ENGINES of CREATION | Chapter 6: The World Beyond Earth

February 21, 2001
K. Eric Drexler

That inverted Bowl we call The Sky; Whereunder crawling coop’d we live and die.

The Rubáiyát of Omar Khayyam

THE EARTH is but a small part of the world, and the rest of the world will be important to our future. In terms of energy, materials, and room for growth, space is almost everything. In the past, successes in space have regularly fulfilled engineering projections. In the future, an open space frontier will widen the human world. Advances in AI and nanotechnology will play a crucial role.  

People took ages to recognize space as a frontier. Our ancestors once saw the night sky as a black dome with tiny sparks, a light show of the gods. They couldn’t imagine space travel, because they didn’t even know that outer space existed.  

We now know that space exists, but few people yet understand its value. This is hardly surprising. Our minds and cultures have evolved on this planet, and we have just begun to digest the idea of a frontier beyond the sky.

 Only in this century did such visionary designers as Hermann Oberth and Robert Goddard show that rockets could reach space. They had confidence in this because they knew enough about fuel, engines, tankage, and structures to calculate what multistage rockets could do. Yet, in 1921 a New York Times editorialist chided Goddard for the notion that rockets could fly through space without air to push against, and as late as 1956 the Astronomer Royal of Britain snorted that “Space travel is utter bilge.” This only showed that editorialists and astronomers were the wrong experts to ask about space hardware. In 1957, Sputnik orbited Earth, followed in 1961 by Yuri Gagarin. In 1969, the world saw footprints on the Moon.

 We paid a price for ignorance, though. Because the pioneers of space technology had lacked any way to establish their case in public, they were forced to argue basic points again and again (“Yes, rockets will work in vacuum…. Yes, they really will reach orbit….). Busy defending the basics of spaceflight, they had little time to discuss its consequences. Thus, when Sputnik startled the world and embarrassed the United States, people were unprepared: there had been no widespread debate to shape a strategy for space.

 Some of the pioneers had seen what to do: build a space station and a reusable spaceship, then reach out to the Moon or asteroids for resources. But the noise of flustered politicians promptly drowned out their suggestions, and U.S. politicians clamored for a big, easy-to-understand goal. Thus was born Project Apollo, the race to land a U.S. citizen on the nearest place to plant a flag. Project Apollo bypassed building a space station and space shuttle, instead building giant missiles able to reach the Moon in one great leap. The project was glorious, it gave scientists some information, and it brought great returns through advances in technology – but at the core, it was a hollow stunt. Taxpayers saw this, congressmen saw this, and the space program shriveled.  

During Apollo, old dreams held sway in the public mind, and they were simple, romantic dreams of settling other planets. Then robot instruments dissolved the dream of a jungle-clad Venus in the reality of a planet-wide oven of high-pressure poison. They erased the lines Earthbound astronomers had drawn on Mars, and with them went both canals and Martians. In their place was a Mars of craters and canyons and dry blowing dust. Sunward of Venus lay the baked rock of Mercury; starward of Mars lay rubble and ice. The planets ranged from dead to murderous, and the dream of new Earths receded to distant stars. Space seemed a dead end.

The New Space Program

A new space program has risen from the ruin of the old. A new generation of space advocates, engineers, and entrepreneurs now aims to make space the frontier it should have been from the beginning – a place for development and use, not for empty political gestures. They have confidence in success because space development requires no breakthroughs in science or technology. Indeed, the human race could conquer space by applying the technologies of twenty years ago – and by avoiding stunt flights, we could probably do it at a profit. Space activities need not be expensive.

 Consider the high cost of reaching orbit today – thousands of dollars per kilogram. Where does it come from? To a spectator at a shuttle launch, shaken by the roar and awed by the flames, the answer seems obvious: the fuel must cost a mint. Even airlines pay roughly half their direct operating costs for fuel. A rocket resembles an airliner – it is made of aluminum and stuffed with engines, controls, and electronics – but fuel makes up almost all its mass as it sits on the launch pad. Thus, one might expect fuel to account for well over half the operating cost of a rocket. But this expectation is false. In the Moon shots, the cost of the fuel needed to reach orbit amounted to less than a million dollars – a few dollars per kilogram delivered to orbit, a fraction of a percent of the total cost. Even today, fuel remains a negligible part of the cost of spaceflight.

 Why is spaceflight so much costlier than air flight? In part, because spacecraft aren’t made in quantity; this forces manufacturers to recover their design costs from sales of only a few units, and to make those few units by hand at great cost. Further, most spacecraft are thrown away after one use, and even shuttles are flown just a few times a year – their cost cannot be spread over several flights a day for years, as the cost of airliners can. Finally, spaceport costs are now spread over only a few flights per month, when large airports can spread their costs over many thousands. All this conspires to make each flight into space dauntingly expensive.  

But studies by Boeing Aerospace Company – the people who brought inexpensive jet transportation to much of the world – show that a fleet of fully reusable shuttles, flown and maintained like airliners, would drop the cost of reaching orbit by a factor of fifty or more. The key is not new technology, but economies of scale and changes in management style.

 Space offers vast industrial opportunities. The advantages of perching observation and communications satellites on orbit are well known. Future communications satellites will be powerful enough to communicate with hand-held stations on the ground, bringing the ultimate in mobile telephone service. Companies are already moving to take advantage of zero gravity to perform delicate separation processes, to make improved pharmaceuticals; other companies plan to grow better electronic crystals. In the years before assemblers take over materials production, engineers will use the space environment to extend the abilities of bulk technology. Space industry will provide a growing market for launch services, dropping launch costs. Falling launch costs, in turn, will stimulate the growth of space industry. Rocket transportation to Earth orbit will eventually become economical.

 Space planners and entrepreneurs are already looking beyond Earth orbit to the resources of the solar system. In deep space, however, rockets swiftly become too expensive for hauling freight – they gobble fuel that itself had to be hauled into space by rockets. Fuel-burning rockets are as old as Chinese fireworks, far older than “The Star-Spangled Banner.” They evolved for natural reasons: compact, powerful, and useful to the military, they can punch through air and fight strong gravity. Space engineers know of alternatives, however.

 Vehicles need no great blasts of power to move through the frictionless vacuum of space. Small forces can slowly and steadily push a vehicle to enormous speeds. Because energy has mass, sunlight bouncing off a thin mirror – a solar sail – provides such a force. The pull of solar gravity provides another. Together, light pressure and gravity can carry a spacecraft anywhere in the solar system and back again. Only the heat near the Sun and the drag of planetary atmospheres will limit travel, forcing sails to steer clear of them.

 NASA has studied solar sails designed to be carried to space in rockets, but these must be fairly heavy and sturdy to survive the stress of launch and unfolding. Eventually, engineers will make sails in space, using a low-mass tension structure to support mirrors of thin metal film. The result will be the “lightsail,” a higher-performance class of solar sail. After a year’s acceleration, a lightsail can reach a speed of one hundred kilometers per second, leaving today’s swiftest rockets in the dust.

 If you imagine a network of graphite-fiber strands, a spinning spiderweb kilometers wide with gaps the size of football fields between the strands, you will be well on your way to imagining the structure of a lightsail. If you picture the gaps bridged by reflecting panels built of aluminum foil thinner than a soap bubble, you will have a fair idea of how it looks: many reflective panels tied close together to form a vast, rippled mosaic of mirror. Now picture a load of cargo hanging from the web like a parachutist from a parachute, while centrifugal force holds the web-slung mirror taut and flat in the void, and you almost have it.

 To build lightsails with bulk technology, we must learn to make them in space; their vast reflectors will be too delicate to survive launch and unfolding. We will need to construct scaffolding structures, manufacture thin-film reflectors, and use remotely controlled robot arms in space. But space planners already aim to master construction, manufacturing, and robotics for other space applications. If we build lightsails early in the course of space development, the effort will exercise these skills without requiring the launch of much material. Though vast, the scaffolding (together with materials for many sails) will be light enough for one or two shuttle flights to lift to orbit.

 A sail production facility will produce sails cheaply. The sails, once built, will be cheap to use: they will have few critical moving parts, little mass, and zero fuel consumption. They will be utterly different from rockets in form, function, and cost of operation. In fact, calculations suggest that the costs will differ by a factor of roughly a thousand, in favor of lightsails.

 Today most people view the rest of the solar system as vast and inaccessible. It is vast; like the Earth, it will take months to circumnavigate by sail. Its apparent inaccessibility, however, has less to do with distance than with the cost of transportation via rocket.

 Lightsails can smash the cost barrier, opening the door to the solar system. Lightsails will make other planets easier to reach, but this will not make planets much more useful: they will remain deadly deserts. The gravity of planets will prevent lightsails from shuttling to their surfaces, and will also handicap industry on a planet’s surface. Spinning space stations can simulate gravity if it is needed, but planet-bound stations cannot escape it. Worse yet, planetary atmospheres block solar energy, spread dust, corrode metals, warm refrigerators, cool ovens, and blow things down. Even the airless Moon rotates, blocking sunlight half the time, and has gravity enough to ground lightsails beyond hope of escape. Lightsails are fast and tireless, but not strong.

 The great and enduring value of space lies in its resources of matter, energy, and room. The planets occupy room and block energy. The material resources they offer are inconveniently placed. The asteroids, in contrast, are flying mountains of resources that trace orbits crisscrossing the entire solar system. Some cross the orbit of Earth; some have even struck Earth, blasting craters. Mining the asteroids seems practical. We may need roaring rockets to carry things up into space, but meteorites prove that ordinary rocks can fall down from space – and like the space shuttle, objects falling from space need not burn up on the way down. Delivering packages of material from an asteroid to a landing target in a salt flat will cost little.

 Even small asteroids are big in human terms: they hold billions of tons of resources. Some asteroids contain water and a substance resembling oil shale. Some contain fairly ordinary rock. Some contain a metal that holds elements scarce in Earth’s crust, elements that sank beyond reach ages ago in the formation of Earth’s metal core: this meteoritic steel is a strong, tough alloy of iron, nickel, and cobalt, bearing valuable amounts of platinum-group metals and gold. A kilometer-wide chunk of this material (and there are many) contains precious metals worth several trillion dollars, mixed with enough nickel and cobalt to supply Earth’s industry for many years.

 The Sun floods space with easily collected energy. A square-kilometer framework holding metal-film reflectors will gather over a billion watts of sunlight, free of interference from cloud or night. In the weatherless calm of space, the flimsiest collector will be as permanent as a hydroelectric dam. Since the Sun puts out as much energy in a microsecond as the human race now uses in a year, energy need not be scarce for some time to come.

 Finally, space itself offers room to live. People once saw life in space in terms of planets. They imagined domed cities built on planets, dead planets slowly converted into Earth-like planets, and Earth-like planets reached after years in a flight to the stars. But planets are package deals, generally offering the wrong gravity, atmosphere, length of day, and location.

 Free space offers a better building site for settlements. Professor Gerard O’Neill of Princeton University brought this idea to public attention, helping to revive interest in space after the post-Apollo crash. He showed that ordinary construction materials – steel and glass – could be used to build habitable cylinders in space, kilometers in length and circumference. In his design, dirt underfoot shields inhabitants from the natural radiation of space, just as Earth’s inhabitants are shielded by the air overhead. Rotation produces an acceleration equaling Earth’s gravity, and broad mirrors and window panels flood the interior with sunlight. Add soil, streams, vegetation, and imagination, and the lands inside could rival the best valleys on Earth as places to live. With just the resources of the asteroids, we will be able to build the practical equivalent of a thousand new Earths.

 By adapting present technology, we could open the space frontier. The prospect is heartening. It shows us an obvious way to bypass terrestrial limits to growth, lessening one of the fears that has clouded our view of the future. The promise of the space frontier can thus mobilize human hope – a resource we will need in abundance, if we are to deal with other problems.

Space and Advanced Technology

By adapting present technology, we could indeed open the space frontier – but we won’t. Along the path foreseen by the current space movement, human civilization would take decades to become firmly established in space. Before then, breakthroughs in technology will open new paths.

 Nowadays, teams of engineers typically take five to ten years to develop a new space system, spending tens to thousands of millions of dollars along the way. These engineering delays and costs make progress painfully slow. In coming years, though, computer-aided design systems will evolve toward automated engineering systems. As they do, engineering delays and costs will shrink and then plummet; computer-controlled manufacturing systems will drop overall costs still further. A day will come when automated design and manufacturing will have made space systems development more than tenfold faster and cheaper. Our progress in space will soar.

 At that time, will space settlers look back on our present space program as the key to space development? Perhaps not. They will have seen more technical progress made in a few years than space engineers previously managed in a few decades. They may well conclude that AI and robotics did more for space development than did a whole army of NASA engineers.

 The assembler breakthrough and automated engineering will combine to bring advances that will make our present space efforts seem quaint. In Chapter 4, I described how replicating assemblers will be able to build a light, strong rocket engine using little human labor. Using similar methods, we will build entire spacecraft of low cost and extraordinary performance. Weight for weight, their diamond-based structural materials will have roughly fifty times the strength (and fourteen times the stiffness) of the aluminum used in the present shuttle; vehicles built with these materials can be made over 90 percent lighter than similar vehicles today. Once in space, vehicles will spread solar collectors to gather abundant energy. Using this energy to power assemblers and disassemblers, they will rebuild themselves in flight to suit changing conditions or the whim of their passengers. Today, space travel is a challenge. Tomorrow, it will be easy and convenient.

 Since nanotechnology lends itself to making small things, consider the smallest person-carrying spacecraft: the spacesuit. Forced to use weak, heavy, passive materials, engineers now make bulky, clumsy spacesuits. A look at an advanced spacesuit will illustrate some of the capabilities of nanotechnology.

 Imagine that you are aboard a space station, spun to simulate Earth’s normal gravity. After instruction, you have been given a suit to try out: there it hangs on the wall, a gray, rubbery-looking thing with a transparent helmet. You take it down, heft its substantial weight, strip, and step in through the open seam on the front.

 The suit feels softer than the softest rubber, but has a slick inner surface. It slips on easily and the seam seals at a touch. It provides a skintight covering like a thin leather glove around your fingers, thickening as it runs up your arm to become as thick as your hand in the region around your torso. Behind your shoulders, scarcely noticeable, is a small backpack. Around your head, almost invisible, is the helmet. Below your neck the suits inner surface hugs your skin with a light, uniform touch that soon becomes almost imperceptible.

 You stand up and walk around, experimenting. You bounce on your toes and feel no extra weight from the suit. You bend and stretch and feel no restraint, no wrinkling, no pressure points. When you rub your fingers together they feel sensitive, as if bare – but somehow slightly thicker. As you breathe, the air tastes clean and fresh. In fact, you feel that you could forget that you are wearing a suit at all. What is more, you feel just as comfortable when you step out into the vacuum of space.

 The suit manages to do all this and more by means of complex activity within a structure having a texture almost as intricate as that of living tissue. A glove finger a millimeter thick has room for a thousand micron-thick layers of active nanomachinery and nanoelectronics. A fingertip-sized patch has room for a billion mechanical nanocomputers, with 99.9 percent of the volume left over for other components.

 In particular, this leaves room for an active structure. The middle layer of the suit material holds a three-dimensional weave of diamond-based fibers acting much like artificial muscle, but able to push as well as pull (as discussed in the Notes). These fibers take up much of the volume and make the suit material as strong as steel. Powered by microscopic electric motors and controlled by nanocomputers, they give the suit material its supple strength, making it stretch, contract, and bend as needed. When the suit felt soft earlier, this was because it had been programmed to act soft. The suit has no difficulty holding its shape in a vacuum; it has strength enough to avoid blowing up like a balloon. Likewise, it has no difficulty supporting its own weight and moving to match your motions, quickly, smoothly, and without resistance. This is one reason why it almost seems not to be there at all.

 Your fingers feel almost bare because you feel the texture of what you touch. This happens because pressure sensors cover the suit’s surface and active structure covers its lining: the glove feels the shape of whatever you touch – and the detailed pattern of pressure it exerts – and transmits the same texture pattern to your skin. It also reverses the process, transmitting to the outside the detailed pattern of forces exerted by your skin on the inside of the glove. Thus the glove pretends that it isn’t there, and your skin feels almost bare.

 The suit has the strength of steel and the flexibility of your own body. If you reset the suit’s controls, the suit continues to match your motions, but with a difference. Instead of simply transmitting the forces you exert, it amplifies them by a factor of ten. Likewise, when something brushes against you, the suit now transmits only a tenth of the force to the inside. You are now ready for a wrestling match with a gorilla.

 The fresh air you breathe may not seem surprising; the backpack includes a supply of air and other consumables. Yet after a few days outside in the sunlight, your air will not run out: like a plant, the suit absorbs sunlight and the carbon dioxide you exhale, producing fresh oxygen. Also like a plant (or a whole ecosystem), it breaks down other wastes into simple molecules and reassembles them into the molecular patterns of fresh, wholesome food. In fact, the suit will keep you comfortable, breathing, and well fed almost anywhere in the inner solar system.  

What is more, the suit is durable. It can tolerate the failure of numerous nanomachines because it has so many others to take over the load. The space between the active fibers leaves room enough for assemblers and disassemblers to move about and repair damaged devices. The suit repairs itself as fast as it wears out.

 Within the bounds of the possible, the suit could have many other features. A speck of material smaller than a pinhead could hold the text of every book ever published, for display on a fold-out screen. Another speck could be a “seed” containing the blueprints for a range of devices greater than the total the human race has yet built, along with replicating assemblers able to make any or all of them.  

What is more, fast technical AI systems like those described in the last chapter could design the suit in a morning and have it built by afternoon.

 All that we accomplish in space with modern bulk technology will be swiftly and dramatically surpassed shortly after molecular technology and automated engineering arrive. In particular, we will build replicating assemblers that work in space. These replicators will use solar energy as plants do, and with it they will convert asteroidal rubble into copies of themselves and products for human use. With them, we will grasp the resources of the solar system.

 By now, most readers will have noted that this, like certain earlier discussions, sounds like science fiction. Some may be pleased, some dismayed that future possibilities do in fact have this quality. Some, though, may feel that – sounding like science fiction” is somehow grounds for dismissal. This feeling is common and deserves scrutiny.

 Technology and science fiction have long shared a curious relationship. In imagining future technologies, SF writers have been guided partly by science, partly by human longings, and partly by the market demand for bizarre stories. Some of their imaginings later become real, because ideas that seem plausible and interesting in fiction sometimes prove possible and attractive in actuality. What is more, when scientists and engineers foresee a dramatic possibility, such as rocket-powered spaceflight, SF writers commonly grab the idea and popularize it.

 Later, when engineering advances bring these possibilities closer to realization, other writers examine the facts and describe the prospects. These descriptions, unless they are quite abstract, then sound like science fiction. Future possibilities will often resemble today’s fiction, just as robots, spaceships, and computers resemble yesterday’s fiction. How could it be otherwise? Dramatic new technologies sound like science fiction because science fiction authors, despite their frequent fantasies, aren’t blind and have a professional interest in the area.

 Science fiction authors often fictionalize (that is, counterfeit) the scientific content of their stories to “explain” dramatic technologies. Some fuzzy thinkers then take all descriptions of dramatic technical advances, lump them together with this bogus science, and ignore the lot. This is unfortunate. When engineers project future abilities, they test their ideas, evolving them to fit our best understanding of the laws of nature. The resulting concepts must be distinguished from ideas evolved to fit the demands of paperback fiction. Our lives will depend on it.

 Much will remain impossible, even with molecular technology. No spacesuit, however marvelous, will be able to rocket back and forth indefinitely at tremendous speeds, or survive great explosions, or walk through walls, or even stay cool indefinitely in a hot isolated room. We have far to go before reaching the limits of the possible, yet limits exist. But this is a topic taken up later.


Space resources join with assemblers and automated engineering systems to round out the case for a future of great material abundance. What this means can best be seen by examining costs.

 Costs reflect the limits of our resources and abilities; high costs indicate scarce resources and difficult goals. The prophets of scarcity have in effect predicted steeply rising resource costs, and with them a certain kind of future. Resource costs, however, always depend on technology. Unfortunately, engineers attempting to predict the cost of future technologies have generally encountered a tangle of detail and uncertainty that proves impossible to untie. This problem has obscured our understanding of the future.

 The prospect of replicating assemblers, automated engineering, and space resources cuts this Gordian knot of cost prediction. Today the cost of products includes the costs of labor, capital, raw materials, energy, land, waste disposal, organization, distribution, taxation, and design. To see how total costs will change, consider these elements one by one.

 Labor. Replicating assemblers will require no labor to build, once the first exists. What use are human hands in running an assembler? Further, with robotic devices of various sizes to assemble parts into larger systems, the entire manufacturing process from assembling molecules to assembling skyscrapers could be free of labor costs.

 Capital. Assembler-based systems, if properly programmed, will themselves be productive capital. Together with larger robotic machines, they will be able to build virtually anything, including copies of themselves. Since this self-replicating capital will be able to double many times per day, only demand and available resources will limit its quantity. Capital as such need cost virtually nothing.

 Raw materials. Since molecular machines will arrange atoms to best advantage, a little material can go a long way. Common elements like hydrogen, carbon, nitrogen, oxygen, aluminum, and silicon seem best for constructing the bulk of most structures, vehicles, computers, clothes and so forth: they are light and form strong bonds. Because dirt and air contain these elements in abundance, raw materials can be dirt cheap.

 Energy. Assemblers will be able to run off chemical or electrical energy. Assembler-built systems will convert solar to chemical energy, like plants, or solar to electrical energy, like solar cells. Existing solar cells are already more efficient than plants. With replicating assemblers to build solar collectors, fuel and electric power will cost little.

 Land. Assembler-based production systems will occupy little room. Most could sit in a closet (or a thimble, or a pinhole); larger systems could be placed underground or in space if someone wants something that requires an unsightly amount of room. Assembler-based production systems will make both digging machines and spacecraft cheap.

 Waste disposal. Assembler systems will be able to keep control of the atoms they use, making production as clean as a growing apple tree, or cleaner. If the orchard remains too dirty or ugly, we will be able to move it off Earth entirely.

 Organization. Today, factory production requires organization to coordinate hordes of workers and managers. Assembler-based production machines will contain no people, and will simply sit around and produce things made to order. Their initial programming will provide all the organization and information needed to make a wide range of products.

 Distribution. With automatic vehicles running in tunnels made by cheap digging machines, distribution need neither consume labor nor blight the landscape. With assemblers in the home and community, there will be less need for distribution in the first place.

 Taxation. Most taxes take a fixed percentage of a price, and thus add a fixed percentage to the cost. If the cost is negligible, the tax will be negligible. Further, governments with their own replicators and raw materials will have less reason to tax people.

 Design. The above points add up to a case for low costs of production. Technical AI systems, by avoiding the labor cost of engineering, will virtually eliminate the costs of design. These AI systems will themselves be inexpensive to produce and operate, being constructed by assemblers and having no inclination to do anything but design things.

 In short, at the end of a long line of profitable developments in computer and molecular technologies, the cost of designing and producing things will drop dramatically. I above referred to “dirt cheap” raw materials, and indeed, assemblers will be able to make almost anything from dirt and sunlight. Space resources, however, will change “dirt cheap” to “cheap-dirt cheap”: topsoil has value in Earth’s ecosystem, but rubble from asteroids will come from a dead and dreary desert. By the same token, assemblers in space will run off cheap sunlight.

 Space resources are vast. One asteroid could bury Earth’s continents a kilometer deep in raw materials. Space swallows the 99.999999955 percent of the Sun’s light that misses Earth, and most is lost to the interstellar void.

 Space holds matter, energy, and room enough for projects of vast size, including vast space settlements. Replicator-based systems will be able to construct worlds of continental scale, resembling Dr. O’Neill’s cylinders but made of strong, carbon-based materials. With these materials and water from the ice moons of the outer solar system, we will be able to create not only lands in space, but whole seas, wider and deeper than the Mediterranean. Constructed with energy and materials from space, these broad new lands and seas will cost Earth and its people almost nothing in terms of resources. The chief requirement will be programming the first replicator, but AI systems will help with that. The greatest problem will be deciding what we want.

 As Konstantin Tsiolkovsky wrote near the turn of the century, “Man will not always stay on Earth; the pursuit of light and space will lead him to penetrate the bounds of the atmosphere, timidly at first, but in the end to conquer the whole of solar space.” To dead space we will bring life.

 And replicators will give us the resources to reach for the stars. A lightsail driven starward only by sunlight would soon find itself coasting in the dark – faster than any modern rocket, yet so slowly that it would take millennia to cross the interstellar gulf. We can build a tremendous bank of lasers orbiting the Sun, however, and with it drive a beam far beyond our solar system, pushing a sail toward the speed of light. The crossing then will take only years.

 Stopping presents a problem. Freeman Dyson of Princeton suggests braking with magnetic fields in the thin ionized gas between the stars. Robert Forward of Hughes Research Laboratories suggests bouncing laser light off the sail, directing light back along the sail’s path to decelerate a smaller sail trailing behind. One way or another (and there are many others), the stars themselves lie within our reach.

 For a long time to come, however, the solar system can provide room enough. The space near Earth holds room for lands with a million times Earth’s area. Nothing need stop emigration, or return visits to the old country. We will have no trouble powering the transportation system – the sunlight falling on Earth supplies enough energy in ten minutes to put today’s entire population in orbit. Space travel and space settlements will both become cheap. If we make wise use of molecular technology, our descendants will wonder what kept us bottled up on Earth for so long, and in such poverty.

The Positive-Sum Society

It might seem that the cost of everything – even land, if one doesn’t crave thousands of kilometers of rock underfoot – will drop to nothing. In a sense, this is almost right; in another sense, it is quite false. People will always value matter, energy, information, and genuine human service, therefore everything will still have its cost. And in the long run, we will face real limits to growth, so the cost of resources cannot be dismissed.

 Nonetheless, if we survive, replicators and space resources will bring a long era in which genuine resource limits do not yet pinch us – an era when by our present standards even vast wealth will seem virtually free. This may seem too good to be true, but nature (as usual) has not set her limits based on human feelings. Our ancestors once thought that talking to someone across the sea (many months’ voyage by sailing ship) would be too good to be true, but undersea cables and oversea satellites worked anyway.

 But there is another, less pleasant answer for those who think assemblers are too good to be true: assemblers also threaten to bring hazards and weapons more dangerous than any yet seen. If nanotechnology could be avoided but not controlled, then sane people would shun it. The technology race, however, will bring forth assemblers from biotechnology as surely as it brought forth spacecraft from missiles. The military advantages alone will be enough to make advances almost inevitable. Assemblers are unavoidable, but perhaps controllable.

 Our challenge is to avoid the dangers, but this will take cooperation, and we are more likely to cooperate if we understand how much we have to gain from it. The prospect of space and replicating assemblers may help us clear away some ancient and dangerous memes.

 Human life was once like a zero-sum game. Humankind lived near its ecological limit and tribe fought tribe for living space. Where pastures, farmland, and hunting grounds were concerned, more for one group meant less for another. Because one’s gain roughly equaled the other’s loss, net benefits summed to zero. Still, people who cooperated on other matters prospered, and so our ancestors learned not just to grab, but to cooperate and build.

 Where taxes, transfer payments, and court battles are concerned, more for one still means less for another. We add to total wealth slowly, but redistribute it swiftly. On any given day our resources seem fixed, and this gives rise to the illusion that life is a zero-sum game. This illusion suggests that broad cooperation is pointless, because our gain must result from some opponent’s loss