Foresight Update 2
page 3
A publication of the Foresight Institute
 Computational
Markets
"Agoric systems" may be of use in developing
intelligent machines, with all that that implies. Mark S. Miller and K. Eric Drexler, working at the Xerox Palo Alto
Research Center, have recently completed a set of
three papers on this subject, to appear in a book titled The
Ecology of Computation (Bernardo Huberman, ed., Elsevier
Science Publishers, North Holland, 1988).
A great challenge in artificial intelligence, and in computer
science as a whole, is to build working, useful systems more
complex than any single human mind can design or fully grasp.
That such systems are possible is clear--living organisms and
societies are examples. Both have emerged through evolutionary
processes, through the variation and selecton of replicators.
Evolutionary principles can be applied in computation as well.
The first of the three papers, "Comparative
Ecology: A Computational Perspective," compares
biological systems and idealized markets from an evolutionary
perspective. Considering the goal of building systems that serve
external purposes, it concludes that idealized markets, based on
cooperative trade, provide a better framework than
biological-style systems, based on highly non-cooperative
predation. A variety of other ecological systems are examined
more briefly.
The second paper, "Markets
and Computation: Agoric Open Systems," considers in more
detail how prices and market forces can shape cooperation among
pieces of software, fostering the evolution of efficient, capable
computational systems. The third paper, "Incentive
Engineering for Computational Resource Management,"
examines the problems involved in using markets to allocate
memory space and processor time to software, and describes a set
of initial algorithms that could help get the process off the
ground.
Agoric open systems are a proposal for a new way to organize
computation, blending design and evolution. If successful, this
approach could encourage the growth of large, capable systems of
software--including automated engineering systems able to speed
the development of new technologies. In the course of the last
year, agoric systems have become a research objective at Xerox
PARC.
Spreading
Memes
Since last issue the concepts of nanotechnology, hypertext,
and agoric open systems have been presented to diverse audiences,
from top technical conclaves to a high school summer program.
(The latter is the Space Sciences Academy at Stanford, which
included nanotechnology in its program again this year.)
Top researchers were introduced (or in many cases reintroduced)
to nanotechnology at the first Artificial Life Workshop, held at
Los Alamos National Lab. This workshop covered a broad range of
interests, from spontaneous order resulting from
biologically-inspired processes to construction of systems with
lifelike characteristics; it included discussion of software,
robotics, and genetic engineering, as well as nanotechnology. A
proceedings volume is in progress and promises to be interesting;
we'll report when it's available.
Other groups learning about the subject included the third
Hackers Conference (a meeting of computer programmers), the
Reality Club (watch for their upcoming publication series by the
same name), and a group at U.C. Berkeley including Prof. Hubert
Dreyfus. Researchers at the Digital Equipment Corporation are
learning more on nanotechnology after having their interest
stimulated by David
Forrest, an MIT doctoral candidate and NSG member who covered
the topic heavily as part of his summer employment assignment to
report on molecular computing.
"Hypertext" is of course a popular buzzword now. FI's
thoughts on what constitutes a desirable system were presented at
the Directions and Implications of Advanced Computing symposium,
sponsored by the Computer Professionals for Social
Responsibility, and were published in the proceedings, DIAC
87.
Design work on an open distributed hypertext system (the
"Linktext" proposal) was presented by designer Robin Hanson
at a Tech Talk at Xerox Palo Alto Research Center. Eric Drexler
introduced Hanson's talk with a brief explanation of the value of
such systems.
Work on agoric open systems, explained elsewhere in this issue,
was discussed by co-authors Mark S. Miller and Eric Drexler at a
workshop on logic programming and open systems held at Xerox PARC
in early September. Miller also presented the ideas at the
Artificial Life Workshop.
Interview:
Eric Drexler
Continued
from last issue
FI: Who's doing the work in this field today?
Drexler: Well, the question of who's doing work in the
field very much depends on what one means by "the
field." If you look at the full range of fields that are
contributing to the emergence of nanotechnology--protein design,
synthetic chemistry, scanning tunneling microscope technology,
molecular modeling on computers--there are dozens or hundreds of
research groups, in industry and academia, in the US, Europe,
Japan and the Soviet Union, and so the numbers are large and
what's going on is very diverse. At the other end of the
spectrum, looking at nanotechnology itself--at what can be done
with real assemblers--that's a development that's far enough off
in the future that it doesn't make sense for industry, for
example, to be working on it. Therefore people are just beginning
to think about it, and at this point hardly anyone is doing work
on assemblers and what can be built with them.
FI: What technical work are you doing now in the field?
Drexler: I'm currently working on a series of papers to
fill in more detail on the design of things such as molecular
machines, molecular mechanical computers, assemblers, and
ultimately cell repair machines. The first paper, which discusses
the details of molecular structure, motion, and thermal noise for
the logic elements of a mechanical nanocomputer, will appear in
the Proceedings of the Third International Symposium on
Molecular Electronic Devices, scheduled for publication in
1988.
- For an extended discussion of the topic of the
above paper, see chapter 12 of Nanosystems.
FI: When will we develop genuine nanotechnology?
Drexler: This is very hard to say, again because we don't
fully understand the ground to cross between here and there, or
just how soon the "exploration parties" will set out,
how lucky they will be, and so forth. A friend of mine, Roger
Gregory of the Xanadu hypertext project, likes to say that his
optimistic estimate is thirty years, and his pessimistic estimate
is ten. Though I should hasten to add that Roger not only worries
about the dangers of this technology but also expects us to
benefit greatly from its positive uses. But in terms of that
range of dates, I don't know of a good argument against it.
FI: What are your greatest hopes and fears for the future?
Drexler: In the long run my greatest hope is that we will
handle the coming revolutions in nanotechnology and the
comparable or even greater ones in artificial intelligence so
that we can benefit from their enormous potential. The fear of
course is that we'll wipe ourselves out or paint ourselves into
some very ugly corner.
In the shorter term, though, my greatest fear is that as this
technology moves forward the debate will polarize between groups
that blindly support the technology--seeing its benefits for
everything from economic well-being, to medicine, to improving
the lot of people in the Third World--and those who blindly
oppose it, seeing its potential for abuse in the wrong hands and
perhaps imagining dangers that aren't even there. I'm afraid that
what we'll see is another round of fruitless public mudslinging,
with opposed sides not really addressing each other's cases and
simply trying to stir up as much emotion as they can for their
side of the argument. You can imagine a "debate"
polarized between the followers of someone like Lyndon LaRouche,
pushing technology with inflamed rhetoric, and the followers of
someone like Jeremy Rifkin trying to block it completely. Because
if that happens, it's unlikely that we're going to be
well-prepared for these developments when they emerge, and it's
even possible that this will paralyze the democracies as these
technologies emerge.
My hope is that we'll see a diverse, quarrelsome, but basically
united center--one that embraces people who fear these
technologies and urge caution, but understand that they're
inevitable and can have great benefits, and people who look
toward these technologies with great hope and optimism, but
understand that there are some dangers that need to be watched.
|
| We need to reach a wide
range of opinion leaders |
|
FI: What do you see as important goals for the next few
years?
Drexler: In the next few years, we need to reach a wide
range of opinion leaders, particularly in the scientific and
technical disciplines and some of the longer-range thinkers in
political and economic policy. We need to have these opinion
leaders exposed to the ideas of nanotechnology, assemblers, and
the rest in such a way that they come away seeing them as
credible concerns and understanding their basic implications. And
we need to develop a family of organizations that bring people
together who are concerned with these matters, so that they can
exchange ideas and work together effectively to influence the
course of events--to influence the way this technology emerges
and how it's used. Our goal is to help that process along and to
provide a way for people to get together and do these things.
FI: What can readers of this newsletter do to help this
goal?
Drexler: Readers can think about how they might be able to
help this effort, and can let us know what role they might be
able to play. And they can inform their friends and colleagues
and try to get them involved as well.
Books
of Note
The Media Lab: Inventing the Future at MIT,
Stewart Brand, Viking, 1987. Vividly describes the Lab's goals
and projects in which computers, broadcasting, and publishing are
merging to give us personalized technologies. A fun read,
accessible to laymen.
The Evolution of Cooperation, Robert Axelrod,
Basic Books, 1984. Describes an elegant computer game that showed
how cooperation can evolve among self-interested, competing
entities. Shows what conditions promote cooperation and the
importance of being "nice, retaliatory, and forgiving."
Molecular Electronic Devices II, Forrest
Carter (ed.), Marcel Dekker, 1987. Proceedings of the Second
International Workshop on Molecular Electronic Devices held in
1983 at the Naval Research Lab. Oriented toward chemistry, it
also includes one paper on nanotechnology. If the nearly $100
price daunts you, have your favorite technical library buy it.
Technologies of Freedom, Ithiel de Sola
Poole, Belknap/Harvard, 1983. A classic work on freedom of speech
and of the press in electronic media, combining history, law, and
technology. Of interest to all who look forward to hypertext
publishing as a new free press. Accessible to laymen, also
available in paperback.
Engines of Creation,
K. Eric Drexler, Anchor Press/Doubleday. Just published in
paperback.
Molecular Mechanics, ACS Monograph 177, U.
Burkert and N. Allinger, Amer. Chem. Soc., 1982. Standard
reference on molecular mechanics, useful for molecule hackers.
Induction: Processes of Inference, Learning, and
Discovery, John Holland et al., MIT Press, 1986.
Inductive reasoning and learning in both organisms and machines
are given a new theoretical interpretation by two psychologists,
a computer scientist, and a philosopher. Yale's Sternberg calls
it "the most important book on induction, and probably on
reasoning in general, that has ever been written."
Information
Available
David Forrest of the MIT NSG has prepared an information
packet, entitled "Nanotechnology Press Kit," which we
understand is available to anyone on request from the MIT News
Office, 77 Massachusetts Ave., Cambridge, MA 02139.
Evolution
in Software
Richard Dawkins's
book The Blind Watchmaker (reviewed last
issue) is due out in paperback in January; for an added
$9.95, buyers can get Blind Watchmaker software for
the Macintosh personal computer. This program simulates the
variation and selection processes of evolution. The
"organisms" are line drawings which have an
"embryology" involving the number and pattern of
branching lines, which are controlled by numerical
"genes." Random mutation produces variations in the
drawings, while users control the selection process--essentially,
picking what looks interesting. A startling range of patterns can
quickly be evolved, from trees to insects to frogs to faces. Blind
Watchmaker literally brings home the power of combining
random mutation with nonrandom selection, and it's addictive fun
too.
Letters
to FI
The Foresight Institute has received hundreds of
letters--from across the US, Canada, Japan, and
Europe--requesting information on our activities. Herewith some
excerpts:
... My primary research interest is metabolic engineering, the
rational analysis, design and construction of biochemical
networks. An eventual extension of this work is the design of
membranes and other cellular components. If you know of others
interested in this area of nanotechnology I would greatly
appreciate a listing of their names and addresses.
Prof. Douglas Cameron
University of Wisconsin
I count my reading of Engines
of Creation among the more inspiring and hopeful
intellectual experiences of my lifetime.... Thank you for writing
it. I've long followed developments in technology--primarily
electronics--and have encouraged an active futures orientation in
our small company here in San Francisco. As an independent
consultant, writer, researcher, and entrepreneur, I'm interested
in continuing to track the evolution of bioelectronics and
related fields. I'm also interested in examining and helping to
clarify the social implications of these developments.
Dean Gengle
San Francisco, CA
I am a chemistry professor at Princeton University whose
research interests are the electronic structure and reactivity of
organic, inorganic and biological systems. We are concerned with
the understanding of the mechanistic principles and structural
patterns of molecules ... have been greatly stimulate by
Drexler's book and believe that our research is taking us
precisely the direction suggested there.
Prof. Leland Allen
Princeton University
... My interest stems from a very intimate participation in
the methodology of molecular engineering which is the basis of
our company.
James Cusumano
Chairman and Chief Technical Officer
Catalytica Associates, Inc.
Engines of Creation
is such a thought-provoking and stimulating book that it took
me almost five weeks to read. Your reasoned analysis and
optimistic forecasts ... required me to "step back" and
sort out my thoughts after each reading of a few pages.... Your
combination of insight, well-organized arguments and prediction
of success for mankind continue to add energy and purpose to each
day of my life....
Henry Lederer
Wayzata, MN
Life-extension research is of great interest to me, as are the
potentials of nanotechnology for establishing an environmentally
safe form of industrial production. Drexler's book ... does an
excellent job of charting the likely road into the future with
both its opportunities and its hazards.
Patrick Milburn
Cambria, CA
... I would appreciate being placed on your mailing list and
receiving any information on [nanotechnology]. As a molecular
biologist I am very interested in the application of
biotechnology to all facets of life, but especially to computer
technology and artificial intelligence.... I would like to keep
abreast of the latest developments in nanotechnology that have
relevance to medicine.
Stephen Grund, PhD
Health Science and Technology Program
MIT
... It's surprising that more futurists have not seen
[nanotechnology] earlier, as it is so much in line with
technological trends of increasing miniaturization and fine
control over matter. I particularly liked your portrayal of the
ease and speed that will apply to future construction of space
hardware. It is refreshing to have a virtually certain proof that
space settlement need not be stopped by economics....
So far as the immediate future is concerned, I think that the
recognition of our need for better institutions is right on. Fact
forums and hypertext will be needed for dealing with the new
technology, as well as providing a competitive focus for new
memes. Building these new institutions certainly seems like a
better tactic than simply flailing away at trying to improve the
old institutions.
David Blenkinsop
Ormiston, Saskatchewan
FI
Questionnaires
Over 200 FI contributors have filled out our five-page
questionnaire. As a result, we now have a good feel for who is
interested in nanotechnology, and for their views on the FI
approach to this and other coming technologies.
FI participants come from all walks of life, from professors and
technical workers--the majority--to truck drivers and clerical
workers. Not surprisingly, many of us work in the computer field.
Almost all agree with FI's assessment of the importance of
nanotechnology, artificial intelligence, hypertext
publishing/social software, fact forums, and space development as
technologies which will drastically alter our future, but some
are skeptical regarding one or more of these ideas.
The vast majority are somewhat-to-very concerned about possible
misuse of new technologies, social and economic disruption caused
by them, and unnecessary delays in development due to fears based
on ignorance. Protecting and restoring environmental quality is a
widespread concern, regardless of political affiliation.
Most respondents are "strongly interested" in all of
the information areas we list, from technical information to
progress reports on organizational work. Naturally, nontechnical
people are less enthusiastic about the former, and some technical
people are uninterested in the latter. Most want both, and
everything we list in between, so until two separate publications
can be justified we'll combine technical and nontechnical
information in Update.
Thanks
Too many friends have helped FI in too many ways to be thanked
here, but we'd like to mention a few who've responded to our
calls for advice on two particular topics. For advising us on
Macintosh accounting software, thanks to (among others) Frederick
Reynolds, Lance Albin, and T. Toth-Fejel. For advising us on
finding an Executive Director, thanks to (among others) Ray
Alden, Pat Wagner, Brian Quig, Gayle Pergamit, Stewart Brand,
M.L. Hanson, Margaret Jordan, and Linda Arc.
From Foresight Update 2, originally
published 15 November 1987.
Foresight thanks Dave Kilbridge for converting Update 2 to html
for this web page.
|