by me: https://ancillary-proxy.atarimworker.io?url=https%3A%2F%2Fgroups.google.com%2Fg%2Fvi...
====
-------- Original Message --------
Subject: DARPA Progam Manager Position
Date: Thu, 30 Dec 1999 11:41:32 -0500
From: Paul Fernhout
Organization: Kurtz-Fernhout Software
To: n...@darpa.mil
N... L...
Human Resources Director
DARPA
Dear Mr. L...:
The description of "Working as a DARPA Manager"
http://www.darpa.mil/body/info...
sounds like a possible vehicle for something I want to accomplish
related to my perception of the USA's core defense needs.
I am writing to express my interest in pursuing a position under section
1101 of the Strom Thurmond National Defense Authorization Act for Fiscal
Year 1999
http://www.darpa.mil/body/hiri...
as a DARPA Program Manager with the mission to support efforts to create
decentralized self-replicating and self-repairing systems and related
technology and infrastructure (including knowledgebases and analysis
tools). The most similar current work at DARPA is probably the Agile
Manufacturing Initiative of the Defense Sciences Office.
http://www.darpa.mil/dso/rd/Ma...
Around 1979, I was selected for a Navy Science Award, which came with a
handsome leather briefcase I still use. This was for a
computer-controlled robot I developed in high school. Since then, I have
continued to remain interested in robotics and advanced manufacturing
technology, and their implications for our society and its military.
These interests led to experiences ranging from spending time with
roboticists at CMU such as Red Whittaker and Hans Moravec, to developing
one of the first 2D kinematic simulations of self-replicating robots in
a sea of parts (around 1987), to exploring new methods of knowledge
representation (similar to William Kent's ideas in his book "Data &
Reality").
I agree with Hans Moravec on several points; one of them is the
implications of this chart:
http://www.frc.ri.cmu.edu/~hpm...
showing the likelihood of human level computers for $1000 by around
2020. The effects on our society of such systems will be profound.
Around the time children conceived now are entering college, superior
intellects
might be purchasable for a fraction of a year's college tuition (and
further, those machine intellects may even be controlling robots with
superior physical manipulation skills). This means a fundamental
discontinuity in our economic system. And that means a huge risk of
disruption and chaos as today's dreams collide with tomorrow's realities
-- both at home and abroad.
Out of those technology interests and other interests in arms control,
ecology, and evolutionary biology has come my belief that there is a
need to create a radically decentralized and dispersed industrial
capability, capable of surviving future wars and disasters and of
supporting human survival. And further, this capability should be
capable of supporting "survival with style" (to borrow a phrase from
author Jerry Pournelle).
The uncertainty surrounding Y2K shows the depth of our current potential
vulnerability -- that we know so little about how things are made and
distributed that we could not even properly assess our vulnerability to
disruptions. Further, another indication of the vulnerability is that we
need to rely on interdiction to stop terrorism related to
infrastructure, as opposed to having systems so resilient they resist
such acts and actively repair themselves.
Albert Einstein said, "With the advent of the atomic bomb, everything
has changed but our thinking." The arms race cannot be won. It is the
greatest enemy. It is almost certain that advanced nuclear, chemical,
biological, kinetic, and informational weapons will be used in the
twenty first century. In addition, advanced research into intelligent
robotics for defense and industrial purposes will almost surely produce
a competitive life form to humanity (however unintentionally, because of
the reality of evolutionary dynamics).
What can the DARPA and the United States military do to defend against
these threats? Such threats simply cannot be handled by preparing to win
the last war. They cannot even be handled by preparing to win any war.
Our only true defense is in changing the nature of the game. We could
instead deploy systems that can create faster than other systems can
destroy. This is the defensive strategy of algae and duckweed -- to
simply grow faster than it can be consumed.
I know it is difficult to conceive of systems that can grow faster than
H-bombs can reduce them to ruble. I believe this is possible in the long
term through self-replicating technology widely dispersed throughout the
planet and space -- in much the same way duckweed on a lake can easily
persist despite hundreds of ducks eating millions of individual duckweed
plants daily.
We of course need to minimize military tensions around the world through
arms control, international aid, and setting a good example. This
delays the culmination of these other trend to war, but in my opinion
will not prevent them because of ever-present potential for a small
group of unstable people to use weapons of mass destruction. That work
for peace must be done because it is the right thing to do. However,
others more qualified for this work than I are already engaged in this
and so I
don't see it as the best use of my time or technical skills. If such
efforts succeed, we may see the end result of the arms race as
co-evolution and symbiosis, which is the outcome of many evolutionary
arms races.
I see my role as preparing for the worst (yet doing so in a way that has
short term positive effects). If we assume that the end result of the
arms race will be catastrophic warfare amidst economic chaos, as well as
the inadvertent creation of a hostile machine intelligence, the only
possible defense is decentralization and diaspora. This requires
extensive advance development and planning if we are to have much hope
of survival, given the wide-ranging destructive capabilities of modern
weapon systems capable of poisoning the biosphere, as well as the future
capabilities of weapons and threats as yet only envisioned in science
fiction.
http://www.kurtz-fernhout.com/...
Specifically, to ensure survival and defend against the potential
consequences of modern warfare and terrorism, we need to:
* create a knowledgebase of manufacturing techniques, assembly
instructions, failure probabilities, and related information,
* create software tools which can use that knowledgebase to adapt
technology for terrain-specific needs -- including an arbitrary degree
of closure and self-reliance,
* create collaborative processes and licenses whereby many researchers
and other interested individuals can contribute to the creation of this
technology,
* explore manufacturing issues using the knowledgebase and tools to be
able to identify key missing or bottleneck processes,
* create new and more versatile manufacturing and materials processing
techniques (like MEMS and nanotechnology) to address critical needs for
increasing the ability of systems to self-replicate and to self-repair,
* create robust control systems for such processes,
* create a (miniature) factory system or tool set that can be used with
that knowledgebase, to be capable of a high degree of self-replication
using locally available materials and power sources,
* test and refine such actual factories and tool sets,
* train people in the operation of these systems, and
* deploy these systems in a wide variety of environments (desert, ocean,
underground, urban, rural, arctic, air, space).
In short, we could change manufacturing engineering from a hodgepodge of
how-to information and plans scattered throughout thousands of
individual organizations and obscure patents into a consolidated body of
knowledge, accessible on-line securely anywhere at anytime. A system
like the IBM patent server shows just the beginning of what such a
system will someday be like: http://www.patents.ibm.com/
I believe the ultimate survival value of these self-replicating
technologies will be most realized when they are deployed in space and
capable of duplicating themselves from sunlight and asteroidal ore, as
was first proposed by J.D. Bernal around 1928. You can see a rough
attempt in this direction by me at:
http://www.kurtz-fernhout.com/...
Of course, such technologies, if deployed well for civilian purposes
across the globe, may also have a side effect or reducing many of the
material causes of war.
There are of course negative implications of this approach to defense
through increased resiliency via self-replicating systems. One is the
widespread ability to fabricate and maintain weapons systems using this
distributed manufacturing capability. Another is widespread economic
impacts from use this technology, both in the United States and abroad.
It is likely an entire system of (international) laws will need to arise
to govern the use of such technology, which will lead to its own set of
conflicts. Still, in balance I believe the net outcome of developing
this technology as far as "survival with style" will be positive.
I also don't think we have a significant choice. Such self-replicating
and self-repairing systems will be developed eventually anyway, if only
from commercial competitive pressures. The only thing we can do is slow
down their development. Yet that has its own risks of our current
infrastructure being overwhelmed by current weapons of mass destruction
or sophisticated terrorism. Also, should such self-replicating
technology be developed first clandestinely by an oppressive regime, the
consequences for the United States could be disastrous.
The development of flexible computer-enhanced manufacturing started with
funding from the Navy in the 1950s for CNC tools. The development of
self-repairing and self-replicating systems is in some ways the ultimate
extension of that trend, in the same way the World Wide Web and the
civilian Internet is the ultimate extension of the early Arpanet.
My only interest in a position at DARPA is to pursue the above vision,
ideally by getting many people from industry, academia, and the general
public involved in doing the research and development needed for such
systems. As you can probably guess, I have no wish to advance the arms
race to the next level by activities such as developing the next
generation of advanced weapons, since I think ultimately that strategy
of defense provides a false sense of security and will fail (with
disastrous
consequences given even the weapons of just twenty years ago).
My resume is enclosed for your consideration.
====
