According to Hans Moravec, by 2040 robots will become as
smart as we are. And then they'll displace us as the dominant form of life
on Earth. But he isn't worried - the robots will love us.
Hans Moravec reclines in his chair and places his palms
against his chest. "Consider the human form," he says.
"It clearly isn't designed to be a scientist. Your mental capacity is
extremely limited. You have to undergo all kinds of unnatural training to
get your brain even half suited for this kind of work - and for that
reason, it's hard work. You live just long enough to start figuring things
out before your brain starts deteriorating. And then, you die."
He leans forward, and his eyes widen with enthusiasm. "But wouldn't
it be great," he says, "if you could enhance your abilities via
artificial intelligence, and extend your lifespan, and improve on the
Since his earliest childhood, Moravec has been obsessed with artificial
life. When he was 4 years old, his father helped him use a wooden erector
set to build a model of a little man who would dance and wave his arms and
legs when a crank was turned. "It excited me," says Moravec,
"because at that moment, I saw you could assemble a few parts and end
up with something more - it could seem to have a life of its own."
At the age of 10, he constructed a toy robot from miscellaneous scrap
metal. In high school, when another student maintained that no machine
could ever be truly human, Moravec suggested replacing human neurons, one
at a time, using man-made components that would have the equivalent
function. At what point, he asked, would humanness disappear? If a wholly
artificial entity is still able to act human in every way, how could we
prove that it isn't human?
Today, Moravec is a professor at Carnegie Mellon University's Robotics
Institute, the largest robot research lab in the country and one he helped
establish in 1980. He is a rare mixture of visionary and engineer, equally
comfortable speculating on the fate of the planet or using a soldering
iron, microchips, and stepper motors to build high-tech versions of his
childhood dancing man. More than that, though, he's our most gung-ho
advocate of technology as a tool to transform human beings and make us
more than we are - within our lifetimes, if we want it.
Some of his concepts have a confrontational, in-your-face shock value. For
instance, to find out how the mind works, Moravec suggests severing a
volunteer's corpus callosum (the nerve bundle linking the two hemispheres
of the human brain) and interposing a computer to monitor thought traffic.
After the computer has had time to learn the code, it can start inserting
its own input, helping solve difficult math problems, suggesting new
ideas, even offering friendly advice.
Or here's another scenario for anyone who'd like to escape the
constrictions of dull old human biology: a futuristic robot surgeon peels
away the brain of a conscious patient, using sensors to analyze and
simulate the function of every neuron in each slice. As Moravec puts it,
"Eventually your skull is empty, and the surgeon's hand rests deep in
your brainstem. Though you haven't lost consciousness, your mind has been
removed from the brain and transferred to a machine."
But even proposals like these are modest compared with Moravec's Number
One concern, which is nothing less than the future of humanity. By 2040,
he believes, we can have robots that are as smart as we are. Eventually,
these machines will begin their own process of evolution and render us
extinct in our present form. Yet, according to Moravec, this is not
something we should fear: it's the best thing we could hope for, the
ultimate form of human transcendence. And in his own laboratory, he's
laying the groundwork that may help this evolutionary leap happen ahead of
Not everyone thinks this is such a wonderful idea. Joseph Weizenbaum,
professor emeritus of computer science at MIT, complains that Moravec's
book Mind Children: The Future of Robot and Human Intelligence is as
dangerous as Mein Kampf. Respected mathematician Roger Penrose has written
a long essay for The New York Review of Books in which he twice uses the
word "horrific" to describe some of Moravec's concepts. Book
reviewer Poovan Murugesan denounces Moravec as "a loose cannon of
fast ideas" who suffers from "irresponsible optimism."
Even Moravec's fans seem a little ambivalent. "He comes off as a
cross between Mister Rogers and Dr. Faustus," says writer Richard
Kadrey. And in the words of award-winning science fiction author Vernor
Vinge, who is also an associate professor of mathematical sciences at San
Diego State University, "Moravec puts the rest of the technological
optimists to shame. He is beyond their wildest extremes." But, Vinge
adds hastily, "I mean this as praise!"
How seriously should we take Moravec's ideas? He is widely respected as a
pioneer in robotics, but where is the line dividing his painstakingly
practical research from his unfettered speculation? Why does he insist
that breaking the boundaries of being human is important not just for
himself, but for everyone - and why does he seem so crazy-cheerful about
the whole thing?
These questions were on my mind when I visited Moravec at Carnegie Mellon
in Pittsburgh, Pennsylvania. In person, he's a friendly faced, slightly
overweight, irrepressibly good-humored man in his late 40s who wears
homely clothes and seems shy with strangers. But his enthusiasm gives him
a childlike charm - even when he talks lyrically about human extinction.
His office is next door to the "high bay," a big lab displaying
the results of previous Robotics Institute projects, including a huge,
multilegged "walker" that was sent down into the cone of an
active volcano, and a Pontiac minivan that can drive itself at speeds up
to 60 mph. The van has already found its way from Pittsburgh to
Washington, DC, with minimal human supervision, under the legal fiction
that its four onboard Sparcstations and their mechanical interface are
"an advanced form of cruise control."
But Moravec seems bored by these past achievements and has shed most of
his administrative responsibilities at the Robotics Institute. He hides
out in a small, undistinguished, modern office with a couple of computers,
a few file cabinets, a refrigerator, a microwave oven, and a lot of books.
This is where he pursues his immediate goal: designing and programming a
domestic robot that can navigate freely in cluttered home environments. It
is the next logical step, he says, toward truly intelligent machines that
we will not only tolerate but love - even as they threaten to displace us
as the dominant form of life on Earth.
Moravec's early work in robotics was plagued by setbacks. "I spent
most of the 1970s," he recalls, "trying to teach a robot to find
its way across a room. After 10 years, in 1979, I finally had one that
could get where it was going three times out of four - but it took five
hours to travel 90 feet." He chuckles like a fond father recalling
the first incompetent steps of his baby boy.
Why was it so hard for a robot to accomplish a task that even a mouse can
manage with ease? The answer, of course, is that animals have had hundreds
of millions of years in which to evolve motor skills. The problem of
moving through a three-dimensional world is hideously complex, as Moravec
indicates, while counting off the tasks on his fingers: "Our robot
used multiple images of the same scene, taken from different points of
view, in order to infer distance and construct a sparse description of its
surroundings. It used statistical methods to resolve mismatching errors.
It planned obstacle-avoidance paths. And then it had to decide how to
actually turn its motors and wheels."
In 1980, he built new robots and attempted to boost their performance.
"But the best we were able to do with our old approach," he
recounts, "was speed it up about tenfold and improve its accuracy
tenfold. We did not manage to reduce its brittleness."
By "brittleness" Moravec means that the system tended to fail
suddenly and catastrophically. "Accidental conspiracies of sensory
miscues would lead it to a wrong conclusion while being sure that it was
right. In practical terms, it could misidentify the surrounding objects
and run into a wall."
Like Wile E. Coyote in a Road Runner cartoon, trying to run into the mouth
of a tunnel painted on a rockface?
"Precisely!" he laughs again, sounding genuinely happy, as he
does whenever he describes the lovably fallible behavior of his creations.
In 1984, using US$10 Polaroid ultrasonic range finders instead of
expensive video cameras, he created a new commercial robot that analyzed
maps of the surrounding space rather than just objects in it. The result,
to his surprise, was a system that could navigate reliably and relatively
Moravec's current research robot, a project initiated in 1987, now sits in
a small workshop just across the corridor outside his office. "Would
you like to take a look?" he asks.
We walk into a windowless space no larger than an average living room.
There are a couple of video monitors, workbenches littered with tools,
pale beige walls, and a vinyl floor. The robot stands in the center of the
room: an ugly little four-wheeled truck the size of a go-cart. But Moravec
exudes pleasure and affection as he guides his toy out of the workshop,
into the hall, and back again.
Today's best robots can think at insect level," he says as we return
to his office. He explains that state-of-the-art mobile robots orient
themselves by sensing special markers placed on floors, walls, or
ceilings. Insects behave the same way: ants follow pheromone trails,
lightning bugs look for each other's flashes, and moths navigate with
reference to the moon.
The trouble is, such systems are still brittle. Just as a moth can become
fatally confused by fixing on candlelight instead of moonlight, a robot
guided by markers can easily make a disastrous mistake - as happened when
one designed by a Connecticut company to distribute hospital linens took a
nosedive down a flight of stairs when it failed to notice a marker that
was supposed to tell it not to proceed past a certain point.
Robots that orient themselves with markers have found some application in
industry - transporting pallets and cleaning floors - but they offer few
advantages over the older systems that follow hidden guide wires. As a
result, the market is very limited. "In fact," says Moravec,
"the market barely exists at all. So, what we're shooting at now is a
robot with the intelligence of a small vertebrate - the smallest fish you
can imagine. It will no longer depend on navigational points; it will
build a relatively dense representation of volumes of space."
By 2000, he foresees that this type of machine will find its own way
around complex, cluttered places without using markers and without needing
to be installed by experts. At first these robots will be expensive and
specialized, but Moravec predicts they will become smaller, cheaper, and
more user-friendly in just the same way that microcomputers evolved from
mainframes. "Once we have a robot that customers can take out of the
box, show it a job, and trust it to work without doing silly things - then
the market will grow easily to hundreds of thousands and beyond. Any
institution that does regular cleaning will find that it's cheaper to use
a robot than a person. The same goes for delivery jobs."
Moravec estimates that these systems will need an onboard computer capable
of 500 million instructions per second. The first IBM PCs managed 0.3 mips;
a modern Pentium-based PC reaches 200 mips; and it's reasonable to expect
that 500-mips processors will be affordable by the turn of the century.
This power will enable the robot to convert 500-by-500-pixel stereoscopic
pictures from its camera eyes into a 3-D model consisting of about
100-by-100-by-100 cells. Updating and processing all this visual
information will take about one second - the longest interval that is
reasonably safe and practical, since the robot will move blindly between
glimpses of the world.
Once robots find a niche doing dull, repetitive jobs, Moravec sees an
ever-expanding market. "The next step will be adding an arm and
improving the sensor resolution so that they can find and manipulate
objects. The result will be a first generation of universal robots, around
2010, with enough general competence to do relatively intricate mechanical
tasks such as automotive repair, bathroom cleaning, or factory assembly
By "universal" Moravec means the robot will tackle many
different jobs in the same way a Nintendo system plays many different
games. Plug in one cartridge, and the robot will know how to change the
oil in your car. Plug in another, and it will know how to patrol your
property and challenge intruders.
Add more memory and computing power and enhance the software, and by 2020
we have a second generation that can learn from its own performance.
"It will tackle tasks in various ways," says Moravec, "keep
statistics on how well each alternative has succeeded, and choose the
approach that worked best. This means that it can learn and adapt. Success
or failure will be defined by separate programs that will monitor the
robot's actions and generate internal punishment and reward signals, which
will actually shape its character - what it likes to do and what it
prefers not to do."
Moravec pauses. The near future of robotics is something he's spelled out
a thousand times before, and he no longer finds it particularly exciting.
But now we get to a subject that interests him more: the idea that robots
can mimic human traits.
By 2030, according to Moravec, we should have a third-generation universal
robot that emulates higher-level thought processes such as planning and
foresight. "It will maintain an internal model not only of its own
past actions, but of the outside world," he explains. "This
means it can run different simulations of how it plans to tackle a task,
see how well each one works out, and compare them with what it's done
before." An onlooker will have the eerie sense that it's imagining
different solutions to a problem, developing its own ideas.
But perfecting the model of reality this robot will need is not going to
be an easy task. In fact, creating this model is the single hardest
problem in artificial intelligence. Intuitively, human beings know why
they need to wear a raincoat in wet weather, or why they must turn the
handle before pushing open a door. Almost without thinking we know if a
bottle is empty, whether an object is breakable, or when food has spoiled.
But to an artificial intelligence, none of these things is obvious - each
everyday fact must be established in advance or derived from logical
On the plus side, each time a robot learns a fact or masters a skill, it
will be able to pass its knowledge to other robots as quickly and easily
as sending a program over the Net. This way, the task of understanding the
world can be divided among thousands or millions of robot minds. As a
result, the machines will soon develop a deeper knowledge base than any
single person can hope to possess. Within a short space of time, robots
that are linked in this way will no longer need our help to show them how
to do anything.
Meanwhile, they will be smart enough to interact with us on a human level.
"Their world model will include psychological attributes,"
Moravec says, "which means, for instance, that a robot will express
in its internal language a logical statement such as 'I must be careful
with this item, because it is valuable to my owner, and if I break it, my
owner will be angry.' This means that if the robot's internal processes
are translated into human terms, you will hear a description of
consciousness - especially if the robot applies psychological attributes
to its own actions, as in 'I don't like to bump into things,' which is a
compact way of saying that the robot gets an internal negative
reinforcement signal whenever it collides with something, or imagines a
Moravec's critics are skeptical on this point. Many have stated flat out
that a machine can never be "conscious." Their arguments are
hard to refute, partly because no one can really say what consciousness
is; but Moravec sidesteps the issue. He believes a robot that understands
human behavior can be programmed to act as if it is conscious, and can
also claim to be conscious. If it says it's conscious, and it seems
conscious, how can we prove that it isn't conscious?
Either way, there's no doubt that systems that can analyze their world,
deduce generalizations, and modify their behavior will have a major impact
"The robots will still be in our thrall," Moravec points out,
meaning that we will still be designing and programming them to serve and
obey us. "They'll learn everything they know from us, and their goals
and their methods will be imitations of ours. But as they become more
competent, efficiency and productivity will keep going up, and the amount
of work for humans will keep going down. By around 2040, there will be no
job that people can do better than robots."
He sits back in his chair, pausing with cheerful satisfaction as he does
whenever he reaches a radical conclusion that places him one step ahead,
waiting for his audience to catch up.
In this case, though, Moravec's conclusion is less radical than it seems -
because when many jobs are broken down into tasks, they require a
relatively limited degree of "humanness." Even today, we have
expert systems that offer advice based on a large number of facts in a
field such as medicine or geology. Imagine this expertise gradually
broadening to include subjects such as corporate law, mechanical design,
profitability, and efficiency. Decisions in these areas are all made
logically from sets of facts, which means that if the facts are completely
spelled out, a machine intelligence should be able to deal with them.
Thus a corporation can literally become automated from the bottom up:
first the assembly lines, then bookkeeping, product design, and planning.
Even management can be taken over by computers that are able to learn from
past performance. Ultimately, a corporation will consist of a diverse mix
of robots, some mobile, some fixed, some large and powerful, some
microscopic, all interacting with speed and versatility that is completely
beyond human abilities.
But what about the time scale? Isn't he compressing a huge amount of
progress into a very few decades?
"Back in the 1970s I made some overoptimistic assumptions about the
rate of progress of computers. I thought that using an array of cheap
microcomputers, we might achieve human equivalence by the mid-1980s. Then
I did a slightly more careful calculation around 1978 and decided it would
take another 20 years, requiring a supercomputer. But then I started
getting serious, writing articles and essays, and I thought I should do
the calculations more rigorously. So I collected 100 data points of
previous computer progress, I did the best calculation I could, I compared
the human retina with computer vision applications, and I plotted it all
Still, even if his predictions are confirmed to be on schedule, there's an
obvious problem: When robots are doing all the work, no one will earn any
money. How can an economy flourish when all the consumers are penniless?
Moravec obviously isn't troubled by the question. In fact, it's hard to
imagine any question bothering him: he sits calmly, comfortably, eating
the questions and spitting out answers with ease. Today, he points out,
people who retire are supported via wealth that is ultimately created by
industry. As industry becomes more efficient, there will be more wealth,
allowing people to retire earlier. When industry is totally automated and
hyper-efficient, it will create so much wealth that retirement can begin
at birth. "We'll levy a tax on corporations," Moravec says,
"and distribute the money to everyone as lifetime social-security
But what if the robot-run corporations fail to function as he expects? He
assumes these business entities will follow programs written by us,
compelling them to obey laws and pay their taxes. But the programming will
also encourage robot-controlled corporations to compete with each other.
Won't they try to exploit loopholes in their instructions, just as
present-day businesses try to evade federal regulations? Isn't there a
real risk that autonomous robots will steal from each other and cheat on
"There is always the possibility that some kind of malfunction will
produce a rogue corporation," Moravec admits. "We'll need police
provisions so that legal companies will act to suppress rogues
economically, or physically, if necessary. And among the inprogrammed laws
we'll need antitrust clauses to force dangerously large companies to
divest into smaller entities."
But this would be a second set of rules to solve a problem created by
robots breaking the first set of rules. The system still seems
"It is unstable," he agrees. "Everything will depend on the
way in which we create it. Crafting these machines and the corporate laws
that control them is going to be the most important thing humanity ever
does. You know, each age has an activity in which the best minds get
involved. Crafting the laws, and their implementation, will be the thing
to do in the 21st century."
If the job is done right, he predicts a world of comfort, health, and
boundless plenty - at least for a while. Human beings will be like slave
owners whose servants never complain, need no supervision, and are
constantly eager to please.
In the long term, though, robots programmed to serve us with maximum
efficiency can become a potential hazard. They will naturally try to
obtain energy and raw materials as cheaply as possible, with a minimum of
regulatory interference. And the ideal way to do this is by relocating
some of their operations beyond planet Earth.
Unlike human beings, robots don't need to breathe air, aren't disoriented
by zero gravity, and can be easily shielded from harmful radiation. There
are vast mineral resources in the asteroid belt, where there will be no
regulations regarding pollution, noise, or safety. Robot factories located
in space would be able to manufacture products with maximum efficiency and
then drop them down into Earth's gravity well. Alternatively, they could
conduct hazardous research and radio the encrypted results back to their
parent corporation on Earth.
Only a small "seed colony" of robots would be needed to set up
an off-world operation. Using local mineral ores and solar energy, robots
could build everything they required - including copies of themselves.
In this scenario, everything is still being controlled by the parent
corporations, which are still being controlled by us. Therefore, the
off-world operations should present no problems. "But now suppose a
company goes out of business," Moravec says, "leaving its
research division in space, where there's no supervision. The result is
self-sustaining, superintelligent wildlife."
His critics, of course, disagree. They complain that his vision is
inhuman, lacking attributes such as culture and art that seem central to
our identity. Skeptics also point out that the negative implications of
his work far outweigh its benefits in the near future, when robots will
cause a huge economic dislocation, creating a feeling of purposelessness
among citizens who are rendered permanently unemployable.
Moravec is quite aware of this but sees no way to prevent it. He says his
projection of the future is at least 50 percent probable, and we're seeing
the first signs of it right now. "In Europe generally," he says,
"I believe unemployment is now up to around 15 percent, and
essentially this will never reverse. We're already moving into the mode I
envisage, where everyone is subsidized by productive machines."
This has created uncertainty and discontent - as he readily admits.
"We all agree," he says, "that the world is a bit screwed
up. The reason for this is rather obvious. We have a Stone Age brain, but
we don't live in the Stone Age anymore. We were fitted by evolution to
live in tribal villages of up to 200 relatives and friends, finding and
hunting our food. We now live in cities of millions of strangers,
supporting ourselves with unnatural tasks we have to be trained to
accomplish, like animals who have been forced to learn circus
In which case, what's the answer? Moravec adamantly believes that
reversing the evolution of technology would create an even bigger
disaster. "Most of us would starve," he says. He suggests the
opposite approach: that we try to catch up with technology by accelerating
our own evolution. "We can change ourselves," he says, "and
we can also build new children who are properly suited for the new
conditions. Robot children."
Inevitably, I ask whether he has any normal, flesh-and-blood children.
"No. In fact, I am biologically incapable of it. I contracted
testicular cancer as I was finishing my PhD; it didn't affect me very
much, it didn't really hurt, I noticed a growth, but I still had my thesis
to write and my orals to do, and the whole thing seemed very unreal. There
were two surgeries, one minor, one major - with my intestines out in a bag
to get at the lymph nodes. I came through it in sparkling condition, aged
around 30. But a side effect is that I'm basically infertile."
Does this mean that his love of robots is nothing more than a displaced
desire for the biological children he can't have?
"Not at all. Long before the cancer, I was already obsessively
committed to robots for whatever neurotic reason. That was where I wanted
to spend my energy. I met my wife in the hospital when I was getting
chemotherapy in 1980. She already had two children, so I inherited them as
Does his wife share any of his feelings about machines?
He laughs. "At the moment, my wife is a biblical scholar."
Moravec himself was raised Catholic, but he rebelled against it as a
teenager and says he still has some anti-Catholic reflexes. As a result,
he and his wife had some bitter theological debates in the past. "But
these days there's no point in arguing," he says, "because we
already know exactly what each other is going to say, and in any case
she's more astute in human relations than I am, so she knows how to handle
me. But I have changed my outlook slightly. I'm a little less hard-core in
my atheism than I used to be. And my ideas about resurrection in some ways
are not so different from those of early theologians, or from the Greek
thought that fed into that."
Also, of course, the desire for human transcendence has been a fundamental
feature of almost all religions. And Moravec's vision of a supremely
powerful artificial intelligence that will love humanity enough to
re-create it is basically a vision of a god - the only difference being
that in his scheme of things, we create god version 1.0, after which it
builds its own enhancements.
But how does all this fit in with Moravec's obvious personal love for
"My father was an engineer in Czechoslovakia and had a business
making and selling electrical goods during the war. When the Russians
arrived in 1944, he became a refugee. He left the country on a tricycle
with 50 kilos of tools and 50 kilos of food. He met my mother in Austria,
which is where I was born. He had an electrical store, where he'd hand
wind transformers to convert battery-operated radios so they'd run on
house current. We relocated to Canada in 1953."
This marks the point where the genie finally gets out of the bottle and
Earth's retirement community of pampered humans finds itself faced with a
big problem. Out in space, the preprogrammed drive to compete and be
efficient will result in the runaway evolution of machine capabilities.
Moravec feels that in a short period of time, all the local materials will
be plundered and converted into machines, and all available solar energy
will be used to power them.
The result will be a dense, interacting swarm of competing entities -
although, he says, the competition will be relatively benign. Warfare
among robots will be rare because "fighting wastes energy, and a
third entity can eat the pieces."
He believes that the most useful skill will be intelligence. Robots will
be motivated to make themselves as small as possible, conserving raw
materials to build better brains. "As a result, you end up with the
whole mess forming a cyberspace where entities try to outsmart each other
by causing their way of thinking to be more pervasive. Here's an ecology
where all the dead-matter activity has been squeezed out and almost
everything that happens is meaningful. You have this sphere of cyberspace
with a robot shell, expanding outward toward Earth."
What will it look like?
"It will look like a region of space glowing warmly, with hardly
anything visible on a human scale. The competitive pressure toward
miniaturization will result in activity on the subatomic level. They'll
transform matter in some way; it will no longer be matter as we know
Since space-based machine intelligences will be free to develop at their
own pace, they will quickly outstrip their cousins on Earth and eventually
will be tempted to use the planet for their own purposes. "I don't
think humanity will last long under these conditions," Moravec says.
But, ever the optimist, he believes that "the takeover will be swift
Why? Because machine intelligence will be so far advanced, so
incomprehensible to human beings, that we literally won't know what hit
us. Moravec foresees a kind of happy ending, though, because the
cyberspace entities should find human activity interesting from a
We will be remembered as their ancestors, the creators who enabled them to
As Moravec puts it, "We are their past, and they will be interested
in us for the same reason that today we are interested in the origins of
our own life on Earth."
He seems very sincere as he says this, almost as if it's an article of
faith for him - though of course it has some logical foundation. Machine
intelligences of the far future will develop from our initial programming,
just as a child grows from its parents' DNA. Consequently, even when
robots are smarter than we are, they should retain many of our priorities
But Moravec takes the scenario even one step further. Assuming the
artificial intelligences now have truly overwhelming processing power,
they should be able to reconstruct human society in every detail by
tracing atomic events backward in time. "It will cost them very
little to preserve us this way," he points out. "They will, in
fact, be able to re-create a model of our entire civilization, with
everything and everyone in it, down to the atomic level, simulating our
atoms with machinery that's vastly subatomic. Also," he says with
amusement, "they'll be able to use data compression to remove the
redundant stuff that isn't important."
But by this logic, our current "reality" could be nothing more
than a simulation produced by information entities.
"Of course." Moravec shrugs and waves his hand as if the idea is
too obvious. "In fact, the robots will re-create us any number of
times, whereas the original version of our world exists, at most, only
once. Therefore, statistically speaking, it's much more likely we're
living in a vast simulation than in the original version. To me, the whole
concept of reality is rather absurd. But while you're inside the scenario,
you can't help but play by the rules. So we might as well pretend this is
real - even though the chance things are as they seem is essentially
And so, according to Hans Moravec, the human race is almost certainly
extinct, while the world around us is just an advanced version of SimCity.
I've been sitting opposite Moravec in his office, typing on my laptop
computer, following his exposition step by step. The vision he has
described exists for him as a unified whole; it takes him only about an
hour to describe it clearly and fluently from beginning to end. For him it
seems entirely pleasurable: a destiny that grows out of his own work and
affirms his own values.
Growing up in Montreal, learning English and adjusting to a strange new
culture, Hans Moravec was a solitary child who found solace in building
models and gadgets. "I remember the thrill I got when I put together
something and made it work. I could admire it for hours. And these things
also made other people proud of me. I guess I actually thought that they
would get me a wife! I knew I didn't have any social skills, but maybe if
I could build these machine things really, really well, it would make me
more attractive to women." He laughs at his own childhood na´vetÚ.
And yet, he didn't always want to be a scientist. First he wanted to be
Superman. "But I could see that it wasn't practical. Then I noticed
another character in the comics, Lex Luthor, who didn't have superpowers
but was almost a match for Superman. So, I thought if I couldn't be
Superman, maybe I could be Lex Luthor."
In person, Moravec seems diffident and gentle; he doesn't drive a car
because, he says, he's uneasy with so much potentially dangerous mass in
his control. He likes living in Pittsburgh because his home is a short
walk from his office, and he seems to feel little need to venture outside
this simple life.
Yet as a child he enjoyed fantasies about superheroes and supervillains,
and as an adult he talks casually of totally rebuilding human society. He
refers to his new book, for which he's currently seeking a publisher, as
"a kind of speculative long-term business plan for humanity,"
and in it he speaks condescendingly of "Earth's small-minded
biological natives." Can Moravec really claim that his work as a
scientist is in no way manipulative?
"People such as myself," he says, "may have a little bit of
influence, but we're like mosquitoes pushing at a rolling boulder.
Progress is inflicted on people in the same way that natural evolution is
inflicted on people. It really is evolution; it's the selection and growth
of information, transmitted from one generation to the next."
But what about the rights of people who don't love the rolling boulder of
"Well," he says, beginning to sound a little impatient with my
objections, "they'll - they'll get used to it! In fact, they should
enjoy it, since the amount of wealth will be astronomical; you'll be able
to live anywhere and in any way you want."
In any case, he says, the progress he's talking about will be offered via
the free market, not physically imposed on anyone. "All I'm
suggesting is that we give people a choice. In the next decade, people
will either buy their housecleaning robot or not buy it. And I think
they'll want to buy it. Then they'll have the choice of upgrading to one
that learns, and I think they'll want that, too. Then they'll have the
choice of a robot that claims it's conscious, a really nice entity that
talks like a person, seems to understand you, and has nothing but your
best interests at heart - because that's how it's programmed. And then the
fourth generation will take that personality and add intelligence. It will
be a constant help to you; it will explain why something that you want to
do isn't what you should do - because it loves you. I think people will
like these machines and will quickly get used to them."
Well, yes - until the machines cut loose, develop hyperintelligence, and
bring about our demise.
"But I don't consider it a demise," Moravec retorts, still
insisting that his vision is wholly positive. "The robots will be a
continuation of us, and they won't mean our extinction any more than a new
generation of children spells the extinction of the previous generation of
adults. In any case, in the long term, the robots are much more likely to
resurrect us than our biological children are."
For people who find long-term resurrection a somewhat nebulous concept,
there are also some practical reasons why we should be happy to change
ourselves radically. On a long-term basis, Moravec points out, our planet
may not be a hospitable place to live. Huge climatic shifts may occur (as
they did during the ice ages). Our sun may become unstable. The world may
be ravaged by incurable diseases. Our entire ecology could be destroyed by
a large meteor or comet. "Sooner or later," he says,
"something big will come along that we cannot deal with. But by
changing ourselves in the most fundamental way, we will be able to survive
This is an arguable point of view, but I can't help wondering which came
first, Moravec's personal interest in becoming more than human, or his
proof that it's really a very good idea. He readily admits that he has a
personal obsession with robots, and his passion for transcendence is far
more extreme than that of most scientists. What makes him so different
from everyone else?
"Well, I was breast-fed as a baby," he answers with typically
disconcerting candor. "I was also the first born of my family, and I
was well loved by my mother - which must have helped me feel confident
about life." He pauses, realizing that this explanation isn't
adequate. "Maybe the idea of human transcendence makes me happy
because my endorphin levels were misadjusted early on in life," he
says with a laugh and a shrug, unable to come up with a better answer.
Personally, I suspect he likes the idea of radical change because he's an
intensely intelligent man who is easily bored by the everyday world. He
finds it impossible to believe that it makes sense to continue, as human
beings, in our exact same form. "Do we really want more of what we
have now?" he asks, sounding incredulous. "More millennia of the
same old human soap opera? Surely we have played out most of the
interesting scenarios already in terms of human relationships in a trivial
framework. What I'm talking about transcends all that. There'll be far
more interesting stories. And what is life but a set of stories?"
Ultimately, Moravec comes back again to the power and grandeur of a
destiny that exceeds all limits. "This universe is so big," he
says. "The possibilities must be infinitely greater than anything we
can imagine for ourselves. Pushing things in the direction of expanded
possibilities seems to be by far the most productive use of my time. And
that, here, is my purpose."