A conversation with Arati Prabhakar, former DARPA director, on how to improve our R&D ecosystem
If anyone has examined our country’s innovation infrastructure from all sides, it is Arati Prabhakar.
The former DARPA director spent decades in both government and private sector roles. Early in her career, Prabhakar worked closely with semiconductor companies and researchers as a DARPA program manager and office director. As director of NIST, she sought to improve opportunities for a wide range of American companies through standards and advanced technology. She then moved into the private sector, first in executive roles in two companies and then as a venture capitalist. She returned to Washington to serve as DARPA’s director from 2012 until last year. Most recently she served as a fellow at the Center for Advanced Study in the Behavioral Sciences at Stanford University, studying how we shape our technologies and how they shape us.
During her William D. Carey Lecture at the AAAS Forum on Science & Technology Policy earlier this summer, Prabhakar focused her remarks on the need for new approaches and advances in the research and development ecosystem. We need these new approaches, she said, not just to keep the U.S. competitive, but to address basic human needs for health, security and equality in a quickly changing world.
Prabhakar remarked that the U.S. is great at basic scientific research. And our market economy is very good at applying talent and resources once a business opportunity is clear. But, she argued, we need ways to bridge these strengths, and she identified Cyclotron Road as one of the organizations helping promising scientists and engineers to directly commercialize their research insights as entrepreneurs.
We wanted, naturally, to delve more deeply into this topic, and were thrilled that Arati Prabhakar agreed to chat more about it. Below is a transcript of our conversation, edited for length and clarity:
Cyclotron Road: Arati, you’ve spent your career at the intersection of R&D, the private industrial sector, and government. And you’ve worked with researchers at the highest levels of science and technology, both in government and private sector. What factors do you think created that gap in the research ecosystem in the first place? And why do you think we can fix it now?
Arati Prabhakar: This is not something that is actually supposed to be easy. The more innovative a new research insight is — the more disruptive to how we think about doing things today — the harder it is to get accepted. The whole history of R&D has been exciting insights and then usually massive resistance from people about making the changes needed to use those new insights.
But we do have places where the whole system is working well today. Sometimes we have a fairly short time and modest cost — and therefore modest risk — between the research insight and the point where the market can say, “that’s a good bet.” When that early risk is manageable, as it often is when you’re dealing in software, that’s where things are actually working great.
I’m sitting in Palo Alto, where everybody thinks things are just hunky dory in the R&D world. But we are leaving a lot on the table because there are a lot of other areas where it’s not working as well.
Secondly, in a really well-meaning push, most research investors in government agencies have over-constrained their work in trying to avoid merely doing more of what the markets do anyway. As a result, I’d argue that federal agencies have contributed to too big a gap between the public role and where private capital can move forward.
CR: You’ve had a chance to visit Cyclotron Road and see what we’re doing in action. What do you find interesting or different about the entrepreneurial fellowship model we’ve developed?
AP: Cyclotron Road is one of the experiments underway that I find very very encouraging. I think it’s going right at this issue. The Cyclotron Road model recognizes that there are really smart people with a lot of passion to deeply understand research advances. But in the way they’ve been trained, there is no reason to expect they have the moves to go commercialize that technology, or know what the best commercialization path might be.
So, bringing them in and giving them access to laboratory facilities they need to do their work, and also immersing them in thinking about business — this is very different from just thinking about the research or the science.
I think that’s an incredibly effective model, so I’ve been a big fan of what Cyclotron Road is doing. One of the big strengths is that it recognizes that there’s not a single recipe to figure out how every new insight progresses. Helping these researchers become entrepreneurs and figure out what the path is for their particular advance…that is extremely powerful.
CR: Are there some institutional barriers — either structural or policy-based — within agencies that you think could be removed or changed in order to improve researchers’ access to government funding or other support systems?
AP: Yes, absolutely. And I think the way things are now is a natural progression of the priorities that we place. Take this specific example: you’re doing academic research in a particular field and can make a small amount of material and do some lab experiments that look pretty dazzling in terms of the material properties. But that does not tell you what a real product might be, how a customer might use it and what they would pay for it and what it would cost you to make.
That’s the vast pile of questions that have to get dealt with and you look around and you say, okay so how would I fund that? With NSF and NIH, the work that they fund reflects what is valued by the people that they fund. It is peer-review driven. So, you have a bunch of people who are in an incentive structure that rewards publication, citation, getting tenure in university, so that becomes a very good self-reinforcing system to accomplishing those things.
But none of those are incentives for learning, say, what is the process technology that goes with your marvelous new material or figuring out anything about what an actual product might look like. So, there is a mismatch between the next set of questions that need to get addressed and the incentives around which a lot of the federal research dollars are aligned.
I had the great fortune of working in one of the few places in government that doesn’t live in that peer-review model. At DARPA we often did work in universities but that wasn’t our mission per se. It was to demonstrate breakthrough technology for national security and as a consequence of that mission, if we saw something that didn’t fit with the university basic research model, we would go do it. Sometimes we’d do that with university researchers. Sometimes with companies large or small. Because the mission is so different, the work DARPA funded ended up being different. So that’s fantastic, but DARPA is about 2 percent of federal R&D spending. ARPA-e is also mission-driven, but it’s a 10th as big as DARPA. So that doesn’t really move the needle. These are some places where we’re doing the work that breaks out of the university mold a little bit but they are very small on the scale of the overall system.
CR: During your time in the private sector did you similarly see barriers that you think make it unduly hard for scientists and engineers to launch companies?
AP: Well the private sector is not very complicated at all, because companies invest in R&D based on the market and product opportunities they perceive. I hear a lot of people say I wish companies would do X or they should do Y…but if you live in a market economy and if you value the way the market works — and we are all surrounded by the blessings that have come from that — then you don’t actually get to say I wish companies did x, y, or z. The incentives that are set for them drive their behavior.
And they’re very good at allocating resources to the things that create profit opportunities. So, to me, private industry and how it funds research is not the issue… There are broader conversations about long and short-term incentives and executive pay. There certainly has been a shift over the past few decades of paying relatively more attention to the stock price and relatively less attention to everything else. So those are the broader issues but fundamentally I don’t think it’s a sensible expectation that companies will do something other than what is in their business interests.
If you look back to the Bell Labs days, compared to now, what we expect of companies has changed. Back then AT&T was a regulated monopoly and a slice of your phone bill went to paying for Bell Labs. And what is so interesting to me is that in the research community there is still this deep mourning about the loss of corporate research of that era — Bell Labs but also DuPont, General Electric, there were many others. But when I talk to people in the business world, there is zero mourning over the loss of that kind of corporate research. It didn’t actually contribute to the business’s success in the direct way that we now think of every penny having to do.
That’s just a fact of life, that if we have this global market economy that’s how companies are going to behave. There are companies today that are very highly profitable and have huge market power but also have enlightened, thoughtful leadership they are trying to figure out new models for research, and you can see that happening at Google among other places. But perhaps we mourn the loss of corporate research labs more than makes sense, given how the market works.
And I don’t think we celebrate enough the fact that universities are actually much better than they once were at commercializing technology. When I started as a young program manager at DARPA in 1986 it was hard to get university researchers to even talk to companies because they would talk about how doing so would taint them. That attitude is hard to find these days. And back when Bell Labs was flourishing, venture capital and entrepreneurship were barely bubbling.
The whole ecosystem needs to continue to change and experiment with new ways that could help us meet the future more effectively.
— -
Our sincere thanks to Arati Prabhakar for sharing her insights with us. If you’d like to view her William D. Carey Lecture, A Better Tomorrow: Renewing R&D’s Promise to America, from the AAAS Forum on Science & Technology Policy, it is archived here.
