|
The introduction of atomic positioners into manufacturingConsider:
When I say how many people, I mean not only the ones who work in the factory itself, but all upstream people, everyone whose presence is required for the factory to operate. The engineer who designed the air filter system, the architect who designed the building, the programmers who wrote the software the engineers use, the other programmers who wrote the compilers and other tools that those programmers use, the professors in the MBA program that the CEO attended, the banker who arranged financing, the scientists who did the basic research that made the whole thing possible... everybody.
And now, my main question:
The idea here is to get the idea of molecular manufacturing in focus by putting it into a realistic economic context. What will happen when the first atomic positioners make their debut in a factory? They will be introduced into a certain building at a certain time. They will have been designed by somebody, over a period of several years, at great cost. (Please don't tell me they will be designed for free, by Genies.) They will be installed and supervised by the same engineers who designed them (and there will only be a handful of people in the world who know how to do this -- such people are not cheap to employ). The atomic positioners will have some kind of supporting apparatus. They will interface with other machines. They will use materials unloaded at a truck dock. They will be owned by somebody. They will be an item in somebody's budget, along with payroll, rent, utilities, interest on debt, cost of training employees, cost of replacing machines, and so forth. How will that budget change, as a result of making the manufacturing process more fine-grained and more automated at a certain point? -- The budget will increase! Using atomic positioners instead of lithography does not change the situation in any essential way. If anything, it will make things worse.
The answer to question h is -- none of this will change. The focal point of the manufacturing process, where a pattern is etched into silicon, is already very fine-grained, and getting more fine-grained all the time. However, this focal point exists within a larger context. If you consider the manufacturing process as a whole, you find this somewhat paradoxical fact: as the focal point becomes smaller, the factory gets bigger, and the cost of the whole operation increases. The focal point will always require jigs and scaffolds. Consider the whole factory, and the whole industrial infrastructure, to be the scaffold. The more exact your control over individual atoms, the bigger and more elaborate the infrastructure has to be. Chip factories that use atomic positioners will still be part of Silicon Valley, in an extended sense -- i.e. those factories can only exist as part of an ever-growing network of other factories, universities, etc. The factories required to make very fine-grained machines will continue to get more expensive with each generation of chips. Atomic positioners are not going to change this situation at all. They are just going to be another step in the same direction we are already going. The cost of retooling is bad enough when we go from one generation of lithography to the next generation -- going from lithography to atomic positioners will require a whole new level of retooling. Atomic positioners will also have generations. Every year new equipment will come into use. There will be new drivers to install, new software upgrades and service packs, new machines that are incompatible with last year's machines, and so forth, just like now. Programmable matter will involve the same hassles as programmable computers. In the absence of Genies to design everything for us, going from one generation of atomic positioning machines to the next generation is going to be just as expensive as going from one generation of lithographic machines to the next. Note added December 25, 2000: The conclusion of the last paragraph above is not necessarily correct. It is true that going from lithography to atomic positioners will involve a whole new level of retooling, and that step will be enormously expensive, but after that step has been taken, going from one generation of atomic positioners to the next may or may not be expensive. It may be like going from one generation of genetically engineered tomatoes to the next, and I can't say how expensive that will be. "Programmable matter" will take the form of programmable cells. Genetic engineering will exist in an economic context, but this implies nothing about how much it will cost. |