This week’s issue of Science profiles the Defense Advanced Research Projects Agency (DARPA) in Arlington, Virginia, the U.S. military’s renowned research and technology development arm, through the eyes of Benjamin Mann, a former program manager. The agency enjoys a stellar reputation as a potent source of technological innovation for the U.S. military, in large part thanks to the unique degree of autonomy, flexibility, and authority given DARPA program managers, who serve relatively brief tenures.
The three stories below examine facets of DARPA discussed only briefly in our magazine story. One presents the exceptions to the rule that most DARPA program managers spend just a few years at the agency. Another reveals the sudden demise of an outside expert panel that advised the agency for most of its existence. The third highlights a rarity in Washington, D.C.: an agency basically satisfied with its annual budget.
Some make DARPA a career
One pillar of DARPA’s management philosophy is that program managers will come for a relatively short time—typically 4 to 5 years—work their tails off to accomplish great things, and then leave. Those short-term appointments, often in the form of a sabbatical from an academic position, pump fresh blood into the 58-year-old agency and keep it at the cutting edge, senior officials say.
But that rapid turnover doesn’t apply to everyone. Materials scientist William Coblenz left DARPA last fall after serving for 25 years as a program manager. He’s part of a generation of program managers who spent most of their careers at DARPA. And those exceptions to the up-and-out rule—including several senior officials with more than a dozen years at the agency—provide important insights into how the semisecretive agency does business.
In 1990 Coblenz was working for a defense contractor when a DARPA program manager called to say the agency “was looking for someone to manage a ceramics program.” The goal of the project was to replace the metal parts on the hottest part of a jet engine, just before the afterburner, with ceramic components. The goal was to manufacture a component that would last much longer and, thus, extend the time between overhauls. And that would save the U.S. Air Force a lot of money in the long run.
It wasn’t a huge technological leap, Coblenz admits. But Air Force officials were not eager to make the switch because converting to the new material—assuming it worked—would require a substantial up-front investment by industry and, thus, the government.
“I like to say you have to be either desperate or crazy to do an insertion like this,” Coblenz explains. “Desperate because you have to and there are no alternatives, or crazy because you’re just thinking how much better this will perform.”
Those barriers made it what Pentagon officials like to call a DARPA-hard problem. And there was already a precedent: DARPA had recently run a successful insertion program for computer chips, replacing silicon with gallium arsenide. The new material made them faster and more resistant to radiation, an important quality for space-based weapons systems. (The program manager on that project in the late 1980s happened to be Arati Prabhakar, the current DARPA director.)
So Coblenz signed on. But because of a recent change in government hiring rules aimed at avoiding conflicts of interest, Coblenz severed all ties with his employer and was brought on as a career civil servant. That set him apart from most of his colleagues, who had taken leave from their current job.
Overall, Coblenz recalls, “when I got hired in maybe one-third of the program managers were career civil service, one-third were [temporary], mostly from universities, and the rest were military billets [active duty service men and women who had been assigned to DARPA].”
Too many familiar faces
But that distribution was not ideal, says Larry Lynn, who became DARPA director in 1995. “I rejoined DARPA 10 years after I had left it [as deputy director],” Lynn recalled in 2006 for an oral history of the agency. “And the thing that startled me most was that I knew most of the people, which says that they had all been there for more than 10 years.” Like several other former DARPA directors, Lynn believes that “the health of DARPA depends on the ability to bring in fresh blood every year. So while 20% [staff] turnover is unhealthy in business, it’s very healthy for DARPA.”
Once he became director, Lynn moved quickly to formalize the unofficial rules favoring short-term appointments. He got a boost from a change in the federal law that allowed DARPA to bring in private sector scientists on temporary assignments. Although he couldn’t fire civil servants like Coblenz without cause, Lynn encouraged long-timers to leave, even to the point of peddling their resumés to prospective employers.
Coblenz says he was told about DARPA’s long-standing policy of short-term employment when he was hired and that he didn’t expect to be an exception. “But I ended up liking it,” he says.
He also felt that he was making important contributions to materials science, to the country’s manufacturing sector, and to national defense. His second program, on solid free-form manufacturing, was a precursor to today’s hot field of additive manufacturing. He even learned to take advantage of the bane of every federal program manager—a congressional spending directive from a legislator, steering funds to the interests of an important constituent—by turning one such earmark into a productive program to expand the use of ceramic bearings.
Another, smaller cohort at DARPA that comes close to matching Coblenz’s longevity is comprised of those senior officials who have moved up the ranks. Few such positions exist in DARPA’s horizontal management structure. But the agency seems to find room for a few people with long institutional memories.
Craig Fields, for example, joined DARPA in 1974 as a program manager and stayed until 1990, when he was ousted after less than 1 year as director. Jane (Xan) Alexander was hired as a program manager in 1989 and left in 2002 after serving as deputy director under Frank Fernandez and Tony Tether. Her successor, Robert Leheny, arrived in 1993 and left in 2009 shortly after Tether departed. The current deputy, Steven Walker, has been with DARPA since 2002 except for a 2-year leave to oversee Air Force science programs.
After spending more than 2 decades as a program manager in the defense sciences office, Coblenz was moved to a support office in 2014 in what he says was a not-so-subtle effect to push him out the door. “I had been there much longer than the time they felt was appropriate,” he says. Coblenz says a colleague who recently suffered the same fate may be the last career employee still at DARPA.
Although he has left DARPA, Coblenz has no plans to stop working, and says he has several irons in the fire. Meanwhile, a quarter-century at DARPA has shaped his approach to doing science. His business card reads: “Inventing the Future one Project @ a Time.”
DARPA kills long-time advisory council
Not long ago, DARPA quietly disbanded a group of prominent outside scientists who have helped the agency peer into the future for nearly a half-century. Members of the Defense Science Research Council (DSRC) say its disappearance deprives DARPA of fresh perspectives on hot new research areas that might interest the military. The decision, by DARPA Director Arati Prabhakar, also offers a glimpse into the peculiar relationship between the agency and the academic research community.
Members of the council say they were never given an explanation for its demise. “I have no insights into why we were terminated,” says the most recent chair, David Walt, a chemical biologist at Tufts University in Medford, Massachusetts. But, perhaps curiously, members also don’t seem to mind being kept in the dark. “DARPA has information that we don’t have,” says Hayden Wadley, a materials scientist at the University of Virginia in Charlottesville, who joined the council in 1996.
Wadley presented an informal history of the council at a November 2014 event that DARPA billed as a reunion of current and former council members. And even though the event was held after the agency decided to defund their work, Wadley left thinking that the council still had a future. “We are on standby, waiting for someone to ask us to work on something,” he says.
But that apparently is not the case. “Accelerating changes in the R&D ecosystem … led the director to conclude that the DSRC model was no longer well-matched to DARPA’s needs,” says Rick Weiss, director of strategic communications, responding to a query from ScienceInsider.
A unique chance to serve
Created in 1968 after DARPA had begun funding research that would lay the groundwork for a new field, the group was originally called the Materials Research Council. Its 20 or so members would assemble each summer for several weeks to present their findings on topics they had chosen to examine earlier in the year. The meetings were an intensive give-and-take between those prominent scientists and senior DARPA leadership. “For most members, this was their most significant outside scientific activity,” says Walt, who in 1998 co-founded Illumina, a genomics equipment company in San Diego, California.
The council slowly broadened its scope over the next few decades to cover more of the physical and computing sciences. But by the mid-1990s, the agency’s growing interest in biological technologies, including those targeted to medicine, bioterrorism, and genetic engineering, prompted outreach to a new constituency. “I had never even heard of DARPA, and I don’t think I was a cloistered scientist,” recalls Patrick Scannon, a founder and chief scientific officer of Xoma Corp., a biotech company in Berkeley, California, who joined the council in 1996 and later served as chair.
Scannon had been recruited by a DARPA program manager touring the country’s biotech hotbeds, and he was one of the few industry members on the council. “I thought that it was incredible to have a federal agency with the foresight to be interested in these problems,” says Scannon, who has also sat on and chaired major biodefense advisory boards for the departments of Defense and Homeland Security during the Bush and Obama administrations.
To make the case for the breadth—and prescience—of the council’s deliberations, Wadley ticks off a partial list of studies conducted in 1997. “Improving human performance, novel applications of VLSI [very large-scale integration of transistors on a computer chip], just-in-time electronics, multifunctional dynamic natural systems [involving biomaterials], uninhabited vehicles—and that was in 1997,” he recites. “Some of those were one-off things, while others, like improving human performance, ran on for several years as we would dig deeper and deeper into the topic.”
Tony Tether, who was DARPA’s director from 2001 to 2009, attended the summer sessions and seemed quite eager to hear what the council had to say, according to several members. “The enthusiasm began to wane under Regina [Dugan, who succeeded Tether], and the council never really recovered from that,” Wadley says. “But I don’t blame her. There are other organizations that can run quick studies for them.”
Walt takes issue with that last point. The council was never a job-shop, he insists. Unlike its more famous counterpart, JASON, which advises the government of sensitive military matters, the council didn’t address pressing problems facing the agency or answer questions about specific technologies. Rather, he says, its scientists tried to look into the unknown and then share with DARPA officials what they saw. “We never provided recommendations, and we weren’t asked for our advice,” Walt says.
The fact that members received an annual consulting fee—in the low five figures, according to Walt—allowed the council to attract younger scientists who might otherwise not have been able to afford time off from their career ladders. But the real reward, members say, was the opportunity to exercise their minds for a cause greater than themselves.
“Our job was to come up with ideas that related to the future needs of the nation,” Walt says. “Sometimes we got it right, and sometimes we got it wrong. But everybody left their egos at home, and there was no huffing and puffing up your own work.”
Budgets not a big deal at DARPA
DARPA directors seem to have a surprisingly blasé attitude toward their annual appropriation from Congress. “I never really felt constrained by money,” says Tony Tether, who led the agency from 2001 to 2009. “I was more constrained by ideas.” In fact, many former directors say that a significantly larger budget than the current $2.9 billion might even damage the agency’s ability to launch new projects and kill those that aren’t working.
“When an organization becomes bigger, it becomes more bureaucratic,” says Larry Lynn, DARPA’s director from 1995 to 1998, who says he successfully lobbied Congress to shrink his budget after the Clinton administration had boosted it to “dangerous levels” in 1993 and 1994 to finance the short-lived technology reinvestment program. By 1999 DARPA’s budget had returned to roughly its 1992 level of $2.5 billion after approaching $4 billion in 1994.
The 9/11 attacks sent DARPA’s stock soaring again. Tether says Defense Secretary Donald Rumsfeld floated a 5-year plan that would have boosted DARPA’s budget to $7 billion right after he arrived at DARPA, causing Tether to raise a red flag. “I said, ‘If you want to put it in, go right ahead. But don’t expect me to justify it. I don’t have the slightest idea how to do that.’” Tether says Rumsfeld laughed and said, ‘OK. We can always get it for you later.’”
One reason DARPA directors may feel less restricted by their annual budgets than the heads of other research agencies is the high turnover of projects. For most agencies, programs continue indefinitely even as individual grantees may come and go. But at DARPA, roughly one-fifth of its portfolio disappears each year, freeing up a sizeable pot of money for program managers to tap into.
DARPA directors have devised various ways to manage that constant flux. “I would bring all the office directors into the room for 2 days, and everybody got the chance to describe their programs,” explains Victor Reis, who served as director under President George H.W. Bush. “Then we rank-ordered them—do you want A or B, and on down the line. We were done within a couple of hours, and every office director got to vote on every program. And we left space for a mulligan.”
The ability to move money around so freely is a luxury not available to the heads of many federal agencies. But it can also have its downside—the risk that directors will get too far down into the weeds in reviewing every project. The problem was acute under Tether, who says he wears the label of “nanomanager” with pride.
“With only a quarter of the budget coming free each year, he was OK the first year,” says one source familiar with DARPA’s budgeting process. “But eventually it was the entire fricking budget. And you just don’t have the time to do that.
“So [Tether] underspent by several hundred millions of dollars every year,” explained the source, who requested anonymity. “And eventually the Pentagon and Congress noticed and started taking the money away.” DARPA’s budget took a steep downturn in the last few years of Tether’s tenure after having risen sharply in the aftermath of the 9/11 attacks.