Acquisition Talk, Ewoks, and more
A recent podcast, Star Wars mission creep, and other news
The Nexus Newsletter
Welcome back to The Nexus Newsletter.
Read this week’s newsletter to hear from our Head of Government in a new podcast, how mission creep doomed the Empire’s AT-ST walkers, recommendations to bring the acquisition process into the 21st century, and other news from the nexus of autonomy and national security.
From our Slack
Look, we know that there’s work to do - Congress needs to fund the government, DOD needs to accelerate innovation, etc. - but it’s also the holiday season. So, before we dive into the serious stuff, let’s start with a few lighthearted pieces causing a buzz in our company Slack channels…
As anyone with experience working in or with the DOD knows, bureaucracy, program compromises, and mission creep result in underperforming platforms, bloated programs, and disastrous outcomes for operators. That is true in Washington, D.C., and, as it turns out, the forest moon of Endor. In this Twitter thread, Marque VC’s Jake Chapman explains how compromises and mission creep made the Empire’s AT-ST walkers vulnerable to our favorite diminutive furry bipeds from the Star Wars universe: the Ewoks.
Developing autonomous systems can be expensive. A major benefit of our development platform is how it enables faster, cheaper, and safer development and deployment of robotic systems for a variety of purposes, from defense to agriculture. Still, developing new capabilities requires significant financial investment and adequate time for virtual and real-world testing to validate performance. As the Army faces tightening budgets, lack of training data, and lengthy development timelines, Duffel Blog writes that the service is evaluating ways to replace their exquisite robotic platforms with a notably lower-tech approach: “Human-Marine teaming.”
Podcast: In the fight – Scaling AI/ML in defense with Colin Carroll
Colin Carroll, Head of Government at Applied, recently joined Eric Lofgren’s Acquisition Talk podcast for a freewheeling discussion on acquiring AI/ML-enabled capabilities, a “Manhattan Project” for autonomy, Project Maven, the Joint AI Center (JAIC), the importance of data to developing autonomous systems, and much more.
On the role of data in realizing autonomy: Program offices have a requirement to deliver some kind of unmanned capability — autonomous, semi-autonomous — this decade to a war fighter where they have zero data. They’ve not collected any data from any of the current, enduring, or legacy systems that are out driving, flying, or sailing. No one spent money to do that.
On splitting hardware and software acquisitions: Now you have to merge those things together. Most of these hardware platform providers are not super accustomed to, or probably willing to integrate, someone else’s software into their vehicle. And the government then becomes the systems integrator. But the people that have designed these programs will all have moved on two years from now.
News we’re reading
Autonomous systems are gaining momentum in the national security space. Below, we’ve pulled key quotes from recent articles of interest, plus brief commentary from Applied Intuition’s government team:
RealClearDefense | The 1960s had Their Day: Changing DoD’s Acquisition Processes and Structures
Key quote: In contrast to DoD’s stove piped and often conflicting objectives, the typical commercial development process starts by the designated CFT defining and agreeing upon objectives. The CFT development model places a premium on getting end-user feedback early and throughout the process by building an initial viable product, getting input/recommended changes from the end user, and incorporating that feedback into the next design. The design is rapidly iterated to meet end user requirements. Within the DoD linear acquisition model, end user participation is minimal, leading to a disconnect between the developing needs of the end user and the delivered capability. Due to the lack of current end user feedback and lengthy acquisition timelines, too often DoD delivers capabilities that do not match current end user needs. In the CFT model, development issues are worked across functional areas by stakeholders concurrently rather than sequentially. New ideas and competitive inputs to a solution are ongoing and iterative. The result is a vast increase in the SPEED of development and a solution that better meets the needs of the end user.
DoD’s experimentation, requirements writing, and requirements fulfillment functions and institutions are entrenched and resistant to change. How can the DoD move to the proven commercial CFT development model? A path forward that offers a minimum of disruption to DoD organizations is to start small by taking two or three high priority problem sets that have existing funding and a transition (fielding) plan. Pull personnel and resources from the three functional areas — experimentation, requirements writing and requirements fulfillment — and combine them into a CFT. Designate a product manager (PM) that has the responsibility, the authority and funding to acquire and sustain the needed capability.
Our take: The current sequential role that requirements plays in the acquisitions process is an oft-overlooked obstacle to acceleration, and the organizational separation of requirements from acquisitions tends to breed stovepiped thinking and the delivery of capability that answers requirements that have long since shifted. Mike Brown and Col. McHenry’s recommendation that the DOD abandon linear product development and acquisition processes and move toward a development process that incorporates constant feedback from the end user rings true. There’s a lot of energy devoted to excoriating the acquisitions side of the process over its long timelines - and rightly so - but highlighting the lack of warfighter feedback during the development process for new capabilities is equally important. The Army has experimented with CFTs over the last few years for some of its major functional modernization efforts; these are not specific to a product but rather a group of similar products, like “ground vehicles.” We think that one downside of this current CFT model is that it enables tens or even hundreds of hands in the cookie jar, each with a potential opinion or agenda - and rule by committee has never allowed acceleration in DOD. Shifting the CFT model to be specific to the product, with absolute control by the Product Manager who is responsible to deliver real capability to the warfighter, is an experiment that the Services should give a chance.
The Atlantic | How to Stop the Next World War
Key quote: Today’s challenges require a new offset strategy—what we call Offset-X. It is not a war plan, nor is it comprehensive or definitive. But we believe that, if pursued as a competitive strategy, Offset-X will lay the groundwork for the U.S. to restore the technological superiority of its military over all potential adversaries. The strategy has three goals, each of which can help deter war in the future. First, to invalidate the investments that China has made to defeat the U.S. military. Second, to generate new capabilities that will increase the political and economic cost of war for China while reducing the cost of war for the United States and our allies. And third, to inject uncertainty into the PLA’s planning by giving the U.S. military a range of options for how to respond to a potential conflict.
Offset-X includes 10 specific initiatives, three of which deserve priority. First, the U.S. military must embrace fighting as a distributed, network-based force. These forces are geographically spread out, and their decision making and operations are decentralized. [...] Deploying such forces in the Western Pacific would allow us to restrict the PLA’s maneuvering space and continue fighting even if we suffer losses or our communication systems are destroyed.
Second, we recommend that the U.S. military fully integrate human-machine collaboration and combat teaming into all of its operations, be they planning and decision making or battlefield combat. When humans and machines form interdependent teams, they can outperform both the best humans and the best machines, capitalizing on their respective strengths and compensating for one another’s weaknesses. Employing them will help U.S. forces penetrate dangerous environments, such as the heavily guarded East and South China Seas, with less risk to human life. Machines can also serve as the “eyes and ears” of their human teammates, particularly in urban environments.
Third, the U.S. military should prioritize the embedding and use of innovative software in all future decision aids, combat systems, and operations. Software is now integral to every component of military decision making, whether sensing a target, selecting a weapon, conducting a strike, or assessing damage inflicted or incurred in battle. Successfully deploying and updating software will determine whether the U.S. military can outthink, outperform, and outgun the PLA.
Our take: Restoring technological superiority is essential to detering and, if necessary, prevailing in a future fight with a peer or near-peer adversary. We agree with the priorities outlined here - networked forces, greater human-machine collaboration, and software-defined capabilities will prove essential in a high-end fight with a capable adversary like China. We are proud to have engaged with the Special Competitive Studies Project (SCSP) to inform their earlier report, “Mid-Decade Challenges to National Competitiveness,” and are thrilled to be a part of this ongoing national effort to develop a strategy for tech-centered national competitiveness.
Read SCSP's Interim Panel Report (IPR), “The Future of Conflict and the New Requirements of Defense,” and their earlier report, "Mid-Decade Challenges to National Competitiveness."
The Economist | Ukrainian ingenuity is ushering in a new form of warfare at sea
Key quote: Uncrewed surface vessels thus seem poised to follow the trajectory of airborne commercial drones, which caught governments flat-footed when they went from hobby-shop curiosity to deadly security threat seemingly overnight. [...]
Not having a crew gives USVs other advantages. With no need for a cabin, they can be built for stealth. The Ukrainian boat rises only a few centimetres above the water’s surface, making it almost invisible to radar and cameras—but, unlike a submarine drone, still able to keep in radio contact with its controllers. (Radio waves cannot penetrate water.) This does not mean a follow-up could not dive completely underwater, for example in order to evade detection on a final attack run, like a German u-boat. The Hamas subs, which are guided on the surface by gps, might already operate on a similar principle.
Skipping the crew also means vessels can be used more brazenly. A group planning a kamikaze-style maritime attack can avoid relying on a human “who may lose their nerve at the last second”, as Scott Savitz, a senior engineer at the RAND Corporation, puts it. In the footage from the attack on Sevastopol, the drone charges through a hail of gunfire with gay abandon.
Our take: Ukraine’s combined USV and UAS attack on Russia’s Black Sea fleet continues to make waves. The low cost associated with newer unmanned and autonomous capabilities means that we can expect them to proliferate across domains, including maritime, and actors, including non-state actors. Building up our arsenal of autonomous surface and subsurface capabilities - and testing concepts of operation for those capabilities, in line with TF 59’s work in Bahrain - will be essential to success in the maritime domain moving forward, as will understanding ways to counter those capabilities when they are wielded by our adversaries.
Breaking Defense | 3 keys to bridging ‘valley of death’: Get involved early, cut out middlemen and pray Congress works
Key quote: “So you can have something that the chief wants that will solve the problem, but if you don’t have a way to sell it in a way that the government can buy it — and then the challenge is not, you know, in many cases the buyer isn’t the customer,” former Navy acquisition executive James Guerts said. Guerts was speaking at a panel here dedicated to overcoming the valley of death, and he zeroed in on cutting out some middlemen.
“And so I think where the services are really working hard is to compress that, so the buyer and the customer are more closely aligned,” [James Geurts] said. “But you’ve got to think about both sides of that. Because here’s why: If you have awesome tech and there’s no budget line for it, you know, it’s going to take you a while to get money for that.” [...]
Once the military knows what it wants, Maj. Gen. Heather Pringle, commander of the Air Force Research Lab, said one key to a smooth process is getting involved on the commercial side early.
She noted that a digital test environment at Eglin Air Force Base in Florida, for instance, allows industry to “basically compete different autonomy architectures in a military scenario, get feedback on that, and then go back and make those adaptations and come back again, and see what the results are.”
Our take: Removing middlemen and ensuring that the buyer and customer are more closely aligned during the entire acquisition process, from request through production, is essential. We’re thrilled to have James “Hondo” Geurts as a new member of our National Security and Defense Advisory Board. We’re also excited by Maj. Gen. Pringle’s comments - using a virtual testing environment to evaluate Air Force autonomy stacks, make stack updates, and evaluate those updates through scaled testing represents an understanding of autonomy development best practices, and is the capability that we are delivering to the Army’s RCV program. Virtual testing is central to autonomy development, no matter the use case or domain, so we’re excited to hear that AFRL is seeing its benefits.
Thank you for reading The Nexus Newsletter. Subscribe for more news from the nexus of national security and autonomy.