Ryan “Stinger” Fishel joined me on the Acquisition Talk podcast to discuss the future of the fighter aircraft community. Ryan is an F-15E pilot for the United States Air Force and a friend whose perspective I’ve benefited from over the past year. The views Ryan expresses in this episode are his own. Throughout the podcast, we discuss a number of timely issues, including:
- Implications of the DARPA AlphaDogfight simulation
- Hot takes on the news of NGAD’s first demonstrator
- Why it’s important to think about self-contained logistics
- How operators fit into the acquisition process
- John Boyd’s legacy on tactics and the mental/moral aspects of war
In the episode, Ryan argues that there is plenty of room for artificial intelligence to automate the tasks of a fighter pilot. However, many aspects of the job seem to require an understanding of the context that’s beyond AI at this point. The battlefield can be a complex place with multiple actors of differing intentions. Escalation has to be managed carefully, as tactical decisions increasingly have operational or strategic implications.
We also discussed the Next Generation Air Dominance (NGAD) program. Recently, the Air Force announced that it built a new demonstrator aircraft in just one year as part of the program using digital engineering and mature subsystems. NGAD isn’t intended to result in just a single aircraft, but progress a family of systems that can be integrated onto a new platform every few years. It hearkens back to how military aircraft used to be developed in the 1940s and 1950s. Though NGAD is controversial, it intends to inject competition into the aerospace industry, which has been going through a long period of consolidation and stagnation.
Podcast annotations.
One of the harder concepts for me to wrap my mind around is how the process of competition generates information that would not otherwise be available. Of course, it would be pointless to run a competition if we know in advance who would win.
Yet the defense acquisition process was built around the idea that outcomes are knowable in advance of system development. The style of decision presumes the data is available in advance. So long as people took the time to articulate a design and check for mathematical/physical consistency, it could be built – much like engineering a bridge.
With this framework, the only purpose of competition is to make sure the incentives are right to adhere to the efficient plan – weed out the loafers. If government monitors contractor performance through data and reporting (e.g., EVMS), then this is a substitute for the regulation of market competition.
But competition does much more than that, or even find out what works. It generates outcomes that (1) couldn’t have been predicted before the start of tinkering and learning; and (2) can produce 10x or 100x improvements rather than in small increments.
I give that intro because I think the idea of competition is what drives a lot of Ryan’s thinking. Here’s Ryan:
“If you take 48 years and step back in time to World War I we went from fabric wing biplane to the Mach 3 SR-71. The primary driver of all of that was competition and a different structure to requirements and acquisition… I hope that NGAD offers this pathway to change that, focusing on iterative technology development and an array of technologies. I think we can gain more competitors and more importantly more second-tier contractors who can break through entry barriers that we have, like the security empire and the tyranny from requirements and waterfall budget processes, and most of all the kind of that tyranny from a big winner taking contracts that spans decades. If you put all of that on the canvas of great power competition, I like to say that if we’re not competitive internally we won’t be competitive externally.”
I think his framing is right. One could argue, however, that aircraft were just going through their natural tech s-curve, where progress starts out slow, hits an inflection point of rapid increase, then slows again. The problem for military aircraft by the 1970s or so became less of airframe and propulsion marriage, and more of electronics and software integration. Air Force acquisition chief Will Roper has taken this view a couple times. But if he thought progress in fighter aircraft was hunky dory — we’re underestimating the progress in system of systems integration — then why would he start the NGAD program?
Ultimately, Roper must think progress should have been much faster if the DoD were able to do competitive experimentation. Acquisition if rife with the paradigm of prediction and control. Roper’s philosophy seems to be more humble. That people have to be let to try things out, deliver, iterate, and improve. The military often doesn’t have the information available to write a requirement until you build it and show it to them. Then the requirement gets written really quickly. But that couldn’t have started without innovation that started essentially without permission.
Here’s Ryan expressing something I take to be similar:
“However, I think that we’ll have to see really what this NGAD comes out as, and what kind of technology that it represents. There’s the innovator’s dilemma: the market doesn’t really know what it wants until it actually gets it. If you survey the market right you may not get a consensus to what the next cool thing is until the market actually gets it. So I think that could be the potential with the NGAD. We’ll have to see what it actually really is before making an accurate assessment on where fighter acquisitions are going to go.”
And here is Ryan on a reason why competitive developments seem to make more sense now than in the 20th century:
“There’s a pretty good book that came out not too long ago Unscaled by Hemant Taneja. The idea was that modern technology advantage is going to go away from economies of scale and toward economies of unscale… One of the things that he argues is that artificial intelligence and advanced manufacturing –and then you have communication that can move large amounts of data – they are trending everything towards capturing specialty markets and being able to do so fairly affordably. Something that he calls unscaled. So here we are with companies like Uber, AirBNB, Warby Park, Stripe, the list goes on, and if you think about that principle how can be applied to defense, I think it’ll be revolutionary.”
Be sure to listen to the whole thing! Ryan gives his view of the F-15EX and how that will fit into the picture with the on-going F-35 production and now the new NGAD program. There’s a lot more, including this interesting bit:
For the f-86 in Korea, they had 250 pages of technical manuals that they had to know… Today there’s something like 8,000 pages that each F-15 pilot has to know about his jet and his weapons, so there there’s that natural, you know, ‘hey, I’ve been doing this well and it’s been working for me, I don’t really want to change it.”
I’d like to thank Ryan Fishel for joining me on the Acquisition Talk podcast. Here he is talking on the Middle East Brief podcast, including a lot more goodness on the AlphaDogfight, and on the War on the Rocks podcast. He has a couple articles there as well, “Lessons Learned from the Air War Against Islamic State,” and “Self-Defense and Strategic Directions in the Skies Over Syria.” You can find more about some of his references here, including Marshall Michel’s book Clashes about the Vietnam Air War, and the fighter ace Francis Gabreski.
Leave a Reply