DARPA has a mixed record on transitioning

Abstractly ARPA model’s design has the explicit intention that technologies get out of the lab and into the real world. Clearly DARPA has successfully transitioned paradigm-changing technology to the government, large companies, and startups, so the model’s design isn’t worthless. However, many DARPA technologies still fall into the The (idea) valley of death and even successful transitions often are not smooth.

DARPA’s history pockmarked with stories that go “and then at the end of the program, the military or industry said ‘that’s cool, but we like the way we do it now.’” Or perhaps worse, they say ‘that looks great’ but completely stop working on scaling up the technology after taking ownership. In some of these cases the outside organizations come around: DARPA funded the development of UAVs in the 80’s, the Navy took on and then killed the program, DARPA continued development until the military paradigm shifted in the 90’s. The story is similar for optoelectronics. While counterfactuals are hard, these common ‘near death’ stories suggest that there are many programs that vaporized on impact with existing paradigms.

ARPA-E and DARPA - Applying the DARPA Model to Energy Innovation cites some of theDARPA commercialization failures:
- High definition display program
- Thinking Machine - Connection machine parallel processing computer
- Gallium Arsenide computer chips

Each transition failure, like unhappy families, happens is unique. Some of the failure modes do rhyme:

  • The technology is great, but it doesn’t slot well into a product line at a big company and venture capitalists can’t see how it could lead to a massively valued company in less than a decade.
  • The people who worked on the technology are great researchers but terrible entrepreneurs. The skills that make you good at seeking+executing on government projects can be quite disjoint from the skills that make you good at building and selling a product.
  • The pieces of the program don’t come together in one place. Each project in a program is a piece of a puzzle and most of the time they’re far more valuable together than on their own. For maximum impact those pieces eventually need a single home after the program ends. Personal computing was such a success story in part because as ARPA took its foot off of the gas pedal, Richard Taylor was able to bring the pieces in-house at PARC. Speculatively, bringing the pieces together may have become harder because the Bayh-Dole act enabled universities to extract licensing fees on research they created with government money, potentially bleeding dry anybody who wants to license from multiple universities who each want their cut. The relationship of government, academia, and technology fundamentally changed in 1980 with the Bayh-Dole act
  • DARPA works where markets dont - markets defined broadly to encompass military acquisition, (if there were market forces driving a technology forward, why would DARPA need to work on it?) A program can fail to address the core reason why there’s no market. In a way, DARPA programs are like cave dives - you want to get somewhere that requires you to decouple from safety/the market but you need to end up at somewhere with oxygen/demand eventually. Decoupling from market discipline is like cave diving. You can miscalculate where you need to end up or the target can move.
  • Often, DARPA works on programs that go against the established paradigm and changing paradigms might endanger someone’s job or just be more work than any advocate is willing to do.

Many of these failure modes illustrate the tension between ‘building something people want’ and ‘building something capable of shifting paradigms.’

Transitions are an area where an organization riffing on DARPA may be able to make big improvements on the original. It is important to first acknowledge that DARPA doesn’t exist in a vacuum. The mechanisms at its disposal are subject to the constraints of the world around it. In the 1970’s that meant large corporations like Xerox, defense contractors, or government labs. The constraints on those organizations have shifted, and VC-backed startups have become an important mechanism as well. DARPA has worked to adjust to this new environment - going so far as to create a commercialization team in 2017, but from person experience the adjustment has been clunky. I don’t have a great answer yet, but it’s important to ask “what would an organization built from the ground up with transitions into today (or tomorrow’s) environment in mind look like?”

  • DARPA isn’t a full vertical technology machine (as opposed to something like Bell Labs didn’t have a unique model.) DARPA doesn’t actually own the technology so they can’t commercialize anything themselves.

DARPA has a 5-10 percent program success rate - so programs that transition are outliers in and of themselves.


Web URL for this note

Comment on this note