From: Jef Allbright (email@example.com)
Date: Sun Feb 15 2004 - 12:52:20 MST
It's interesting to note the near-religious fervor that accompanies the
discussion on this list of saving humanity via a single savior AI. It's a
common pattern, when apprehension of a serious threat to survival coincides
with an apparent single solution that will require a particularly narrow
focus to achieve. It's the kind of pattern cults form around, but it's
clear that Eliezer is aware of this, and equally clear he has no intention
of forming one. I admire his altruism, and readiness to admit that his
thinking changes as he grows in knowledge and experience.
However, I am concerned that humanity is approaching a critical time, when
the appropriate choices and actions will make a big difference, and that the
narrow approach is not the one true way.
A single solution is not natures's way, and I see all around evidence that
we are fundamentally enmeshed in nature, that there is a backgound process
of increasing complexity that drives all progress, and that complexity with
long-term survival value is accompanied by *increasing diversity*.
As the members of this list will generally agree, our greatest threats to
survival arise from humankind's relative immaturity and therefore lack of
wisdom in dealing with the next higher level of human interaction and
capability. We are now effectively a global organism but we don't have any
experience keeping a global organism alive and growing. Naturally, at the
same rough time frame that we achieve global connectivity and influence, we
also have access to technology that (for good and bad) has an inherent
global impact. Nature does not guarantee humans survival, but if we are to
survive, it is by using the tools nature has provided at this time and level
of complexity that we find ourselves in and by applying them in a way that
is most effective.
The tools that are now becoming available at this time and level of
complexity are information tools. Tools for sharing information widely
across the global organism so the organism can make effective decisions for
survival. In effect, tools for increasing wisdom. It's already happening
with the Internet, a global economy, and many other examples that members of
this list are well aware of.
I have little doubt that progress will continue, but there is great
uncertainty as to how painful it will be for humanity. Great changes
generally involve a process of breakdown before further growth can proceed,
and this is where we can focus our efforts. Most effective change occurs
when all the components are positioned for change at the right time.
My point is that we can contribute most effectively to this process by
increasing intelligence *at all functional points* within the global
organism. It's happening anyway, with human intelligence already becoming
augmented with media streams, personal computers, PDAs, mobile phones, all
networked, with the trend toward ubiquitous intelligence all around and
within us. We can't ignore the exponentionally enhancing power of networks.
Rather than put all energy and hopes into a great Friendly AI to take over
and protect us as in the myths of old, we would do better to place our faith
(yes, faith of the Bayesian kind) in spreading knowledge over a much wider
network of nodes and work with nature. Not one great AI, but innumerable
We'll be changed in the process, but we'll look back and see that it was
both natural and "good."
This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:00:45 MDT