Table of Contents
Also by Neil Gershenfeld
When Things Start to Think
The Nature of Mathematical Modeling
The Physics of Information Technology
For Kalbag
and
Kelly, Meejin, Dalia, Irene, Saul, Alan, Grace,
Eli, Mel, Kyei, Anil, Frank, Larry, Etienne, Seymour,
David, Kenny, Amy, Haakon, Vicente, Terry,
Sugata, Arjun
and
Lass, Susan, John, Bakhtiar, Amon, Joe, Aisha,
Chris, Caroline, Manu, Sanjay, Debu,
Jrgen, Benn, Milton, Joe, Ike, Scott, Seth,
Alex, Mitch, Marvin
How to Make...
Mainframe computers were expensive machines with limited markets, used by skilled operators working in specialized rooms to perform repetitive industrial operations. We can laugh in retrospect at the small size of the early sales forecasts for computers; when the packaging of computation made it accessible to ordinary people in the form of personal computers, the result was a unprecedented outpouring of new ways to work and play.
However, the machines that make computers (and most everything else) remain expensive tools with limited markets, used by skilled operators working in specialized rooms to perform repetitive industrial operations. Like the earlier transition from mainframes to PCs, the capabilities of machine tools will become accessible to ordinary people in the form of personal fabricators (PFs). This time around, though, the implications are likely to be even greater because whats being personalized is our physical world of atoms rather than the computers digital world of bits.
A PF is a machine that makes machines; its like a printer that can print things rather than images. By personal fabrication, I mean not only the creation of three-dimensional structures but also the integration of logic, sensing, actuation, and displayeverything thats needed to make a complete functioning system. With a PF, instead of shopping for and ordering a product, you could download or develop its description, supplying the fabricator with designs and raw materials.
Programmable personal fabricators are not just a prediction, theyre a reality. The world of tomorrow can be glimpsed in tools available today. Fab tells the stories of these remarkable tools and their equally remarkable users around the world. It explains what can be made, and why, and how.
I first encountered the possibility of personal fabrication through the unexpectedly enthusiastic student response to a class that I teach at MIT, modestly titled How To Make (almost) Anything. At MIT I direct the Center for Bits and Atoms. CBA comprises fifteen or so faculty from across campus: physicists, chemists, biologists, mathematicians, and mechanical and electrical engineers. They all, like me, never fit into the artificial separation of computer science from physical science.
The universe is literally as well as metaphorically a computer. Atoms, molecules, bacteria, and billiard balls can all store and transform information. Using the discrete language of computation rather than the continuous equations of calculus to describe the behavior of physical systems is not only leading to the practical development of new and more powerful kinds of information technologies, such as quantum computers, its also leading to new kinds of insights into the nature of the universe itself, such as the long-term behavior of black holes. If the world is a computer, then the science of computing is really the science of science.
At the intersection of physical science and computer science, programs can process atoms as well as bits, digitizing fabrication in the same way that communications and computation were earlier digitized. Ultimately, this means that a programmable personal fabricator will be able to make anything, including itself, by assembling atoms. It will be a self-reproducing machine. That idea has been a long-standing science fiction staple for better or, sometimes, much worse.
In Star Trek: The Next Generation, the replicator is an essential plot element that is capable of making whatever is needed for each episode. It looks like an overgrown drinks dispenser, but it has the useful feature of being able to dispense anything. In theory, it does this by following stored instructions to put together subatomic particles to make atoms, atoms to make molecules, and molecules to make whatever you want. For Captain Picard, that was frequently a steaming mug of his preferred tea, obtained from the replicator with the command Tea, Earl Grey, hot.
The less fortunate Arthur Ford in the Hitchhikers Guide to the Galaxy had to contend with the infamous Nutri-Matic machine to obtain his cup of tea. Rather than storing the molecular specification in advance, the Nutri-Matic attempted to personalize Arthurs beverage by performing a spectroscopic analysis of his metabolism, and then probing the taste centers in his brain. As with Captain Picards tea, Arthurs drink is synthesized by assembling its molecular constituents. However, in the Nutri-Matics case the inevitable result was a plastic cup filled with a liquid that was almost, but not quite, entirely unlike tea.
None of this violates any physical laws, and in fact such atomic-scale programmable assembly is already possible in the lab today (as long as your tastes dont run to anything much larger than a few atoms).
To develop real working personal fabricators that can operate on a larger scale, my colleagues at MIT and I assembled an array of machines to make the machines that make machines. These tools used supersonic jets of water, or powerful lasers, or microscopic beams of atoms to makewell, almost anything. The problem we quickly ran into was that it would take a lifetime of classes for students to master all of the tools, and even then the students would get little practical experience in combining these tools to create complete working systems. So, we thought, why not offer a single-semester course that would provide a hands-on introduction to all the machines?
In 1998 we tried teaching How To Make (almost) Anything for the first time. The course was aimed at the small group of advanced students who would be using these tools in their research. Imagine our surprise, then, when a hundred or so students showed up for a class that could hold only ten. They werent the ones we expected, either; there were as many artists and architects as engineers. And student after student said something along the lines of All my life Ive been waiting to take a class like this, or Ill do anything to get into this class. Then theyd quietly ask, This seems to be too useful for a place like MITare you really allowed to teach it here?
Students dont usually behave that way. Something had to be wrong with this class, or with all the other classes I taught. I began to suspect the latter.
The overwhelming interest from students with relatively little technical experience (for MIT) was only the first surprise. The next was the reason why they wanted to take the class. Virtually no one was doing this for research. Instead, they were motivated by the desire to make things theyd always wanted, but that didnt exist. These ranged from practical (an alarm clock that needs to be wrestled into turning off), to fanciful (a Web browser for parrots), to profoundly quirky (a portable personal space for screaming). Their inspiration wasnt professional; it was personal. The goal was not to publish a paper, or file a patent, or market a product. Rather, their motivation was their own pleasure in making and using their inventions.
