Hans Moravec
Robotics Institute
Carnegie Mellon University
Pittsburgh, PA 15213
May 1992
Exploration and colonization of the universe awaits, but earth-adapted
biological humans are ill-equipped to respond to the
challenge. Machines have gone farther and seen more, limited though
they presently are by insect-like behavioral inflexibility. As they
become smarter over the coming decades, space will be theirs.
Organizations of robots of ever increasing intelligence and sensory
and motor ability will expand and transform what they occupy,
working with matter, space and time. As they grow, a smaller and
smaller fraction of their territory will be undeveloped frontier.
Competitive success will depend more and more on using already
available matter and space in ever more refined and useful forms.
The process, analogous to the miniaturization that makes today's
computers a trillion times more powerful than the mechanical
calculators of the past, will gradually transform all activity from
grossly physical homesteading of raw nature, to minimum-energy
quantum transactions of computation. The final frontier will be
urbanized, ultimately into an arena where every bit of activity is
a meaningful computation: the inhabited portion of the universe
will transformed into a cyberspace.
Because it will use resources more efficiently, a maturecyberspace
of the distant future will be effectively much bigger than the
present physical universe. While only an infinitesimal fraction of
existing matter and space is doing interesting work, in a well
developed cyberspace every bit will be part of a
relevantcomputation or storing a useful datum. Over time, more
compactand faster ways of using space and matter will be invented,
and used to restructure the cyberspace, effectively increasing the
amount of computational spacetime per unit of physical spacetime.
Computational speedups will affect the subjective experience of
entities in the cyberspace in a paradoxical way. At first glimpse,
there is no subjective effect, because everything, inside and
outside the individual, speeds up equally. But, more subtly,
speedup produces an expansion of the cyber universe, because, as
thought accelerates, more subjective time passes during the fixed
(probably lightspeed) physical transit time of a message between a
given pair of locations -- so those fixed locations seem to grow
fartherapart. Also, as information storage is made continually more
efficient through both denser utilization of matter and more
efficientencodings, there will be increasingly more cyber-stuff
between anytwo points. The effect may somewhat resemble the
continuous-creationprocess in the old steady-state theory of the
physical universe ofHoyle, Bondi and Gold, where hydrogen atoms
appear just fast enoughthroughout the expanding cosmos to maintain
a constant density.
A quantum-mechanical entropy calculation by Bekensteinsuggests that
the ultimate amount of information that can be stored given the
mass and volume of a hydrogen atom is about a megabyte. But let's
be conservative, and imagine that at some point in the future only
"conventional" physics is in play, but every few atoms stores a
useful bit. There are about 10^56 atoms in the solar system.I
estimate that a human brain-equivalent can be encoded in less than
10^15 bits. If a body and surrounding environment takes a thousand
times more storage in addition, a human, with immediate
environment, might consume 10^18 bits. An AI with
equivalentintelligence could probably get by with less, since it
does without the body-simulation "life support" needed to keep a
body-oriented human mind sane. So a city of a million human-scale
inhabitants might be efficiently stored in 10^24 bits. If the atoms
of the solar system were cleverly rearranged so every 100 could
represent a bit, then a single solar system could hold 10^30 cities
-- far more than the number (10^22) of stars in the visible
universe! Multiply that by 10^11 stars in a galaxy, and one gets
10^41 cities per galaxy. The visible universe, with 10^11 galaxies,
would then have room for 10^51 cities -- except that by the time
intelligence has expanded that far, more efficient ways of using
spacetime and encodingdata would surely have been discovered,
increasing the numbermuch further.
Mind without Body?
Start with the concepts of telepresence and virtual reality. You
wear a harness that, with optical, acoustical, mechanical and
chemical devices controls all that you sense, and measures all of
your actions. Its machinery presents pictures to your eyes, sounds
to your ears, pressures and temperatures to your skin, forces to
your muscles and even smells and tastes for the remaining senses.
Telepresence results when the inputs and outputs of this harness
connect to a distant machine that looks like a humanoidrobot. The
images from the robot's two camera eyes appear on your "eyeglass"
viewscreens, and you hear through its ears, feel through its skin
and smell through its chemical sensors. When you move your head or
body, the robot moves in exact synchrony.When you reach for an
object seen in the viewscreens, the robot reaches for the object,
and when it makes contact, your musclesand skin feel the resulting
weight, shape, texture and temperature.For most practical purposes
you inhabit the robot's body -- your senseof consciousness has
migrated to the robot's location, in a true"out of body"
experience.
Virtual reality retains the harness, but replaces the remote robot
with a computer simulation of a body and its surroundings. When
connected to a virtual reality, the location you seem to inhabit
does not exist in the usual physical sense, rather you are in a
kind of computer-generated dream. If the computer has access to
data from the outside world, the simulation may contain some "real"
items, for instance representations of other people connectedvia
their own harnesses, or even views of the outside world,
perhapsthrough simulated windows.
One might imagine a hybrid system where a virtual "central station"
is surrounded by portals that open on to views of multiplereal
locations. While in the station one inhabits a simulated body, but
when one steps through a portal, the harness link is seamlessly
switched from the simulation to a telepresence robot waiting at
that location.
The technical challenges limit the availability, "fidelity"and
affordability of telepresence and virtual reality systems today --
in fact, they exist only in a few highly
experimentaldemonstrations. But progress is being made, and its
possible toanticipate a time, a few decades hence, when people
spend more time in remote and virtual realities than in their
immediate surroundings, just as today most of us spend more time in
artificial indoorsurroundings than in the great outdoors. The
remote bodies we will inhabit can be stronger, faster and have
better senses than our "home" body. In fact, as our home body ages
and weakens, we might compensate by turning up some kind of "volume
control." Eventually, we might wish to bypass our atrophied muscles
and dimmed senses altogether, if neurobiology learns enough to
connect our sensory and motor nerves directly to electronic
interfaces. Then all the harness hardware could be discarded as
obsolete, along with our sense organs and muscles, and indeed most
of our body. There would be no "home" experiences to return to, but
our remote and virtual existences would be better than ever.
The picture is that we are now is a "brain in a vat,"sustained by
life-support machinery, and connected by wonderfulelectronic links,
at will, to a series of "rented" artificial bodies at remote
locations, or to simulated bodies in artificial realities.But the
brain is a biological machine not designed to functionforever, even
in an optimal physical environment. As it beginsto malfunction,
might we not choose to use the same advanced neurological
electronics that make possible our links to theexternal world, to
replace the gray matter as it begins to fail? Bit by bit our brain
is replaced by electronic equivalents, which work at least as well,
leaving our personality and thoughts clearerthan ever. Eventually
everything has been replaced by manufacturedparts. No physical
vestige of our original body or brain remains, butour thoughts and
awareness continue. We will call this process, and other approaches
with the same end result, the downloading of a human mind into a
machine. After downloading, our personality is a pattern impressed
on electronic hardware, and we may then find ways to move our minds
to other similar hardware, just as a computerprogram and its data
can be copied from processor to processor.So not only can our sense
of awareness shift from place to place at the speed of
communication, but the very components of our minds may ride on the
same data channels. We might find ourselves distributed over many
locations, one piece of our mind here, another piece there, and our
sense of awareness at yet another place. Time becomes more flexible
-- when our mind resides in very fast hardware, one second of real
time may provide a subjectiveyear of thinking time, while a
thousand years of real time spent on a passive storage medium may
seem like no time at all. Can we then consider ourselves to be a
mind without a body? Not quite.
A human totally deprived of bodily senses does not do well. After
twelve hours in a sensory deprivation tank (where one floats in a
body-temperature saline solution that produces almost no skin
sensation, in total darkness and silence, with taste and smell and
the sensations of breathing minimized) a subject will begin to
hallucinate, as the mind, somewhat like a television tuned to a
nonexistent channel, turns up the amplification, desperately
lookingfor a signal, becoming ever less discriminating in the
theories itoffers to make sense of the random sensory hiss it
receives. Eventhe most extreme telepresence and virtual reality
scenarios we havepresented avoid complete bodylessness by always
providing the mindwith a consistent sensory (and motor) image,
obtained from an actualremote robot body, or from a computer
simulation. In those scenarios,a person may sometimes exist without
a physical body, but neverwithout the illusion of having one.
But in our computers there are already many entities thatresemble
truly bodiless minds. A typical computer chess program knows
nothing about physical chess pieces or chessboards, or about the
staring eyes of its opponent or the bright lights of a
tournament.Nor does it work with an internal simulation of those
physicalattributes. It reasons instead with a very efficient and
compact mathematical representation of chess positions and moves.
For the benefit of human players this internal representation is
sometimes translated to a recognizable graphic on a computer
screen, but such images mean nothing to the program that actually
chooses the chess moves. For all practical purposes, the chess
program's thoughts and sensations -- its consciousness -- is pure
chess, with no taint of the physical, or any other, world. Much
more than a human mind with a simulated body stored in a computer,
a chess program is a mind without a body.
So now, imagine a future world where programs that do chess,
mathematics, physics, engineering, art, business or whatever, have
grown up to become at least as clever as the human mind. Imagine
also the most of the inhabited universe has been converted to a
computer network -- a cyberspace -- where such programs live, side
by side with downloaded human minds and accompanying simulated
human bodies. Suppose that all these entities make their living in
something of a free market way, trading the products of their labor
for the essentials of life -- in this world memory space and
computing cycles. Some entities do the equivalent of manual work,
converting undeveloped parts of the universe into cyberspace, or
improving the performance of existing patches, thus creating new
wealth. Others work on physics or engineering problems whose
solutions give the developers new and better ways to
constructcomputing capacity. Some create programs that can become
part of one's mental capacity. They trade their discoveries and
inventionsfor more working space and time. There are entities that
specializeas agents, collecting commissions in return for
locatingopportunities and negotiating deals for their clients.
Others act as banks, storing and redistributing resources, buying
and selling computing space, time and information. Some we might
class as artists, creating structures that don't obviously result
in physical resources, but which, for idiosyncratic reasons, are
deemed valuable by some customers, and are traded at prices that
fluctuate for subjective reasons. Some entities in the cyberworld
will fail to produce enough value to support their requirements for
existence -- these eventually shrink and disappear, or merge with
other ventures. Others will succeed and grow. The closest present
day parallel is the growth, evolution, fragmentation and
consolidationof corporations, whose options are shaped primarily by
their economicperformance.
A human would likely fare poorly in such a cyberspace. Unlike the
streamlined artificial intelligences that zip about, making
discoveries and deals, reconfiguring themselves to
efficientlyhandle the data that constitutes their interactions, a
human mind would lumber about in a massively inappropriate body
simulation, analogous to someone in a deep diving suit plodding
along among a troupe of acrobatic dolphins. Every interaction with
the data world would first have to be analogized as
somerecognizable quasi-physical entity: other programs might be
presented as animals, plants or demons, data items as books or
treasure chests, accounting entries as coins or gold. Maintaining
such fictions increases the cost of doing business, as does
operating the mind machinery that reduces the physical simulations
into mentalabstractions in the downloaded human mind. Though a few
humansmay find a niche exploiting their baroque construction to
produce human-flavored art, more may feel a great economicincentive
to streamline their interface to the cyberspace.
The streamlining could begin with the elimination of the body-
simulation along with the portions of the downloaded mind dedicated
to interpreting sense-data. These would be and replaced with
simpler integrated programs that produced approximatelythe same net
effect in one's consciousness. One would still view the cyber world
in terms of location, color, smell, faces, and so on, but only
those details we actually notice would be represented. We would
still be at a disadvantage compared with the true artificial
intelligences, who interact with the cyberspacein ways optimized
for their tasks. We might then be tempted to replace some of our
innermost mental processes with more cyberspace-appropriate
programs purchased from the AIs, and so, bit by bit, transform
ourselves into something much like them. Ultimately our thinking
procedures could be totally liberated from any traces of our
original body, indeed of any body. But thebodiless mind that
results, wonderful though it may be in its clarity of thought and
breadth of understanding, could in no sense beconsidered any longer
human.
So, one way or another, the immensities of cyberspace will be
teeming with very unhuman disembodied superminds, engaged in
affairsof the future that are to human concerns as ours are to
those of bacteria. But, once in a long while, humans do think of
bacteria, even particular individual bacteria seen in particular
microscopes. Similarly, a cyberbeing may occasionally bring to mind
a human event of the distant past. If a sufficiently powerfulmind
makes a sufficiently large effort, such recall could occur with
great detail -- call it high fidelity. With enough fidelity,the
situation of a remembered person, along with all the minutiae of
her body, her thoughts, and feelings would be perfectlyrecreated in
a kind of mental simulation: a cyberspace within a cyberspace where
the person would be as alive as anywhere. Sometimes the recall
might be historically accurate, in other circumstances it could be
artistically enhanced: it depends on the purposes of the cybermind.
An evolving cyberspace becomes effectively ever more capacious and
long lasting, and so can supportever more minds of ever greater
power. If these minds spend only an infinitesimal fraction of their
energy contemplating the human past, their sheer power should
ensure that eventually our entire history is replayed many times in
many places, and in many variations. The very moment we are now
experiencing may actually be (almost certainly is) such a
distributed mental event, and most likely is a complete fabrication
that never happened physically. Alas, there is no way to sort it
out from our perspective: we can only wallow in the scenery.
