Roper’s vision for digital acquisition: The tech stack, the edge, and digital engineering

Getting a lot of play is Air Force acquisition executive Will Roper’s article, “Take the Red Pill: The New Digital Acquisition Reality.” Some of that interest might not be what you think. The “red pill” is a concept used by people in the manosphere which is in strong opposition to feminism. It includes all sorts of lingo that I had never heard before, and is indeed quite disgusting. Now, Roper was likely as ignorant of all this as I was, but it will be interesting to see if the fallout tarnishes Roper or accelerates his exit from Pentagon leadership. That would be an unfortunate outcome for someone who has been a national asset.

I’d just like to list a few things in the Air Force that probably would not have happened (at least as we know them) if it weren’t for Roper: (1) digital century series concept and the next generation air dominance program — there wouldn’t be that rapidly demoed fighter jet; (2) Agility Prime effort for eVTOL; (3) Advanced Battle Management System; (4) the scaling of Kessel Run idea into numerous Software Factories, along with Nicholas Chaillan’s DevSecOps CI/CD pipeline; (5) AFVentures and the STRATFI program; (6) emphasis on budget reform in Space Force acquisition report as well as creating five space PEOs where there was just one. I’m sure there’s several things I’m leaving out, but that’s really impressive.

OK, so back to Roper’s article. He focuses on three principles, and I’ll give my thoughts on each:

(1) Own the tech stack.

Yes! This is something CSO Chaillan has been working toward, reversing 70 years of practice to outsource program stovepipes to industry leading to a lack of interoperability. Three things: (a) the government no longer has in-house talent to do a lot of this work. Software factories are trying to build that up, so we will see how that works out. (b) Should it really be the “whole” tech stack? A lot of the infrastructure doesn’t need to be government owned or operated… e.g., cloud. JEDI has been a nightmare. (c) There shouldn’t be one tech stack. Let the services compete and overlap and experiment.

(2) Push to the edge.

Not much to report here. This kind of real-time focus on DevSecOps deployment is definitely important. I guess my one add would be the importance of cybersecurity in this environment, and the usefulness of blockchain.

(3) Digital engineering.

I think digital engineering works well for incremental change, e.g., like NGAD, GBSD, B-52 engines, etc. The Space Force also says it wants to digitally engineer everything before prototyping. But did SpaceX digitally engineer the Falcon or Starlink?

I really worry about creating this systems analysis barrier to experimentation, particularly on frontier deep tech like hypersonics and hyper-maneuverability. We don’t have models to predict that perfectly due to resonance between frequencies. DoD must invest in enabling tools (like wind tunnels) and doing small, iterative, rapid real-world experiments.

This is a huge lesson from ballistic missiles in the 1950s. Systems analysis didn’t work for the Air Force Atlas program. Von Braun and the Army solved many problems by trial-and-error. For example, on reentry, no one could have predicted ablation. The Army just put every material and shape into the blast of an engine. So, does the DoD shut itself off to 1,000x gains by forcing digital engineering?

[Update: Roper released another paper, There Is No Spoon.]

1 Trackback / Pingback

  1. MSBAI GURU Community Crowdfunding updates - 101 Incubator

Leave a Reply