We consider our human size to be normal—except, perhaps, for the odd 10 pounds or so that we can’t seem to get rid of around the waist. But set against most creatures, we’re swollen, lumbering giants. As Steven Vogel, a professor of zoology at Duke University, has pointed out, the “typical” species is about an eighth of an inch long. That’s the size that evolution apparently deems works best on this planet.

Naturally, when we Brobdingnagian exotics started to build machines, we found it convenient to build them to approximately our proportions. But we’re learning that in most cases smaller machines, like smaller critters, work better. All else being equal, small machines are stronger (per unit of weight), more precise, faster, quieter, more durable, cooler, safer, and cheaper to make and run. Although we use large machines like bulldozers to excavate foundations, we know that an equivalent weight of robot ants, each removing a grain of sand at a time, could lift far more.

Small machines open up such amazing new applications that it’s easy to foresee the day when doctors perform surgery by injecting tiny machines into the patient’s bloodstream, which would then be controlled remotely. Furniture could be made of “intelligent sand”—teeny devices that could be ordered to form a chair one day, a table the next. In fact, despite the general feeling that we live in a digital age, tiny machines hold the prospect that we may be on the verge of a new mechanical era.

The reason that small machines work so well stems from simple math: As objects shrink, their volume declines much faster than their surface area. Reducing the length of the sides of a cube by a factor of 1,000—the sort of change that researchers are attempting—would shrink volume by a factor of one billion, while cutting surface area by just a factor of one million. As Vogel nicely puts it, “a whale is big inside with little outside, while a bacterium is big outside with little inside.”

Changing the ratio between volume and surface area fundamentally changes the behavior of an object. As an object shrinks, the properties associated with surfaces (drag, heat, and the ability to have chemicals pass between the inside and outside of the object) become steadily more important than those associated with volumes (weight, inertia, friction). This is a good trade, because the surface properties are usually the useful ones, while the properties associated with volume include increased material costs and higher likelihood of flaws.

Already, Texas Instruments is selling a digital movie projector comprising thousands of tiny mirrors that cooperate to light up a commercial movie screen. DuPont is thinking about replacing its system of chemical production (a few large reactor vessels, each handling thousands of liters of product) with millions of microreactors, each processing a few milliliters of product. The new approach would be simpler and cheaper to build. It would be easier to control, and the microreactors would run at higher, more efficient temperatures. There would also be less danger: If one of these microvessels blew up, the problem would not be to contain the explosion but to find it at all.

Perhaps the example that could change our lives the most, soonest is a battery being developed at MIT. The battery would be a complete electrical power plant built on the same principles as those that power factories or small towns, but shrunk to the size of a collar button (not counting the fuel tank). Theoretically, the device should have 50 times the energy density of a conventional battery. Laptops would run for months on a single charge. Appliances from computers to refrigerators might come with their power installed. Because making the batteries should be cheap—micromachines are designed to be printed out in large numbers, just like computer chips—the problem of bringing electricity to remote areas would also be solved.

The battery’s potentially stunning performance comes because the power produced by a generator is directly related to how fast its rotor spins. The MIT device’s tiny mass means it can spin at a terrific speed (two million rpm) without being torn apart by centrifugal force. Another plus of small size is that the rotors spin so fast they can’t be heard by the human ear.

The microturbine can even be used for propulsion, like an ordinary jet engine. Aerodynamical engineers are investigating the possibility of powering jetliners by embedding thousands of microturbines in the planes’ wings. This would reduce turbulence by removing those big engines that now hang down into the airflow.

Like all technologies, micromachines do have a troubling side. Tiny rocket engines might be used to power fleets of cheap intercontinental ballistic missiles, each the size of a human leg. The U.S. Advanced Research Project Authority, which brought us the Internet, is financing research into microairplanes the size of birds (and, eventually, large insects). These would be platforms for cameras, microphones, or molecular sensors sniffing for pollution plumes, forest fires, or the pungent smell of smoldering marijuana. While various government agencies love the idea, the technology raises real privacy issues. Imagine if tabloids sent out fleets to cruise for celebrities.

The military is interested in making bullets that could ride a laser beam to their target exactly like a smart bomb. These “smart bullets” would come with a jacket of microthrusters, an array of small packages of ignitable material, and an onboard microcomputer that would track the laser beam and then fire groups of thrusters in the patterns needed to keep the bullet holding the beam right to the target. In other words, the military’s new command could be: Fire, ready, aim!

We often divide our society’s history into “industrial” and “postindustrial.” The sense of this distinction is that the mechanical age is over, that the future belongs to smarter computers and biotech. But we might actually be sailing into a second mechanical era, one in which our houses will be dusted and polished by mechanical snails, our gardening chores aided by mechanical ants, and our home security taken care of by nests of mechanical wasps.

Hapgood, a free-lance writer based in Boston, describes himself as an "intellectual property provider." He can be reached at hapgood@pobox.com.