If Moore's Law about amount of transistors per die or whatever doubling every so many months holds true we will keep seeing smaller more efficent processors...
But it won't (I know you agree)
In order to reduce the process size, you have to manipulate smaller and smaller elements. For example, photolithography has been a primary way to produce circuitry on chips. But it is already to to point where the wavelength of regular light is longer than the design details of the circuits, and the process has to be reengineered with shorter and shorter wavelengths of radiation. The current state of the art is 29 nm in the lab and later this year 45 nm in production. However the cost of production rises dramatically as you have to use more exotic processes.
Other problems with continued shrinking of the process are
leakage, a host of resistance and capacitance issues, heat dissipation, and greater problems in purity in the materials. Once you get really small, 'normal' macro-world physics (which by and large relies on the
average properties of large collections of atoms) starts to get distorted by atomic-level forces, and variations among individual atoms.
http://en.wikipedia.org/wiki/Moore's_law#Future_trends
In Gordon Moore's own words:
"In terms of size [of transistor] you can see that we're approaching the size of atoms which is a fundamental barrier, but it'll be two or three generations before we get that farbut that's as far out as we've ever been able to see. We have another 10 to 20 years before we reach a fundamental limit. By then they'll be able to make bigger chips and have transistor budgets in the billions."
