If you look around you today, the contest of flesh and blood versus machine seems to be over. The machines won.

John Henry "died with a hammer in his hand," trying to keep up with a steam-powered pile driver. No horse can compete with car or train in speed or comfort for a long journey, no ox can plow like a John Deere tractor or reap like a combine harvester, no tree can grow as tall as the Empire State Building, and no human can beat even the most modest handheld calculator in arithmetic speed and precision.

At another level, however, the competition is far from over, and much of the smart money is backing biology. The level I'm talking about is the world of the very small, the world of the individual cell, where every human is a lumbering giant.

It takes a lot of cells to make a human. Your body contains many trillions of them, each one about a hundred micrometers across, just about visible without using a microscope. In humans, and in all multi-cellular plants and animals, cells have become specialized and depend on each other for the whole organism to survive; your liver cells need your lung cells, and your brain cells need your blood cells.

However, single-celled organisms were here billions of years before us and our multi-celled cousins, and they managed a complete life style. Each single cell is a fully equipped chemical factory. They eat, grow, move, excrete waste products, reproduce, and compete for living space, just like their larger cousins. It is a fact that we could not survive without single-celled organisms, whereas they could and can live very well without us.

Until very recently, biology had the world of the very small all to itself. When we began to make machines, it was natural at first to make them about the same size as ourselves, so that we could drive and control them. And when we did think of changes, it was natural to make things larger, so that we could make bigger and bigger ships, bridges, superhighways, and skyscrapers. Only recently, about forty years ago, did anyone suggest exploring the downward direction. Human-scale machines could be used to make yet smaller machines, which would construct smaller ones yet, and so on until we could at last build devices as small as living cells, or even as small as molecules.

We might call this the "top down" method. Another way to proceed would be to build these microscopic machines "bottom up," manipulating the exact placement of atoms, one at a time, to create the whole structure. The advantage of this approach is that, since individual atoms are identical, machines could also be absolutely identical in both structure and performance. We will no longer have the irritating variability, which at our usual scale of manufacturing guarantees that every now and again an automobile assembly line will produce a lemon.

Whether we build down and down, or whether we build from atoms up, the resulting machines will be of microscopic size. The general name for the building and use of molecular-level machinery was coined by Eric Drexler. It is nanotechnology. The devices made in this way are nanomachines.

Nanomachines, like the machines in many of today's automated factories, can build a wide variety of structures. They need to be given a specification (usually, today, that's in the form of a computer program) as to what their output must be. There is, in principle, no reason why large numbers of nanomachines should not work in concert to produce anything from a ship's hull to a house to a broadcasting tower. In fact, any structure or material that can be completely specified can also be built. We could, for instance, direct our nanomachines to produce a material identical to wood, but without the knotholes or imperfections of ordinary timber.

Since a nanomachine can, in principle, produce anything that a biological organism can make, and more, you might feel that in the world of the very small, as in our everyday world, the machines will win hands down. We have, however, omitted one crucial point. A house, or a ship big enough to carry useful amounts of cargo, is made up of trillions of trillions of molecules. If we make our nanomachines one at a time and set them to work, it will take them forever to painstakingly assemble, molecule by molecule, our final product.

The nanomachine must be able to do one other thing. In addition to contributing to the assembly of a final product, it must be able to make an exact copy of itself. Only when it is self-replicating, as all living cells are self-replicating, can it hope to compete with living microscopic factories.

This is a real challenge. Reproduction, which all forms of Nature accomplish without apparent effort, turns out to be a highly complicated operation. Living organisms make it look easy only because they have had close to four billion years of practice. The idea of making self-replicating machines is less than half a century old. At the moment we are able to make very small machines, and are becoming more and more skilled at doing so; but we cannot yet make them produce copies of themselves.

At the same time as progress is made on self-replicating machines, molecular biologists are learning more and more about the natural chemical factories inside living cells. Already, we manipulate them to produce enzymes needed for medical treatments, while Nature has no difficulty putting together an object the size of a blue whale, or a giant redwood tree. Why not instruct the manufacturing machinery within a cell to reproduce as usual, and as a side-product grow and leave behind a useful constructed object, just as in nature a living animal grows an oyster shell, or a coral reef? That may prove easier than building machines able to make exact copies of themselves.

A hundred years ago, it seemed that the race between the mechanical and the biological approach was over and done with. Now that result is not so clear. The next half-century may swing the pendulum back in favor of Nature.

"Borderlands of Science" is syndicated by: