Foresight Update 7
page 2
A publication of the Foresight Institute
 Profile:
Marvin Minsky
by Dan Shafer
Marvin
Minsky may be one of the few human beings alive today who can
say, with a straight face, that he's been interested in
nanotechnology his whole life and not produce knowing chuckles in
his listeners. Minsky, who is Toshiba Professor of Media Arts and
Sciences at Massachusetts Institute of Technology, is undeniably
an elder statesman of American science and technology. Widely
viewed as one of the fathers of Artificial Intelligence, Prof.
Minsky has been a mathematician and scientist since the
mid-1950s.
But he still says he was "always interested in
nanotechnology, even as a child. I got a chemistry book when I
was quite young. I learned a lot about organic chemistry and
almost nothing about inorganic chemistry," he says with a
deep chuckle. "I was always intrigued by the little machines
one studied in organic chemistry."
Prof. Minsky, who is a member of the Foresight
Institute Board of Advisors, is a strong supporter of the
Institute and of the work Eric
Drexler and his associates are doing today.
"Traditionally, you've had scientists doing things and other
people opposing them. Eric's attempt to be the responsible
scientist, to combine both development of an important idea and
thoughtful consideration of its consequences in one person, is
admirable. I respect his view and his uniqueness," he said
during an interview for Update. "Generally, the
dream of the responsible scientist seems to me to emerge only
accidentally, if at all."
In college in the 1940s, Prof. Minsky became interested in the
nervous system and in how brain cells are connected to one
another. He invented the first scanning microscope, for which he
still holds the basic patent. This microscope was significant in
large part because it permitted scientists to study thick
sections rather than confining research to thin sections of
specimens. He also developed a uniquely precise and fast
electromagnetic micromanipulator. In the mid-1950s, he says,
"I was hot on the track of what is now called the atomic
force microscope" as a result of his efforts to increase
resolution of the microscopic work he was doing.
He became increasingly interested in Artificial Intelligence,
however, and did some of the early work in that field, notably
the 1957 essay, "Heuristic Aspects of the Artificial
Intelligence Problem." He and John McCarthy co-founded the
MIT Artificial Intelligence Laboratory and in 1961, he published
"Steps Toward Artificial Intelligence," a work that was
to influence research directions in that field for the next 15 or
20 years.
His most recent work is a popular book, Society of Mind,
which promulgates a new concept of human psychology based on his
more than 30 years of research into intelligences of all kinds.
Prof. Minsky offered his views on a number of intriguing topics
during the wide-ranging conversation. Here is a sampling.
On nanotechnology:
The most important single thing about nanotechnology can't be
singled out. If you can actually build an assembler,
then every aspect of the physical world will be touched and
altered. The inability to do atomic rearrangement in the past has
closed so many doors.
On Artificial Intelligence and
molecular computing:
The two disciplines will complement each other in significant
ways. Nanotechnology will help us to get at the brain's
architecture, for example. Today, we use relatively large
micron-sized needles to probe and analyze the brain's structure.
We need to be able to use nano-sized needles so we can get at the
neurons and the synapses. On the other side, AI will be important
in developing better theories about what happens at the atomic
bond level. Today, we can get at the structure of a protein but
we can't understand yet how it interacts with its substrates.
Heuristic chemistry is going to become a major factor in
understanding how smaller molecular structures work. We need
expert assistance with these complex problems and AI will help by
suggesting how things work, proposing and testing new designs for
the understanding and, ultimately, the manufacture of these
little machines.
On Software:
It's hard to say if nanotechnology will have any specific
impact on software. What is clear, though, is that we must move
to more parallel software architectures and structures if we are
going to solve the very complex problems involved in, say,
simulating the behavior of even a relatively simple cell.
Parallel computers like the Connection Machine are speeding
things up about as fast as we want, so we just need to understand
better how to make them work for us. New algorithms are emerging
every couple of months from the research people are doing into
parallel computer architectures. Once some of these very
expensive processes like complex sorts become cheap, they will
help us gain new insights into how we can model the naturally
complicated processes of life. Then there's the whole issue of
how memory will work for such structures. Clearly we need to do
more work in associative memory, and such memory systems are
going to be so massively parallel that we need to get them as
small as possible to make them really usable.
On the heat dissipation problem in
molecular computing:
The whole brain only generates 40 watts. We should be able to
find a way to dissipate that little bit of heat once we
understand how to model the real thing.
|
| "The first large AIs will all
be insane in one way or another." |
|
On the possibility that people might
place too much trust in Artificial Intelligences as they become
real:
The first large AIs will all be insane in one way or another.
I don't think there's much danger of people trusting them too
much.
On Foresight Institute's
performance and mission:
The simplest thing, which it is doing quite well, is to act as
the public explainer of what nanotechnology is, what
technological developments are needed, where to put the emphasis.
If our goal is to speed the development of this technology, we
need a place for someone to go who wonders, "Is there
something in here for my company?" The Institute's approach
should also ultimately prevent crackpot and science-fiction ideas
from taking hold. This is a difficult task, particularly because
this technology has such potential to impact everything.
In the long run, it is doubtful if any one group is going to be
able to have much influence. What we can and must do in the near
term is help people focus on the important issues and provide a
forum for their discussion and clarification.
On biostasis:
It's sort of a long shot, but it's not very costly to try. The
real problem is that it's best to do the freezing just before
death, but that raises all kinds of questions--gambling
questions, let alone the ethical ones. A number of researchers
have suggested that we may not, in the long run, have to keep the
brain alive, just preserve enough of it to enable molecular
machines to reconstruct it in another environment. Some people
wonder if this reconstructed brain is really the same
"person" as the cryogenically preserved one. For me,
that's no problem. I've made up my mind that a person is a
machine. If you make a copy of the machine, the copy is as good
as the original. Would I want to be the subject of biostasis?
Well, if it's done too late, I've lost a lot of what makes me who
I am and makes me want to live on. But if we find a way to make
periodic backup tapes, I'd go for it in a minute. In some number
of decades, this subject will be a normal topic of discussion;
the question won't be whether we should preserve people
through biostasis, but when and how.
On preparing for a career involved
in nanotechnology:
That's easy. Do all the math you can until you are sick of it.
Do some computer hacking. Get involved with a research group
that's actually making something so you get your hands
into that process. Major in computer science, physics, biology,
anything but nanotechnology. Most importantly, get rid of the
idea that you have to choose a field. Change it instead to a
feeling that you want to get good at something. That way you can
expect to learn 20 or 30 important fields in your life instead of
deciding what you want to be.
On modern philosophers of science:
The things these people say, I'd give a C to a sophomore for
producing. Technical philosophers don't seem able to think about
the really important issues at all.
Dan Shafer is an author and consultant in computation and
emerging technologies.
The
Economics of Rapidly Changing Technology (Part I)
Opinion by Jeffrey C. MacGillivray
Technological change at an accelerating rate is not a new
phenomenon:
"It is an extraordinary era in which we live. It is
altogether new. The world has seen nothing like it before. I
will not pretend, nobody can pretend, to discern the end. But
everyone knows that the age is remarkable for scientific
research.... The ancients saw nothing like it. The moderns
have seen nothing like it till the present generation."
These words apply to today and, with equal validity, to the
1950's, the 1920's, the 1890's, and the 1850's. They were spoken
by Daniel Webster in 1847. The experience of 150 years of
accelerating technological change, all at previously
unprecedented rates, teaches us something about the likely
economic effects of technological change, of whatever nature and
magnitude.
The pace of technological change will continue to accelerate for
sound economic reasons. The rising standard of living resulting
from previous technological progress makes more resources
available to research, develop, and produce technology. The high
standard of living allows us to safely perform a large number of
experiments, unlike a subsistence economy which must be
conservative and tradition-bound lest a single failed experiment
result in the destruction of an entire society.
Nanotechnology, like any other rapid change in technology, will
dramatically change relative values and, as a result, lifestyles.
However, technological change will not bring an end to economic
activity. Economic exchanges will still occur.
It is extremely difficult to predict the effects of many
simultaneous technological changes; it is even difficult to
predict the effects of a single change. Larry Niven, in his story
"Flash Crowd," describes a plausible society resulting
from the introduction of teleportation--a technological change
which lowers transportation costs, both in economic value and
time, by several orders of magnitude. He describes how a serious
investor, with full knowledge of the forthcoming technological
change, could fail miserably in predicting the economic
ramifications of the new technology.
Introducing only part of nanotechnology would not completely
disrupt our current economic structure. Consider the economic
effects of one unlikely partial breakthrough in nanotechnology.
Assume that cell-repair machines exist, and can be tailored at
zero additional cost for additional people; and assume that
cell-replication machines exist, making food and clothing
virtually free. However, assume no breakthroughs in design
capabilities--no new genetic engineering, no new materials, no
nanocomputers, and no macroscopic replicators. The cost of
medical care, food, and clothing would drop drastically. However,
most other material goods would not change dramatically in value;
capital assets in unaffected industries such as housing and
transportation would still be of significant value, and economic
life might still largely resemble today's.
But nanotechnology promises much more: new materials created from
common, plentiful building blocks such as carbon atoms, and
self-replicating machines able to produce other machines that
build skyscrapers as well as cells. Thus the production cost of
non-biological material goods such as housing and transportation
will also drop dramatically. When all physical goods can be
produced at negligible cost, their economic value will become
very small; when even custom on-site material handling and
construction becomes literally dirt-cheap, many elements of
today's economic structure will disappear.
What will retain value when materials are virtually free? When
the cost of complicated constructed objects is not much more than
that of raw materials, and plentiful raw materials are as useful
as relatively rare materials? What forms of human labor will
still be of value after self-replicating molecular machines
provide material goods in virtually unlimited quantity at almost
zero cost?
The most valuable types of labor will be very different from
today's.
Production labor will not retain value; the traditional value
of labor arising from its ability to contribute to the production
of valuable material goods will disappear. When goods can be
produced at virtually zero cost with no labor, neither labor
which produces goods nor labor which further improves production
technology will be of any significant economic value.
Production-design labor will also lose its value. Once the
technology has been created, further advances are not
economically necessary and will have no significant economic
value. This will be true even without artificial intelligence
advances inspired by nanotechnological computers; but it is
extremely unlikely that there will be no further advances at all.
Only labor which contributes to entertainment--such as artistic
design labor--will be of any value. Fortunately, the desire for
such labor in any society is virtually unlimited.
In a nanotechnological society, the amount of human effort
required to acquire necessities will shrink to a minute fraction
of individuals' lives. Freedom from the necessity of acquiring
basic human needs will give individuals more control over their
lives, and more of their time will be at their own disposal.
What will people choose to do with their additional free
time?
Two extremely different subcultures in which virtually all of
the basics of life were provided externally, with little prospect
of change in the material standard of living regardless of the
efforts of the individual, illustrate the wide range of likely
human response to such freedom. One subculture is the more
scholastic of the medieval monasteries, where resources provided
from outside allowed the monks to pursue and preserve knowledge.
A very different example is today's inner city welfare culture,
where a much higher standard of living is provided from outside;
virtually all effort is devoted to entertainment, which too often
is violent or destructive.
We will have an entertainment society, not an information
society. Knowledge will be pursued only for its own sake,
not because the student expects to be able to convert the
knowledge into physical wealth. Knowledge will still be sought,
since the presence of physical libraries and electronic
databases, or even implanted access to electronic databases, will
not eliminate the desire and need for a framework of knowledge in
the individual brain. In the absence of such a framework to
provide context, the utility of facts from a database will be
limited, like the limited utility of a word-by-word foreign
language translation using a pocket dictionary.
Individuals' wants will take many forms, just as they do
today.
Self-directed people will pursue knowledge and entertainment
for their own pleasure. Some will accumulate knowledge for the
joy, satisfaction, and challenge of the pursuit. Others will take
up artistic activities. Some will preserve and continue
traditional and creative means of construction and production.
Most people will still prefer live performances of the arts, such
as ballet or concerts, over home viewing, even if five-sense
reproductions of such events could be directly wired into the
viewer's brain. There is a definite psychological value to
experiencing such performances directly, and sharing the
experience with friends. Entertainment will be a major part of
everyday consumption. Some restaurants already offer a look at
this entertainment society of the future: your food is cooked for
you at your table by your personal chef, who provides a
performance along with the food and traditional meal service.
These creative and recreational activities, so similar to the
leisure pursuits of self-directed people today, will expand to
consume virtually all of these individuals' time. Many of these
activities will also require the purchase of specialized labor.
Another group of people is driven by a need to be accepted by and
impress others. These people follow fads in the clothes they
wear, the types of food they eat and offer to others, the places
they go, and the activities they participate in. This will create
a demand for copies of certain artistic designs, for tickets to
performances of certain performers, and for close contact with
fad leaders. One contemporary example of this is the desire not
just to travel, but to travel with a celebrity.
A similar class of people wants to be noticed by others. Being
influential, well thought of, or noticed in a small group is not
enough. They want to feel important in a larger group, and are
willing to pay for the privilege, just as people have been buying
their way into the Social Register for decades.
Technology will create myriad new forms of this want. One will
increase as electronic databases allow information to be produced
even faster than today. People are already unable to handle
today's volume of information without using filters--friends, or
editors of journals--to point out interesting items. As the rate
of information production increases, people will become more
dependent on such filters to select items out of the mass of
information. The increasing importance of these filters as the
rate of information production increases will encourage people
wishing to be noticed to bribe editors.
Another group of people want to push other people around, and
will probably be willing to pay for the privilege. This desire,
merged with the desire to impress others, could lead to
monumental displays of extravagance, employing enormous numbers
of servants for short periods of time.
In other words, the economics of production will change;
human nature won't. Some believe that when people are freed
from the stresses involved in earning a living, their
personalities will change. The lack of any signs of such large
scale changes in the last 100 years leads me to believe that
future changes of this sort will be insignificant.
There will be no shortage of a demand for labor--even
unskilled labor--in such a society. The largest rewards will
go to talents different from those most rewarded in the past. We
are already seeing some signs of this change; some of the largest
monetary rewards in today's society already go to a few
performers who are either clearly the best at a particular
artistic endeavor, or who are the current entertainment sensation
of the masses.
Land will retain its value in a nanotechnological society.
Some people will want land to isolate themselves from their
neighbors. This desire will be increased by the elimination of
the traditional obstacle to living far apart--the high cost of
daily transportation to a place of production.
With the increased amount of time available for travel and the
decreased cost of transportation, scenic locations will attract
tourists in great numbers for esthetic reasons. Some scenic
locations, some historic locations, and some popular recreational
locations will also derive a special value from being current or
long-standing favorites of fad followers.
The creation of additional scenic locations by large-scale
terraforming will be technologically feasible. Terraforming is
not a new concept; new land was created during the 19th century
in the Boston area by leveling several hills to fill bays and
marshes. Future terraforming is more likely to create hills than
flatten them; however, public sentiment will likely limit the
extent of this process. Therefore, little enough terraforming
will occur that this will not significantly change the value of
land which is not terraformed.
Summarizing the likely items of value in a nanotechnological
society:
- Material goods in general will have very little value.
- Special items of artistic merit will have temporary
value.
- Land will have enduring value.
- Production labor will have very little value.
- Entertainment labor of many kinds will retain value.
(In the second
installment, the author discusses media of exchange
and emigration pressures in a nanotechnological society, relates
some current economic trends to the transition to a
higher-technology society, and examines the political economics
of a nanotechnological society.)
Dr. MacGillivray is a member of the MIT Nanotechnology Study
Group with a background in physics.
Foresight
Losses
Accidents have recently caused the deaths of two Foresight
supporters:
George Koopman
George was President of the American Rocket Company, a firm
working to launch the first privately-developed space
transportation system. Long a supporter of the ideas underlying
the Foresight Institute and other forward-looking groups such as
the L5 Society and the National Space Society, he was one of the
Institute's major contributors. George was honored in the
Congressional Record: "Like American pioneers through the
years, he would not let his dreams be denied. He was, by his
actions, a true space pioneer." George was on his way to the
site of an engine test when the auto accident occurred.
Kirk Kelley
Kirk worked at Sun Microsystems, where one of his primary
concerns was to ensure that hypertext systems developed by
different groups would be compatible and able to share data. Kirk
shared Foresight's goal of an open hypertext publishing system
and was actively working to make it a reality. He was on his way
to a hypertext meeting in Italy when the accident occurred.
From Foresight Update 7, originally
published 15 December 1989.
Foresight thanks Dave Kilbridge for converting Update 7 to
html for this web page.
|