear in their collective *extrapolated* volition.
>
> I'm not going to hook up a superintelligence to the decisions of
> Earthlings the way we are now. Geeze, do I look that suicidal?
>
"Collective Volition" is a nice idea on the path to better ones. It
breaks with reality in at least two major ways. (1) The extrapolation
error (2) The value of diversity.
Comprehension of the current collective volition of humanity is a useful
ideal, and attainable to a practical approximation, to guide social and
political growth in an incremental manner. However, we cannot predict
with any degree of confidence beyond the current horizon, in what
direction the growth of human culture and society will take us. Just as
a child's clear and simple wish of growing up to be a
"nurse/policeman/fireman/scientist" evolves into something richer and
more complex in actuality, we are fundamentally unable to extrapolate
the future of such a complex systems due to bounded knowledge and
constrained computational resources inherently inadequate to deal with
the combinatorial explosion and cumulative error that characterizes such
predictive problems.
However, there are principles of the growth of dynamical systems,
illuminated by an understanding of the current, not extrapolated,
collective volition of humankind, that can effectively guide us on the
next step of our journey.
One such principle is the value of diversity to the growth and
robustness of complex systems. While a collective vision is essential
for setting the direction of policy, it is vital to a robust
implementation that diversity be promoted within the overall policy
framework. Just as a robust immune system cannot be predesigned in it's
entirety to meet the challenges of an uncertain and evolving
environment, a single, complete solution to the "problem" (I consider it
a challenge) of human growth and survival must evolve rather than be
designed.
I have argued before, and will argue again here, that the more effective
way forward, at this current point in time, is to create and build tools
for effective augmentation of human awareness and intelligence within a
global network. Of course, this is already happening, but will benefit
from increased participation by intelligent and passionate minds such as
we find in this group.
I am encouraged by Eliezer's more recent comments that he is no longer
searching for a "complete" solution, but rather a "correct" partial
solution that won't blow up, that perhaps the thinking on SL4 is
gradually converging to something that may bear fruit of a practical sort.
- Jef
http://www.jefallbright.net
This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:00:49 MDT