Perhaps you've heard of MEMS, microelectromechanical systems, a field being invested in heavily by governments and corporations. In MEMS, the components are usually between 10 and 100 microns in size. Using MEMS, you can build gear systems smaller than a dust mite. The military is looking into MEMS to build spy-bots the size of the smallest bugs.
Beyond MEMS there is NEMS, nanoelectromechanical systems, an area scientists and engineers are just beginning to investigate. NEMS are about a 1000 times smaller than MEMS, with components between 10 and 100 nanometers in size. With NEMS, you could build a complex machine the size of a red blood cell or smaller. Transhumanists hope to use NEMS to improve our health and expand our sensory and motor capabilities.
The Holy Grail of nanotechnology is designing a NEMS that can build other NEMS. This goal has been called molecular nanotechnology (MNT), and it is a topic of controversy within the nanotechnology community. Some futurists and scientists believe MNT is impossible, while others consider it very likely.
Here are some feasibility arguments for molecular nanotechnology:
1) We already have working examples of molecular nanotechnology: living things. Every organism depends on nanoscale assemblers called ribosomes to synthesize all their parts, including copies of the ribosomes themselves. Specialized organelles, like the Golgi apparatus, may process these proteins further. This is similar to an assembly line in a factory, where a series of tools perform different collaborative functions to achieve a predetermined outcome. Nanotechnologists seek to duplicate this scheme in an inorganic medium.
2) Positional placement of individual atoms has already been demonstrated numerous times. In 1999, researchers at Cornell University synthesized single molecules of iron carbonyl (FeCO) from iron and CO2 precursors using an exceptionally precise STM. What is lacking here is not a proof of principle, but the need to miniaturize the manipulation apparatus and make it more reliable. This is primarily an engineering challenge, albeit a difficult one.
3) At the nanoscale, proteins tend to be floppy, while an inorganic material like diamond can be relatively rigid. In the 1992 book Nanosystems, nanotechnologist Eric Drexler offers numerous designs for a broad range of diamondoid nanoscale machine components, including motors, generators, pistons, rods, interlocking structures, gears, bearings, belt-and-roller systems, rachets, clutches, sorters, and many others. Drexler shows how these systems are physically feasible and could work at acceptable speeds without overheating. In the 16 years since its publication, no one has yet found a mathematical error in Nanosystems.
4) Mechanosynthesis -- the synthesis of chemicals through mechanical action alone -- is a desired capability for a "dry" molecular nanotechnology system that uses NEMS to build NEMS. As mentioned above, researchers have already been able to synthesize individual molecules from atomic constituents. What is needed next is to extend these techniques to carbon. In the next decade, nanotechnologists hope to demonstrate diamondoid mechanosynthesis -- the mechanical synthesis of complex carbon structures. Many thousands of hours of computing time has already been spent simulating diamondoid mechanosynthesis, and experimental work is just beginning.
5) Many rudimentary molecular machines and components have already been built. These include Nadrian Seeman's DNA walker robot (2004) and other nanomechanical devices, the Rice University nanocar (2005), molecular logic gates, and more. Some nanomachine components, like the bacterial flagellar motor, already come pre-built from nature. Many nanotechnologists see inspiration from biology as key. Obviously, there is no lack of available nanoscale machines -- the challenge is putting them together into reliable and reprogrammable systems.
As we can see, we are much closer to the goal of molecular nanotechnology than we were only 10 years ago. Going back further, to 20 years ago, very few scientists could have even imagined what we'd be achieving now. Our goal for the future should be to push the envelope of nanotechnology research, devoting more money to research in molecular nanotechnology, while carefully studying the potential benefits and risks that could arise from a major breakthrough in the area.
See also: Six challenges for molecular nanotechnology, by EPSRC Senior Strategic Advisor Richard Jones.