Kalil on Google and America’s Innovation Policy
Tom Kalil has a nice column that explains the importance of federal support for fundamental research in the creation of Google (and makes the case that current US policy is hurting the environment that allows companies like Google to spawn and grow). The Google story is just one of the more recent examples of long-term, government-supported fundamental research helping develop and grow billion dollar industries and markets. It’s a story that has been repeated a lot in information technology. The National Academies Computer Science and Telecommunications Board even put together this somewhat hard-to-read graphic that shows 19 different IT-related technologies that, with government-support, each grew into billion dollar industries. (Note to self: redesign CSTB chart to make it clearer to read).
Kalil’s article notes some warning signs — we’re not producing enough students with science and engineering degrees, we’re relying too much on foreign students to fill the gap and tighter visa restrictions are affecting the supply, US share of publications in top science journals is declining — but he doesn’t delve into some of the specific causes, other than to note that in the President’s most recent budget “science funding in 21 of 24 science agencies would would be cut over the next five years…including NSF, NIH, and DOE Office of Science.” I’d add that I think the problems go beyond raw funding levels. I think we’re approaching the funding of fundamental research in a way different than in years past, especially in IT R&D, and especially at the Department of Defense. DOD and DARPA have always been crucially important to the development and advancement of computer science, and university researchers, in turn, have been crucially important to DOD and DARPA. However, changes in the way DARPA does business — from its moves to classify most of its computer security research, to its recent move to a ‘milestone’ based approach to funded research, where programs are evaluated on a 12 to 18 month cycle with ‘go/no go’ decisions at each step — have had the effect of discouraging university researchers from participating in DARPA-sponsored research. This is significant for a couple of reasons. First, it means some of the brightest minds in the country won’t/can’t work on DARPA’s important research problems. Second, it means university researchers have a hard time participating in maybe the most important aspect of DARPA-sponsored research, the community building around particular problems.
Computing research (and the country as a whole, I’d argue) has been well-served historically by having a two significant, diverse sources of funding in NSF and DARPA. NSF continues to be primarily a place for the single investigator — modest grants for small numbers of individual researchers. DARPA’s real strength historically, however, was different. DARPA program managers could identify a particular problem, then bring together and nurture communities of really smart people devoted to working on the problem. It was a very successful approach — DARPA is credited with between a third and a half of all the major innovations in computer science and technology (according to Michael Dertouzos). Between the two of them, the NSF and DARPA models have led to everything from graphical user interfaces, the Internet, and, well, Google.
So it concerns me that DARPA’s is discouraging (intentionally or unintentionally) university-based researchers from participating in their programs…maybe even more than the declining share of basic research in the DOD science and technology portfolio concerns me. And I think Kalil is right to be concerned with what we may reap in the future as a result of these policies today.