By 2020, and potentially as early as 2010, we will know enough about carbon chemistry, kinematic self-replication, and nanoscale positional control to build a desktop nanofactory - a machine that uses many trillions of tiny arms to put together macro-scale products. Because tiny arms can move incredibly fast, they will be radically productive. It has been estimated that a 100 kg nanofactory will be able to manufacture its own weight in product in about three hours, perhaps less.
Nanofactory technology will begin with an assembler - a reprogrammable molecular machine capable of making a copy of itself. An assembler would be extremely small, composed of maybe a couple million atoms. This is about the same as a ribosome. For a reference, see this picture of some nanoparts next to a virus:
An assembler would basically be an artificial ribosome. Ribosomes are the little machines in the cell that manufacture every protein in your body. Its basic design hasn't changed in over a billion years.
Feasibility arguments for molecular nanotechnology (MNT) are well-documented in the literature. Its not a question of if, but when. The technological and sociological impact of personal nanofactories (PNs) is certain to be extreme. If regulations permit it, you will be able to construct, right in your very home, just about any structure allowed by the laws of chemistry and available feedstock. All current manufacturing, communication, and transportation processes will be fundamentally restructured over a period of mere years or even months. The first nanofactories are likely to use carbon feedstock, meaning most of the products will be made out of diamond. Water may be used as a ballast for some diamond products.
Products built using MNT will be extremely cheap: around the cost of their raw materials. This is because human labor, the primary cost of manufacturing today, is largely subtracted from the equation. Carbon is extremely cheap, and can be mined by the megaton from practically anywhere. Power requirements are modest. Made of diamond, a nanofactory will not require much maintenance.
Quickly, typical products made of plastic, ceramic, or metal will be redesigned to accommodate the new diamondoid medium. There will be diamond plates, diamond tables, diamond cutlery, ovens, coffee makers, microwaves, tiles, walls, chairs, televisions, cameras, printers, scanners, shelving, windows, computers, pens, notepads, pottery, showerheads, and so on. Something like 90% of all manufactured products will be replaced by diamondoid versions. This is what Neal Stephenson was thinking when he wrote a book called The Diamond Age.
The father of nanotechnology, Eric Drexler, lists a few things which would become possible with MNT on his website:
desktop computers with a billion processors
inexpensive, efficient solar energy systems
medical devices able to destroy pathogens and repair tissues
materials 100 times stronger than steel
superior military systems
additional molecular manufacturing systems
MNT has been called "magic", and the word choice is not entirely inappropriate. We will be able to build products with greater performance and more diverse functionality than anything you or any university Ph.Ds have imagined. All shortages of energy, food, water, and shelter will be rapidly solved, as long as nanofactories are made available to developing countries. Subdermal heaters, nanoproducts designed to do little more than generate waste heat, will eliminate the problem of obesity practically overnight. The size and range of products will be limited only by whatever regulations are built into the first round of nanofactories. And I hope that these regulations are extremely strict. You see, nanofactories will be the most dangerous technology that mankind has ever faced, thousands of times more dangerous than nuclear weapons.
Given an unrestricted nanofactory and a few million dollars worth of programming and engineering, here are a few products that I could manufacture in almost arbitrary quantities, given a couple months manufacturing time:
sniper rifles that weigh less than 5 kg, capable of firing a lethal projectile at Mach 10 towards any target within my line of sight.
extremely light and strong armor capable of stopping 10 kg explosive shells moving at faster than 10 km/sec.
Metal Storm systems which fire as many as 1,000,000 projectiles per minute through ballistics arrays.
UAV swarms capable of actively neutralizing very large rockets, providing comprehensive area denial, working together to disassemble buildings, etc.
highly maneuverable VTOL craft able to destroy almost any number of F-22 Raptors or F-35 Lightnings.
gigawatt-class, solar array or nuclear-powered microwave beams capable of completely melting tanks, aircraft, destroyers, incoming missiles, etc. from hundreds of miles away.
isotope separation systems that enrich uranium efficiently, at great speeds, giving enough fissile material to make bombs in days rather than years.
gigantic lenses capable of redirecting sunlight towards arbitrary coordinates in extremely high concentrations; a solar furnace.
missile swarms composed of individual missiles about 1 meter long, carrying 1 kg warheads, manufactured by the millions, capable of traveling through the upper atmosphere and surviving reentry.
Because products made out of diamond can be extremely strong and light, 100 kg of carbon gives you a very large bang for your buck. For example, a Mercedes S-class today weighs about 2,000 kg, but with diamondoid building materials, this weight could be reduced tremendously, if desired - the primary motivation to preserve the vehicle's current weight would be the preservation of inertia, rather than engineering limitations. An automobile made out of nanodiamond could have an absurdly low weight, on the order of a hundreth of an ounce, not including fuel. If this sounds fantastic to you, take a look at what is already possible today:
This tiny block of transparent aerogel is supporting a brick weighing 2.5 kg. The aerogel's density is 0.1 g/cm^3.
Anyway, the point of all this is simple: nanofactories need to be extremely restricted in the products they can build, or there is going to be big problems. The open source, anti-digital rights management, P2P-generation needs to get this. Information may want to be free, but if weapons designs are readily available and manufacturable in the post-MNT world, there are going to be problems of the likes we've never seen. To minimize the risk of danger, the safest option is to have all product designs authenticated by a central authority. Yes, that scary phrase, "central authority". This central authority needs to be capable of determining which designs are safe, maintaining an extremely high level of nanofactory security, and enforcing the law when people try to circumvent it. The libertarian dream of minimalist government, unfortunately, must be discarded.
Now, in general, I'm extremely against big government. It can be a huge waste, and extremely inefficient relative to market-driven competition. But when it comes to managing magic, decentralized solutions simply won't do. There needs to be a global standard and global regulations. Rogue states won't do, either. One rogue nation could use MNT to manufacture enough weapons to turn the capitals of any opposing nation, no matter how large, into a series of smoking craters. This is a risk we shouldn't be willing to take, and once the potential of MNT starts to sink in with higher-level government officials, they won't.
Life extensionists: realize that the greatest risk to living longer is not actually aging, which we will eventually defeat cleanly, but existential risks of the type I frequently discuss, including superintelligence and nanotech arms races. You can extend your expected future life more by lowering the probability of these disasters than through any other means.