In case you missed it, the Thiel Foundation, founded by Peter Thiel of PayPal fame, just announced the first batch of winners for its "20 Under 20" fellowship program to pay students to drop out of college for two years to pursue entrepreneurial ideas.
Any number of people have criticized this venture for encouraging kids to drop out of college. Personally, while I think it would be terrible thing if this sort of thing becomes too hip and causes deluded average kids who think the next Google is locked inside their heads to drop out and ruin their educations, it appears that the group chosen is sufficiently elite that a) they may well actually do something useful and b) will succeed no matter what.
Nonetheless, I still think this program is a bad idea, and unlikely to achieve its intended purpose of enhancing innovation in the U.S. (My apologies to the foundation if its purpose is otherwise.) Why? Because its essential strategy consists in smoothing the path of individual geniuses, and this is simply not where the bottleneck to innovation lies in America today.
It's easy to be distracted by the glamorous entrepreneurs who appear on the covers of business magazines into thinking that they are the sole essence of innovation. Obviously, what they do is important, and I hope they continue to do it. But unfortunately, individual technological entrepreneurship is only the third stage of an innovation "pipeline" whose stages are, roughly:
- Advances in basic science.
- Advances in infratechnologies. (I'll explain what these are in a minute.)
- Development and sale of new technologies by entrepreneurs.
In the U.S. today, the principal bottleneck in the pipeline is stage two, not three.
What are infratechnologies? They are the crucial, but unpatentable and thus unprofitable, advances that must take place before salable new technologies can be developed. Because they are unpatentable (or if patentable, infeasible to commercialize directly for other reasons) private industry has little interest in developing them.
Because they are not pure science, academic science funded by the National Science Foundation isn't that interested either.
America simply doesn't have a bottleneck at stage three. Our culture and institutions are friendly to for-profit business generally, and entrepreneurship especially, in just about every meaningful way. (Granted, one can quibble about imperfections, but by any reasonable international comparison, we're just about as friendly as one can get, pace politically motivated whining.)
Similarly, America is probably not where it should be with regards to basic science, but we still lead the world. So there's not really a bottleneck there, either.
To understand infratechnologies, let's take an example reported by Gregory Tassey, the economist at the National Institute of Standards and Technologies (NIST) who is America's foremost expert on them. In his book The Technology Imperative, he writes:
Measurement-related infratechnologies are a prominent example. For example, a fundamental measurement method called isotope dilution mass spectrometry (IDMS), developed by NIST scientists and others, led to infratechnologies and associated standards in such diverse applications as measurement of sulfur in fossil fuels for compliance with environmental regulations and test methods for cholesterol and other blood elements.
America's investment in such things for civilian purposes, despite exceptions like NIST, the National Institutes of Health, and the Advanced Research Projects Agency - Energy, is relatively small, and far below what it should be.
What happens when infratechnologies don't get enough development? Tassey provides a case study:
In 1982, Dr Ronald Levy and colleagues at Stanford University succeeded in treating a chemotherapy-refractory patient with low-grade follicular lymphoma by using high doses of MABs [monoclonal antibodies]. This initial success created hopes that the "magic bullet" against cancer had been found.
However, subsequent efforts at developing therapeutic MABs for various cancers failed. The problem was that the generic mechanism of action was not adequately understood. Many guesses were made in order to rationalize proceeding with drug candidate development. For example, some researchers thought that the monoclonal antibody somehow activated the patient's immune system because successful treatment provided protection long after the antibody was eliminated from the patient, but no proof of this conjecture was developed.
Such guesses were forced by the fact that the generic technology of MABs was only vaguely understood. Without the underlying technology platform in place, subsequent drug development efforts failed. The result of multiple failures was that both companies and investors lost interest in MABs as a promising therapy. The risk of further failure for additional drug candidates was prohibitively high and therefore they became unattractive candidates for venture capital...
Approximately 10 additional years of government funding by NIH [National Institutes of Health] were required to eventually advance the generic technology to the point that once again private capital was induced to flow into antibody drug development.
Multiply this bioscience example across all our emerging technologies, from nanotech to biofuels, and you have the true bottlenecks to American innovation today. And this is a problem that nurturing young geniuses -- whatever else that may accomplish -- has very little to do with.
Infratechnologies are, in the language of economics, quasi-public goods. That is, they fall into a difficult category between pure public goods like, say, national security, which it is impossible for any one individual to appropriate, and the pure private goods that are the province of ordinary profit-seeking businesses. Because they are partly public goods, there is a legitimate argument for -- horrors! -- big government to be involved in supplying them.
This fact tends to drive Silicon Valley libertarians crazy. But if they would only be honest with themselves about the ultimate bases of their own fortunes, they would see it quite clearly.