| The words to content ratio is high here. Finally, after about 100 meters of semi-infinite scrolling, he says something important. "Why Does DARPA Work? " covers DARPA program managers extensively, but the key points are worth reiterating both as a refresher and to emphasize that empowered PMs are the core thing one should not mess with. "What about managing programs through a committee?" NO. "What if performers just submitted grants and coordinated among themselves?" NO. "What about a more rigorous approval process to make sure money isn't wasted?" NO. "What if people could be career program managers?" NO. You get the point." The point being that DARPA program managers are people who've done something good but are not primarily managers. Their role is somewhat like VCs, without the greed. The problems DARPA works on tend to be rather specific. Many reflect specific military goals. Here's the current project list.[1] Examples: Automated air combat. A transportable linear accelerator. "Persistent, wide-area surveillance of all UAS operating below 1,000 feet in a large city." "Gun-hard, high-bandwidth, high-dynamic-range, GPS-free navigation." Note how different this is from a list of start up companies. These are technical goals, not business goals. Most are hard engineering problems, not pure science. They reflect specific military problems. The author's article doesn't reflect that tight focus. DARPA has a customer - the US Department of Defense. The purpose of DARPA is to solve hard problems for DoD, sometimes if DoD doesn't know it has them yet, and sometimes because DoD has a big problem and needs it dealt with. A "private ARPA" as proposed by the author has no customer. That defocuses the organization. It's not clear what a "private ARPA" is for. Even after reading all that verbiage. Then there's the problem, who does the work? DARPA funding generally goes to small parts of companies that do major work for DoD, or who have expertise making some specific thing. Not post-doc researchers. Not startups. DARPA does not create organizations to work for them. They use little pieces of existing organizations. The paper looks almost entirely at US institutions. This needs more reach. China has been opening lots of research organizations. Some are boondoggles, some produce good results. Take a hard look at that. Look at what Korea is doing. Figure out why Japan's R&D stagnated. There are two good books about DARPA - "The Pentagon's Brain", and "The Imagineers of War". I've read the first, but not the second. The first gives a good sense of how the organization works. (It's been a long time, but I've worked on a DARPA program.) [1] https://www.darpa.mil/our-research |
This is a really important point. DARPA is one of the few funding organizations that can--and will--pay for experienced technical staff.
Most other funders like the NIH and NSF have an odd split. The principal investigator is rewarded for having a track record (e.g., a prof with a history of similar papers), but the bulk of the work is to be done by trainees. This is sometimes baked into the grant itself: involving students is a good way to meet the NSF's "Broader Impact" requirements. Other times, it's a de facto restriction based on the budget. The standard NIH grant, a modular R01, will barely stretch past the PI + one staff scientist; trainees are cheaper and can often be offloaded onto other fellowships or other training grants entirely. Training future scientists is obviously important, but we've also got to make good use of the ones we already have.
Other agencies are dropping that ball. The NIH has a staff scientist program, but it funds literally a few dozen people per year vs. thousands of trainees. Dramatically expanding this would, I think, produce better science AND ease the academic job market crunch at the same time.