Stephen Senturia | MIT | Center for Materials Science and Engineering | sensing devices

Like many other DARPA program managers, Dubois was recruited from industry, in his case Bell Labs. Others come from universities, national laboratories, and (a few from) the military. The more diverse the mix, he says, the bigger the rewards. "If you have people who are all raised from the same background," he says, "you tend to get a little stale."

NO ABSENT-MINDED PROFESSORS
DARPA's investment strategy has changed dramatically over the years. Once, the agency handed out big block grants to create new academic departments - like the Center for Materials Science and Engineering at MIT. Now, DARPA is tightly focused on tangible, fast results. To make sure there's a real market tugging at every project, and not just ivory-tower interest, DARPA encourages research teams drawn from both universities and industry. And it does intense research out in the field, talking to the military about what kind of killer apps - literally and figuratively - would make their jobs easier.

Once selected, the research teams start by defining their goals - so that their program managers can rapidly assess what worked or failed within a few years. Each team is given a definite end date, so the researchers know they have to deliver quickly. All DARPA's projects are typically forced to wrap within three to five years.

"You actually have to build something that works," says Stephen Senturia, now retired from teaching at MIT and an alumnus of DARPA's funding process. That can be a double-edged sword, Senturia is quick to add, but it persuaded him to assemble a team from private industry and academia to develop sensing devices that could be used on the battlefield to detect chemical weapons and protect troops.