I recommend this piece very strongly to anyone who is interested in rigorous innovation. It’s quite long, and a very exhilarating look at what it takes to build an industrial research program on part with DARPA. To my understanding there has been such an attempt in progressive politics - Mikey Dickerson’s New Data Project - which unfortunately faltered after a couple of years. I think there is more than enough reason to give this kind of idea a second run. But I believe that such an effort would require a tremendous amount of funding and some very deep thinking about organizational structure, along the lines given in this essay.

There is one point in this essay which strikes me as incorrect. The notion is that DARPA is historically unique because it is established to ward off existential threats to the United States. Furthermore, that there is really only one category of governmental activity which seeks to address existential threats - the Defense Department. There are no further threats to address, the thinking goes, and therefore there is no more room for government-funded DARPA clones. That’s a point I disagree with, because we are witnessing existential threats on a number of fronts today - especially, a collapse of civil-society support for democracy. This threat is so significant that it has nearly eclipsed normal political debate, to the point that there is a viable and rather strong authoritarian party with a very plausible path to tremendous power. If that’s not an existential threat I’m not sure what is! The threat of climate change is probably a close contender, and it’d be interesting to hear why ARPA-E has foundered so much through this lens.

The larger point is that there is probably ample room for building DARPA clones which might address some of these problems, especially because they are both existential threats and so heavily influenced by technology.