Share

First off, I am glad this article exists because it gives me an excuse to talk about something that is near and dear to me: mission command. The fundamental issue is that the author does not understand mission command. To him, mission command is simply centralized planning and decentralized execution against a commander’s intent. This is a fraction of what mission command means. Mission command is when a leader tells subordinates what to achieve but not how to achieve it. At its core, it is about making sure that subordinate leaders can make the best possible decisions in the absence of higher command. This seems like it should apply to agents, but it does not, because inherent in this relationship are risk acceptance and personal accountability. Those two things are intricately tied together. They must be. Owning the consequences is core to risk acceptance. For a subordinate commander, this could mean completely redesigning my mission, abandoning one set of actions to go another way because that gets me to the actual mission objective. In mission command, I must assume the risk of that decision, and if I am wrong, more than my life is on the line. An agent cannot own this risk because it cannot morally, professionally, or legally own the consequences. This is decentralized execution without decentralized accountability.

“Agents are imperfect and occasionally strange, but so were the airmen I commanded.” Agents are not airmen. We don’t built trust with them, but trust is one of the core elements of mission command. He is asking us to replace interpersonal trust with engineering assurance. The author wants to be able to open the black box and poke around. That is not trust, not in the human sense; that is attempting to ensure understanding, which, as we see from the Hugging Face incident, even the best in the world do not have. What he is describing is an automation bias that delegates responsibility to automation without the accountability.

This paper shows that the author has an odd risk identification process. He is using Chinese open-weight models and says, “being a Chinese model doesn’t make me trust or reject it—I understand its behavior enough to bound risk.” He spends time talking about abliteration of refusals and stripping models of refusal behavior. This brings up two issues. One is that refusal behavior in humans is a core part of mission command and the law of armed conflict. We expect humans to refuse unlawful orders. Agents do not have this expectation and, even if they did, machine refusal is not conscientious refusal of an order. Then the author admits his approach is not surgical. His interventions degrade constraint-following broadly and unpredictably. This means he is creating, and arguing for the Pentagon to create, agents that are unevaluated and have been modified in unpredictable ways from their prior evaluations. But that Commanders should have the ability to “understand its [the agent’s] behavior enough to bound risk.” That does not make sense to me.

The author notes DoDD 3000.09 excludes autonomous cyber capabilities, but we should run these pilots “anyway.” Operations in denied communications environments are the point where autonomous agents are the most dangerous. This is also the point where the author wants to use them the most. When humans are in denied environments, they still exist in a system that holds them accountable for their actions. In degraded and denied environments, agents cannot coordinate, you cannot update them, and you cannot recall them, and cannot remotely terminate them. This means commanders will have lost corrective control. Denied environments also mean guardrails reliant on external control vanish. This is where agents working under a perversion of mission command could do untold harm. Distinction and proportionality are part of a military action, even in cyberspace. The inability to control a system that can make consequential decisions based on mission parameters but cannot be held accountable for those decisions is a road to perdition.

He is right about one thing, and it is a big thing. A commander owns everything his forces do and fail to do. Similarly, commanders who use agents are responsible for everything they do and fail to do. This should not scare them from using them, but it should push them toward a deeper understanding of both the uses and limitations of agents in warfare. They have to understand that they are not delegating responsibility to agents.

Leave a comment

I still remember being in the gym when the first news about the World Trade Center started to tickle into my high school. I only had a vague sense of what the Trade Center was or why it mattered. By the next period, the TVs were out in the classrooms. Even in rural Alabama, we knew we were watching history. Before lunch, the office administrator called my brother and me to the office for a phone call. My dad told us he was being placed on orders. It started a year-long period of him being in the U.S. but constantly gone, providing security for arms shipments. The U.S. was afraid terrorists would attack the shipments. As a nation, we were starting to see ghosts. Even though I was a teen, I remember being angry that anyone would attack us. I was proud that our military quickly took over Afghanistan in November 2001. We showed that nations and individuals should think before messing with us. We were venting our righteous anger on the wicked. That pride swelled as, with my mom, I watched George W. Bush deliver, with all the eloquence he could muster, the State of the Union. I was glad to see him label Iran, Iraq, and North Korea as “The Axis of Evil”.

Subscribe now

Looking back, I have no idea why I was glad to see the Axis of Evil labeled. I did not know what it really meant. I was a teen, and not a particularly politically active teen. I played football, went to the movies, and spent too much time thinking about girls and not enough time thinking about anything else. By the next year, the U.S. was preparing to knock off the next on the Axis of Evil. We were going to make the world, and ourselves, more secure. This time, we would make Saddam Hussein cower in fear of our power.

Iraq was another quick victory. Our military showed once again that it was world-class and without peers. This time, though, the luster of the victory did not last. Iraq turned quickly. We had inadvertently released decades of hate, repression, and fear. We had not replaced Iraq’s government; we had just broken it and stood around looking at the broken pieces, wondering why everything was not fixing itself. I was still not politically active, but my dad was gone again, and I was becoming disenfranchised. My whole life up to that point, I had wanted to be in the military. It was at this point that I first said I wanted nothing to do with the military.

Not to fast-forward too much, but I did end up joining the Army. I did not join the Army because of something born out of pure patriotism. It was not memories of September 11th that drove me to raise my right hand. I did, and do, love America, but I was not blind to what it was by the time I joined. Politicians asked the Army I grew up in to do things that an Army cannot do.

The core objectives of Operation Enduring Freedom (OEF) were to destroy al-Qaeda, remove the Taliban, capture Taliban and al-Qaeda leadership, and prevent safe havens. These were difficult objectives, and we did not initially meet all of them, but they were objectives that a military could meet. By 2005, the objectives changed to dismantling remnants of al-Qaeda and the Taliban, building host-nation capacity, supporting democratic governance, and expanding stability. It became our job to “extend the reach of the central government.” The first objectives are clearly military in nature. They have distinct and observable end states that can be measured. The second set of objectives are completely open-ended. A commander could easily ask “building their capacity for what and for how long?” Those are objectives that invite mission creep.

I need to make it clear that building a host-nation’s capacity is a legitimate military mission. This occurs every day around the world. The U.S. provides training and helps nations improve their internal security via a mission called foreign internal defense (FID). FID is fundamentally different from creating a durable Afghan state capable of independently governing all regions. This was not only not a military mission but was an ahistorical mission. This was a nation that had never had a strong central government. For years, the king of Afghanistan was known as the “Mayor of Kabul.” We did not care about that history because we had a role in denying safe havens, so that meant we had to have friendly eyes and ears in every region. The objectives required the military to become all things to the Afghanis. They had to be security, provide infrastructure, settle disputes, and provide employment, all in a culture they did not understand.

Our military became obsessed at this time with counterinsurgency. Commanders put Eating Soup with a Knife and The Accidental Guerilla on every reading list. We were also becoming convinced that Afghanistan was nearing a conclusion. RAND, in a stunning bit of analysis, provided an assessment of Operation Enduring Freedom (OEF) in December 2005 that shows that the U.S. Military was already seeing the end of the war and preparing for accolades and lessons learned. As RAND writes, “In all, Operation Enduring Freedom earned far more deserved accolades than demerits.”

That ended up being the conclusion from one small part of the war. It was likely one of the last rosy retrospective analyses of the war. U.S. operations in Afghanistan stretched on until 2021 and concluded with the U.S. evacuating, much the same way it evacuated Vietnam.

The lesson of Afghanistan is not “don’t get involved in a land war in Asia.” Its core lesson is: do not ask your military to be more than a military. War is politics by other means. In 2001, al-Qaeda chose war with the U.S. We cannot relitigate our response. What we can do is, in the future, recognize when we reach the end of military utility and transition. This requires the military to know when to say, “we cannot do what you are asking.” That is not the military’s way. Soldiers are taught to soldier on, never give up, and never surrender. Our leaders, political and military, have to be able to see the forest for the trees and know that our impressive capabilities have limits. Eventually, force of arms can only go so far.

I ended up spending twenty years in the Army. The end of Afghanistan aligned with the twilight of my time in uniform. I was already seeing the Army, and the military, willing itself to forget the war. Units stopped discussing counterinsurgency and counterterrorism. Those missions once again became the purview of special operations. Large-scale combat operations were the future. The post-Vietnam cycle was repeating itself, and this time we would make sure that we did not sign up for too much. Fast-forward just a few years, and our political leaders are again asking our military to meet an objective the military might not be able to reach. Can we destroy support for the “axis of resistance” more fully than we did the “axis of evil?” Or are we forcing our military into mission creep to achieve something that was impossible from the start?

Leave a comment

The clearest case in point I can make is innovation. A few weeks ago, I was talking to a former Innovation Cell director at a three-star command. He told me, “I have no idea what innovation means, and I was an innovation director.” This is a man who oversaw not only the assessment of new technology but also helped secure funding to pilot new initiatives and pull new programs of record into existence. He was an acquisition professional who worked at a very capable special operations unit, charged with innovation as his mission and title, and he could not define what it meant. This might sound like a silly semantic diatribe against using imprecise words, but it goes to the point of theater rather than purpose.

A core idea in the military, and one that applies to the larger world, is that missions use action words to describe what a unit is doing. They are attacking, defending, securing, clearing, holding, and retreating. Time and space provide boundaries to these actions. How does one go about innovating? How do you “do innovation?” Stanford has a quick reference for types of innovation. They list seven types: product innovation, process innovation, disruptive innovation, continuous innovation, sustaining innovation, business innovation, and radical innovation. It is not entirely clear that many of these are different. What they do highlight is that we have gotten exceptionally good at naming ways to optimize our existing industrial systems. To do this, we have systematically stripped out the redundancy in systems necessary to give a system resilience.

For example, Stanford offers that process innovation is, “Making enhancements to methods or procedures is known as process innovation. This type of innovation aims to increase efficiency and quality while reducing the costs of production.” Business innovation is, “Business innovation features enterprises executing new ideas services, products, and offerings to enhance their bottom lines.” From a defense tech case, these definitions show that innovation is not good in and of itself. Its quality depends on what we are optimizing.

Innovation is not invention. That part should be clear and something that everyone can agree on. Inventions are something new; innovations create value by changing existing products. Sometimes innovations shift a unique invention into a mass-producible product. Sometimes they make a product cheaper, faster, or easier to produce. That is the purpose of innovation at the enterprise level: to increase efficiency and, for companies, to drive stakeholder value (read: profits). So, when defense tech companies say they are innovating, is this what they mean? Likely not. I imagine they are saying that they are inventing new products, which some are, or that they are innovating by making existing tech better, more scalable, or cheaper.

The issue here is that innovation is not something that is easy, occurs on a timeline, or can be reliably counted on. More importantly, unwinding optimization to build toward a more robust system takes time. Sublevel suppliers that would have existed if companies had not pruned more robust supply chains in the name of efficiency have died or moved on to new products as the primes streamlined their existing product lines. The issue we see now with quickly making more Long Range Anti-Ship Missiles (LRASM) is that Lockheed Martin optimized for the world it existed in. That was a world where American companies, defense included, focused on process and organizational innovation and efficiency, in the form of Lean Six Sigma and just-in-time supply, to protect profits and right-size. These systems all worked. They made U.S. businesses efficient and allowed companies to move up value chains and focus on higher-margin work. They also made them fragile. Now, we are innovating at this upper level, the top of the value chain, but resting on a fragile production base. I am not sure that you can innovate out of this. For the production of the highest-end weapon systems, we need to unwind some of the efficiency and expand the base supporting them so, in the end, we can make more critical systems. It’s easier to subtract in the name of efficiency, and call that innovation, than it is to add in the name of capability.

No one would blame a company CFO for looking at their company and saying, “we need to eliminate excess capacity, reduce inventory, consolidate our vendors, outsource all low-margin part fabrication, and move most of our capital to our high-margin activities.” That is a completely rational action, happens in every business, and runs directly counter to building a robust defense industrial base. and have produced two great pieces over the last two weeks that touch on how we are compounding this issue with our defense approach. The first, Are the Tech Bros Blockheads, makes an argument I have made in the past. Namely, there is a defense tech bubble centered around the increased number of new companies competing over the very small OTA portions of the defense budget. Those small sums, relative to the overall defense budget, mean that companies already have to be as efficient as possible to be profitable. Their second piece, out this week, is Headlines: Supply Chains Will Determine Dominance. It points to the start of an overreaction to years of innovating up the value chain but, in my mind, creates potential perverse incentives.

Deputy Secretary Feinberg signed a memo on 18 August directing supplier cost and pricing transparency at all tiers of the supply chain. The end of the memo states, “The Department does not intend to limit profitability when contractors realize efficiencies after negotiating a fair and reasonable price based on a realistic baseline.” Again, Matt and Pete have a good take that I agree with:

This could radically change the DoW-industry relationship by requiring extensive cost data at prime and sub level, regardless of contract type. The expanded expectation of direct access to industry ERP systems and the department establishing fair and reasonable margins will be… interesting. While the intent is to provide acquisition professionals better data to negotiate, it could have a chilling effects for many companies to enter or stay in the defense industry if they have viable commercial businesses.

Taken together, both the idea that Tech Bros are blockheads and the fact that the Department is now going to have access to all costs up and down the supply chain for all deals mean that innovation focused on process efficiency is going to remain a focus. This is how we got here, and we are building back in a focus on incentivizing efficiency, which normally means creating less robust systems.

In other news, we have a new entrant on the defense tech Takeability List: Blake Scholl at Boom Supersonic.

Since every weapon used by the U.S. military is funded by taxpayer dollars, I guess we can send out valor awards to everyone. Good job heroes.

Share Mind of Things

Martin was born in 1933 and grew up in a rural part of Ohio. He was the second son and third child in a family of eight children. His older brother left home at 17 to join the Navy and fight in the Pacific during World War II. He followed him as soon as he was able. He left home at 17 to enlist in the Army. The year was 1950, and the Korean War was threatening to pull the world back into a global conflict. The recruiters offered him the chance to go to officer candidate school upon enlistment because of his test scores, but he was afraid he was going to miss the Korean War. After basic training and infantry training, he volunteered for service in Korea, and thirty days later he was on a ship, heading first to Japan, then to Korea in 1951. He would spend nine of the next eighteen months on the front lines with the 5th Regimental Combat Team in Korea.

The 5th Regimental Combat Team had the thankless job of being an independent unit. Larger units used them to plug gaps in their lines. This was hard fighting and hard living in those trenches. Despite all the stories he would tell of his time in the military, he almost never talked about Korea, with rare exceptions. He told me how he used his helmet as a cistern so he could bathe himself once a week in the trenches. He talked about the cold. Those were the only stories that he told of that time.

After Korea, he lived in West Germany, Alaska, Hawaii, Fort Bliss, Texas, Bangkok, Thailand, and Fort Benning, Georgia. He was married to my grandmother for 50 years. They raised five children together. He went on to serve two tours in Vietnam. Before his second tour, he was pulled aside by his battalion commander in Alaska and told that he was going to officer candidate school (OCS). He was a 31-year-old first sergeant, well past the age of most soldiers going to OCS. He told me how the interaction with “the old man” went.

“First Sergeant, you are going to go to OCS.”

He replied, “No sir, I’m not.”

“First Sergeant, either you are going to OCS, or you are never leaving Alaska.”

“Then I guess I am going to OCS.”

He said, “Your grandmother hated Alaska, so I started the car, because it was still winter there, and let it run for three days as we packed to leave because I was afraid it would not start again if I turned it off.” Then they drove from Alaska to Fort Benning, Georgia.

During OCS, he was known as Ol’ Sarge because he was the oldest by more than a few years. Sometime after graduating, he was assigned as a Special Forces officer. I wish now that I had spent more time gathering his stories. I feel like I am trying to recall a dream the next morning when I search my mind for clues about his military life. In his personal life, he was a man of deep faith. He helped build the church I attended as a child, served on the board, helped with the Christmas pageant, and taught the men’s Sunday School classes. He was quick-witted until the very end; even at 93, he was able to tell jokes and make snappy comebacks.

He was, in many ways, the most constant person in my life for much of my childhood. He was a teacher and mentor. He always had a book or a story with some meaning that fit the situation I was in. Before I left for Iraq on my first deployment, he sat with me at his dining room table and said, “Do you know what a bear trap is for?”

“To catch a bear?” I replied with a question.

“That’s right. Do you know what a mousetrap is for?” he asked, looking right into my eyes, and I could tell he was serious.

“To catch a mouse,” I replied.

“Yes, so a booby trap is to catch a boob. Don’t be a boob, son. Don’t go over there and get caught in a trap for a boob.”

He was saying something to me that I didn’t fully appreciate at the time. He was worried. He could not just say that he was worried that his first grandchild could be hurt or killed; that would have put his worries on me, which is something he did not do. This was his way of showing his care and his worry. It was Ol’ Sarge talking to a young soldier.

Luckily, I proved not to be a boob and came back. He saw me through all the stages of my career. He was there when I graduated Airborne School and he pinned his jump wings on me. He saw my graduation from the Special Forces course and was still there to see me retire from the Army last year. I should have asked more questions along the way. I was lucky I had so much time with him.

Share

To start, the idea that the U.S. Army is unable to adopt drones for warfare is incorrect. In many ways, soldiers in the U.S. Army led the world into this new era of drone warfare. Special Forces teams, working with conventional artillery, aircraft, and HIMARS, employed drones extensively to direct artillery attacks, attack enemy positions, and target individuals. Below is a photo of one of my guys in Syria launching a PUMA that we would use to direct 120mm mortar and artillery fire on ISIS fighters and support Kurdish fighters. We were using these in the early and mid-2010s.

Second, and equally important, is the fact that 3rd Armored Brigade has not been engaged in four years of large-scale combat operations. The Ukrainians have. We, as a military, place a high premium on experience. Why would we expect an armored brigade, one that has not been in combat recently, to perform well against a force operating in the defense, which is the stronger position, and with a vast advantage in experience? We should not. What is telling, buried in the reports, is that the Army reran the training scenario, and with each iteration over the course of two weeks, the brigade got better. They were more efficient in employing their electronic warfare, they were more survivable, and they adapted to their new reality.

I have written about it before; I am a sucker for training. and I wrote a piece earlier this year titled We Love Training. One of my strongly held beliefs is that units should fail the majority of training scenarios. So, 3rd Armored Brigade losing, even losing handily, to a more seasoned, more experienced force in the defense is unsurprising. It is a story, but the story is that there are considerable lessons to be learned. Of course, 3rd Brigade lost; they should have. More importantly, they learned and showed the ability to learn and adapt quickly, even over a two-week period.

My real worry is not the Army’s capabilities or ability to adapt; it is the U.S.’s ability to produce the number of munitions, drones, and electronic warfare systems the military needs and to ramp up production rapidly in a conflict. We are already seeing struggles with producing top-end interceptors, air defense systems, and long-range standoff weapons. Are we going to find ourselves reliant on Chinese parts? If so, we may find ourselves in a best-case scenario where we are sending data to our number one strategic and geopolitical rival. In the worst case, being reliant on Chinese parts means our defense industrial base will not be there in the event of a conflict with China.

This is our real risk. It is not that we are going to be unprepared tactically; it’s that we are not going to have the industrial capacity because we have continually focused on the wrong problems. The problem is not training soldiers. The problem is producing the weapons and systems that we need for the next conflict. If we do not solve that problem, all the training in the world will not matter.

Subscribe now

Are we selecting the best? I would argue that, for the most part, we are, because part of being the best in a military sense is a desire to be in the military. This means that high performers in athletics, business, or academics who have no desire to serve cannot be the best. The job, like any career, takes desire. Are we educating them properly? This, I feel, is where we struggle, and there is a risk, on the geopolitical stage, of wasting the capability we recruit by not pushing their education hard and far enough.

The social media response may just be one of those vapid, tongue-in-cheek responses seeking laughs and reactions. I mean, clearly it has made me react, even if only to use it as a foil. But what if it is not an n=1? What if it is indicative of a systemic issue with professional military education and with the general academic caliber of our academies? SAASS’s critical competence is to encourage “students to rigorously test their own ideas, interrogate conventional wisdoms, and engage in serious introspection. Graduates are educated to avoid intellectual complacency and institutional bias in assessing and characterizing the contemporary security environment.” Well, not all graduates, it would seem.

In a world where we are already seeing our number-one geopolitical rival not only investing in building military capability but also drawing a larger number of high-caliber entrants to its ranks, we must critically analyze how we are educating our force. To make the comparison more direct, I am going to focus my discussion on initial officer training, selection, and education during college. Plenty of writers have published reams of pieces on how to improve the mid-career and later stages of professional military education, but the military builds on a foundation laid during college.

One of my past commanding officers once asked me how I would change officer education. My assumption is that he asked me because I would routinely challenge officers in our task force on their assumptions in a way that resonated with him. My answer was something like, “I don’t care as long as they get out of the way so I can do my job.” Honestly, as a young soldier and noncommissioned officer (NCO), I meant that with every ounce of my being. The officers I knew were nice guys, but they were barely older than me or, very quickly, were younger. To put it mildly, they did not know their ass from a hole in the ground.

Eventually, I left the ranks of the NCOs and joined the warrant officers, which is a pretend officer rank that no one knows what to do with because you know all the NCO tricks and have seen what the expectations of an officer are. But I did not change my opinion. Many of my fellow officers would spend a minimal amount of time in a tactical formation, then head off to staff and education billets for “broadening.” After those assignments, once they were majors, with about 10 years in the Army and four years of tactical time, they would head to ILE, where they would read and learn strategy or, in the case of our commenter, read someone else’s distilled reading of someone else’s ideas on strategy. I did not really understand any of this when he first asked for my thoughts on how to improve officer education.

Asked the same question now, I have a better idea of what I would say. I think that officers, at the start of their careers at West Point or Annapolis, or in the myriad ROTC programs around the country, should have greater technical literacy and should require officers to show proficiency in STEM fields. I do not think this goes as far as requiring ROTC and academy graduates to have a degree in engineering, math, or science degree, but it does mean that they need to have taken the courses, and actually taken then, which we will get to. This is not because I do not see worth in other fields. I am a firm believer in the power of the humanities, but I believe that to command and employ new and emerging technology, they need to have a baseline understanding of the science and ideas that underpin it, enough, at least, to ask hard questions. To ensure this, education is the best solution early.

This is not a radical idea. In fact, West Point was a world-class engineering school through the Civil War and into the 20th century. Prior to the Civil War, officers were taught engineering that allowed them to lead and build railroads and telegraphs. These were core capabilities that allowed armies to move and communicate quickly. Today, the telegraph and railroad are analogous to the combination of software, technology, and physical hardware that powers and runs our munitions, vehicles, and communications. Understanding this will prepare officers to think critically about claims in this space when they reach higher ranks and, importantly, will make them better program managers and portfolio leaders during the middle of their careers, when they are called to work in acquisition.

This goes directly to a point quoted from a Financial Times piece in her look at the dystopian new warfare developing today: “Although not all pentagon procurement officers (or indeed Pentagon leaders) know it yet, nobody is prepared, technically or psychologically, for a new era of intelligence, deadly machines. Warfare is on its way to being transformed beyond recognition.” I agree with this, and I think the way you prepare is by addressing the technical issues, which are real.

Warfare is changing, and because of that, the way we educate officers and leaders has to change to allow them to understand the direction and path technology is taking. Still, war, even drone war, is a human endeavor, and our education needs to be more than just engineering, of course, because officers need more than technical competence. Young aspiring officers should also have a good knowledge of military and geopolitical history. This could take the form of a minor in history, or it could be a required set of courses in history and world religions. This should be married with physical conditioning and an introduction to tactics.

However, I do not think that officers should be educating officers in tactics. That is not who does it at Ranger School, or at Small Unit Tactics in Special Forces, it is senior NCOs. Senior NCOs should also teach tactics to our up-and-coming officers. Not just any senior NCOs, but NCOs who are going to continue being leaders in the Army.

There are two negative traits in officers that I observed over my career that I think education can help push out. One, maybe the worst trait, is hubris. This is not being self-assured, you want self-assured people, but you do not want people who have an overblown sense of themselves and their capabilities. Young officers must know that they own all decisions, but they rely on their NCOs, who often have decades of experience in tactical settings, to inform those decisions. This would be a marked change from the West Point of the past, when the Army taught young officers that “enlisted men are stupid, but extremely cunning and sly, and bear considerable watching.”

The second negative trait I encountered over the course of my career is a lack of intellectual curiosity. This manifested itself most obviously in an unwillingness to read. An actual interaction I had more than once went something like this:

“What do you like to read?”

“I don’t really read outside of work.”

“What do you mean? Like, you don’t read anything?”

“No, I figure I spend so much of the day reading emails and stuff that when I am off, I just like to turn off and watch something.”

Don’t get me wrong: I have finished a long day of work and just wanted to sit down and veg for a bit, but I cannot imagine being so incurious about the world that I would not want to open a book, article, magazine, or anything else and try to learn something about a time, place, or technology. These were also men, and they were all men, who did not have any hobbies outside of running and going out for drinks. It sounds like college, only now, in their careers, the required assignments are emails and PowerPoints. This is not intellectual curiosity, and it is not healthy for the force to build officers who are not intellectually curious.

Not every Reserve Officer Training Corps (ROTC) program at every college across the country can meet these requirements because not all small liberal arts colleges can, or do, provide a STEM education. That is okay. These are steps toward the type of force that can work with the technology of the future while developing officers who have a grounding in history. Not every college will be able to meet them, that does not mean they are not strong colleges, just that they are not colleges meant to educate future military leaders.

One last thing I would address: ROTC and the military academies should be the last places where grade inflation and lax standards exist. wrote about how disappointing current students are in a recent Scriptorium Philosophia post. Aside from saying that most are functionally illiterate, which is incredibly sad but something I wrote about before, he said,

“This is not an educational system problem, this is a societal problem. What am I supposed to do? Keep standards high and fail them all? That’s not an option for untenured faculty who would like to keep their jobs. I’m a tenured full professor. I could probably get away with that for a while, but sooner or later the Dean’s going to bring me in for a sit-down. Plus, if we flunk out half the student body and drive the university into bankruptcy, all we’re doing is depriving the good students of an education.”

You know who should be able to sit through a 50-minute class and do all the reading? The next generation of officers who are going to hold people’s lives in their hands. If they cannot, we can find others who can. To be clear, I do not think this is something that is affecting the military academies to the same degree yet, but I think it clearly must exist inside our ROTC programs, which comprise 70 percent of the active-duty officer corps.

I want to see our officers and our military remain the best. Part of that is recognizing that we must be selective, and the traits we are selecting for are not just intelligence and physical prowess; they are motivation, humility, intellectual curiosity, and a willingness to make hard decisions and own the outcomes. We do that, at least in part, by providing officers with an education and not just credentials. Selecting better, training them earlier, and forcing them to learn hard subjects that allow them to understand the future of technology and strategy without needing someone else’s CliffsNotes is critical.

Share Mind of Things

Subscribe now

To underpin their argument, the authors point to Central Command (CENTCOM) which, in early 2025, was able to run its entire classified network on just four NVIDA H100 GPUs. They expanded this with 24 additional GPUs over the course of 2025. This pales in comparison to commercial data centers and, to the authors, shows a strategic limitation facing the U.S. military. Even if compute becomes strategically important, it is not the limiting factor in U.S. military power. As of today, we are facing shortages of strategic weapons, critical minerals, semiconductor production, skilled labor, and worries about all levels of supply and production in the defense industrial base. Each of these issues are more pressing and more present than constrained compute’s impact on U.S. military operations.

There is no risk inherent in a “lack of compute” for the Pentagon. First, the authors are basing their logical underpinning on a weak foundation. The foundation of their argument revolves around a future conflict where the U.S. needs vastly more compute than it currently has and has no means of access in the crisis. However, the U.S. is building massive volumes of compute. ’s most recent post, The Four Horsemen of the AI Bubble Apocalypse, lays out the direct counterargument to this point. The hyperscalers are taking on billions of dollars of debt, and spending all their warchests, on building compute access. The SCSP team acknowledges we are already seeing massive build out. Companies like Oracle and Amazon are leaders in this space and already serve government and military compute needs. So, it is not directly clear the Pentagon will not benefit from this compute buildout. More importantly, in a peer fight with the PRC where the U.S. is openly targeting China, which is a scenario the authors posit, the idea commercial compute will be sacrosanct and reserved for civilian use is wrong and it underpins their logic. We are not talking about after September 11th when President Bush said, “Do your business around the country. Fly and enjoy America’s great destination spots. Get down to Disney World in Florida. Take your families and enjoy life, the way we want it to be enjoyed.” We would be talking about a WWII-style full mobilization of the nation. The Pentagon does not need to own all the GPUs to be able to pull the compute necessary to support the strategic need. My argument is the SCSP is focusing on the wrong kind of compute the pentagon need to build strategic depth in.

Ukraine shows us electronic warfare and challenges to the electromagnetic spectrum are going to be realities in future conflict. In the last several weeks we have seen reports indicating Russia is beginning to understand how to attack Starlink downlinks on weapons. Ukraine shows us that delivering compute where it can survive is the bottleneck. This means any compute we build must be edge compute capable of surviving in a denied environment. This is something defense tech startups, the primes, and the Pentagon are already chasing. This does not look like building a strategic surplus of data centers, which would do little other than obligate an increasing amount of the U.S. defense budget to something is already at the peak of its market and is itself starting to be constrained by power, political will, and high-bandwidth memory needs. This is a solution to a problem that is not central, and not peripheral, to today’s military risks.

The U.S. miltiary is building data centers, or attempting to if the political will lasts. They are building expanded compute. Saying that they need to do more to denies what we are seeing today. When we look at the Iran War, we do not see the U.S. struggling for targets, we see the U.S. struggling for a lever to compel Iran to submit to its will. We see a nation struggling to build the volume of defensive weapons necessary to protect its troops and meet allies’ requirements globally. We see a nation trying to mitigate risk with exquisite weapon systems and platforms with fragile supply chains. These are all issues facing the U.S. defense industrial base. None of those are because CENTCOM lacks compute. Choosing to focus on a strategic reserve of compute in the face of these issues would leave us wandering in the forest.

Leave a comment

There are deep structural questions about the limits of military power that underpin questions of policy and the use of military force. One of the clearest questions is, does the military understand its own limits? Does it understand what it needs to win a conflict? Let me start by saying this: the U.S. is the strongest and most capable military on the planet today, which, by virtue of our economic and technological advancements, does mean that it is the most powerful army ever to exist. This is not a value judgment. I am not saying it is the most right; I am saying that, objectively, if you put the U.S. military of today against the Allies at Normandy, Zhukov’s Soviet Army, or even against the Coalition from Desert Storm, the U.S. military would wipe the floor with them. Why then has the U.S. Army struggled to win in Iraq (for a time), Afghanistan, and force Iran into a settlement? Because none of those outcomes are things that the U.S. Army does.

In systems thinking there is a heuristic, “The purpose of a system is what it does.” Geek speak shortens this to POSIWID, which is unhelpful to the uninitiated. The purpose of this idea is not to make a simplistic view of the world or, as a form of snark, suggest that home warranties exist to deny claims. Instead, the purpose of this is to help strip out intentions, which can cloud observations and judgments when analyzing a system, and allow people to take a more unbiased look at the structural behaviors that underpin a system. Clearly, the U.S. military is not a system meant for state building because, instead of building states, it engaged in counterterrorism operations.

Justin’s Theory of Targeting

I have a simple theory of military targeting. My theory is that military targeting has four hierarchical stages. The pyramid is only four levels, and I would have made it three, but the fourth serves as a warning.

Level 1: Strategic Bombing – The base level is “strategic” bombing. No, there is nothing strategic about it. That name is something people use to make themselves feel good. We called the B-17, B-24, and B-29 strategic bombers, and the bombing in Germany and Japan was anything but strategic. It could not be strategic because we did not have the technology to ensure we only hit what we were aiming at. To make up for that lack of precision, we focused on volume. The targets were strategic: German ball-bearing production, artificial oil and fuel production, and assembly lines. Those targets were strategic, but none of them, no matter how many bombs we dropped, ended the war. It took the capture of the Nazi leadership, the seizure of Berlin, and the destruction of the German Army to do that.

Level 2: High Value Targets (HVT) – This is where we lived during the Global War on Terrorism. We hunted High Value Targets. The most famous HVT was Osama bin Laden. By the time Navy SEALs landed in Islamabad and stormed his compound, he was not running Al Qaeda, but he was important for us as a nation, if not for the enemy. This gets at the issue with HVT hunting. You are looking for targets that “we believe are valuable,” not that we think will have an impact on the actual fight. I will give you an example from the Ukraine War. Volodymyr Zelenskyy is absolutely a Russian HVT. Do you think that if they were able to kill or capture him the Ukrainians would stop fighting? Absolutely not.

An HVT is not supposed to be a target that is just worthwhile to us. An HVT is supposed to be vital to the adversary’s strategy, course of action, or tactical plan. That requires us to have a deep knowledge of our enemy, their courses of action, and to be able to take those options away from them. Instead, HVTs have become manhunts, and everything gets lumped into the package of HVT. During the GWOT, we would build HVT packets that sometimes included a courier. Are they important? Sometimes. Are they HVTs based on the definition? Often, no. They were not vital to a terrorist network’s operation. They could become vital if targeting them caused a change in behavior that allowed access to an actual HVT. Still, HVT is only the second of four rungs on this ladder.

Level 3: High Payoff Targets (HPT) – For this definition I will turn to the U.S. Air Force Doctrine. An HPT is “a target whose loss to the enemy could significantly contribute to the success of the friendly course of action. Given their importance, the Joint Forces Commander and components prioritize these targets.” To paraphrase Sun Tzu’s maxim, targeting HVTs requires understanding your enemy; targeting HPTs requires understanding your enemy and yourself. This is where commanders want to spend their time because this drives toward the actual end states the civilian leadership of the Pentagon and country have provided them. This is also where an objective that the military can achieve is important.

What are the military objectives around “make Afghanistan a stable democracy?” That was the end state that evolved after the successful overthrow of the Taliban in 2001. The mission became a political project to create a Western-style democracy and build national institutions and a strong central government in a country that had never had that. What are the enemies of functioning national institutions? If you do not know, do not worry; neither did military commanders, so they went HVT hunting.

Level 4: The Perfect Strike – This is a completely theoretical event. It is hitting just the right part of just the right enemy structure so that the entire house of cards collapses. To be clear, I do not actually think this exists. There are people who think it exists. Those people are part of, though not all of, the idea that you can bomb your way to national policy outcomes.

The issue with the idea that we have to fill the unforgiving minute with tactical action is that it robs us of the time and energy to stop, sit back, and reassess. It does not let us gather ourselves and make the right moves. Patton would have had the U.S. and British Armies cross the Alps like some modern reverse Hannibal to take the fight to Germany. The Allies were not ready for that. The same is true of anyone in the military who thinks military action can achieve every or even most national policy goals. It cannot because that is not the purpose of the system. A better understanding of that is essential to ensure we are using the right tools against the right problems as a nation.

Share Mind of Things

Share

So, what is a bubble? In this case, I define a bubble as 1) a growth of the industry at a speed and size that outstrips the Pentagon’s ability and willingness to award contracts, and 2) investors expectations of software-like returns for their capital investments. These both tie to one of BCG’s key takeaways:

Profits are more sustainable for exquisite systems. We mapped the contractor-addressable profit pools for two archetypal platforms: the exquisite F/A-18E/F Super Hornet fighter jet and the expendable AeroVironment Switchblade 300 loitering munition. That analysis shows that the shape of profit pools for expendable systems is not just smaller but fundamentally different.

This is the idea that underpins its “bubble,” and it is something that is both obvious and not indicative of a bubble. Of course, the Pentagon is going to spend more on exquisite systems. The research and development that underpins them is not borne by the company; it is shared or even fully funded by the Pentagon as it searches for the leading edge of technology. These are hypersonic missiles, next-generation stealth technologies, and the productization of basic research into applied sciences and technologies. This fact alone, that the government holds the R&D funding risk, means companies are able to make more stable profits, as each item they sell has a higher contribution margin than a system built on company R&D dollars. This is also not indicative of a bubble.

The more pressing issue is that investors are driving for software like margins and profit in an industry that does not support that model. There is more pricing pressure on defense tech companies, caps to how much profit they can actually make on sales to the government, and higher marginal costs that just running a tech stack. The belief that we will completely Silicon Valley-ize defense tech is wrong. They will compete and make some great products, but they are also bound by rules and facing a customer that literally holds all of the power in the relationship. That is even more true with regard the Pentagon’s use of Other Transaction Authority (OTA) which are what many startups are relying on.

There is a bubble. Of course there is a bubble, because there is money to be made. Potentially, a lot of money. The bubble is a swelling number of defense tech companies seeking to win contracts with the Pentagon to produce everything from rifles and bullets to autonomous drones. The clearest example I have that there is a bubble comes from a recent defense tech conference I attended. It was hosted by a prestigious East Coast university. It was one of the first times I had seen a high number of undergraduates. Normally, at these types of events, you are around MBAs and PhD candidates from the university who are trying to find companies and funding for their ideas. The number of undergraduates struck me enough that I had a conversation with some of them and others at defense tech companies. I was told, in very clear economic terms, “Of course they are here. This is one of the areas of the economy. It’s getting harder for these guys to find entry-level positions at the traditional tech companies.” Defense tech is a growth industry.

Is that a problem? When people talk about a bubble, it brings to mind images of the 2000 dot-com crash. People think about the number of internet companies that went bankrupt during that time and the human misery that arose from it. What they forget, or maybe do not connect to that era, is that the dot-com era laid the foundation for our modern economy. Yes, it was painful, but it is also the reason that we have ubiquitous high-speed internet, fiber, and data centers. It gave us the infrastructure that serves as the backbone of the current economy. The defense tech bubble will, hopefully, do the same thing for defense production capability. It will build the infrastructure necessary to produce attritable weapon systems at scale.

This is actually another thing that gives me pause about BCG’s analysis. It sees growth in defense spending through 2033. We are already at historic highs in total dollars devoted to defense. I think it would be unwise to assume that situation will continue for the next six years. That would require the next Congress and the next presidential administration to continue along the same spending track, which is far from a foregone conclusion. Even if you laid that aside, the global position of the U.S. defense tech stack has changed and is under pressure from European companies for market share in a way that it has not been in the past. This further restricts the potential future market for U.S. defense companies.

These are all downwards pressures on the bubble. A bubble that can either pop dramatically or deflate slowly. No matter which happens, the bubble dissipating is nothing more than a return of capital discipline from the frothy highs of the current market. The worst thing that could happen, in my opinion, is for this to happen now, in these heady early days. Now much of the money from the OTAs is not going to production capacity. It is going to software development, prototypes, paying salaries, and building facilities. If the bubble pops before these companies reach full rate of production all of that investment is lost. The metric of how good or bad the bubble ends up being is what and how much it builds before it burst. In this case, we should be hoping the bubble results in more factories, improved production capacity, and onshored supply chains. If the results of the bubble lay along this route, the U.S. will be better off.

Leave a comment

Share

Design is an important part of product creation and development. Steve Jobs demanded the layout of circuit boards and internals, things that customers would never see, should be beautiful and thoughtful. His mantra is something that has dug into Silicon Valley and some of the emerging defense tech companies. Even the Pentagon realized that design was important. In 2025, they adopted Human Readiness Levels. This a 9-point scale where each level of design corresponds to steps companies need to take to consider the size, weight, ergonomics, and form of devices that, at the end of the day, must be used in combat by humans. This scale tracks along the same path as technology readiness levels (TRL), another 9-point scale, that establishes that a technology has proven it does what it says it does. Together, these scales combine to produce functioning and functional devices that warfighters can use out of the box. Right?

Enter the IVAS.

Just look at that monstrosity. Leaving aside the size of the googles and clear impact to peripheral vision, the cabling alone should be cause for the program director that bought these to be called on someone’s carpet. This is the IVAS, the Integrated Visual Augmentation System. An all-weather fighting goggle developed from 2018 to 2025, fielded to 10,000 soldiers at a cost of nearly $2 billion all to end up on a trash heap.

This was never ready for fielding, or built to be fielded, but to the tune of 10,000 units it was fielded. The IVAS an instructive case in how defense companies, new and old, struggle with design, despite that being a Silicon Valley position of pride. Design is important and, honestly, the military is pretty easy to design for because we already know things will be bulkier and have to be more rugged than a consumer electronic. I don’t routinely put my iPad or phone in a 90+lbs sack and toss that into the back of truck with other people on top of it, but I did that all the time with night vision goggles, Toughbooks, and radios. The expectation was they would take a beating.

The IVAS is big, bulky, power hungry, and could get people killed. I do not say that glibly. That wire harness, with its ability to get caught on literally everything, limit head movement, and restrict vision, would have caused U.S. military personnel in combat to die. It offered no added benefit because, as the Government Accountability Office said, “they did not perform better than existing equipment.”

Designing for military users means understanding military users and their needs. That is not to say that the only people that are qualified to design are veterans, at all. What it does take is a willingness to go to where they are, listen to them, and not start with a solution in search of a problem. The IVAS was that. Even now, the U.S. military continues to look for ways to enhance soldier situational awareness, which is what the IVAS purported to do, but what is the situational awareness issues that we are solving?

This is the opening issue the GAO cites. There was a weak business case and fail slow approach. The weak business case meant the acquisition programs do not balance the technology, design knowledge, funding, and time necessary to transform a need into a product. Because of this, you end up with a product that does nothing for anyone.

I can only hope the next iteration of the IVAS, whatever that is, has a real business case. That case has to start with an actual problem. It would be amazing to see in smoke and through walls, but not at the cost of never being able to actually see the battlefield with my own eyes. The idea that I could rapidly move through all of the available drone footage is nice, but not at the cost of never being able to focus or always being connected to a network that can give away my position. These are all the tradeoffs that you have to consider when you are designing. Power, weight, connectivity, heat, those are engineering issues, and you can solve those all day and never build a device a soldier will carry. If they won’t carry it, it might as well not exist.

Products and production start with design and design has to start with either 1) a vision or 2) a problem. The IVAS started with a good idea, that we are going to make soldiers superheroes by giving them greater powers. The issue is soldier vision is not why we have lost a war or even a battle. We have lost because of poor planning, lack of tactical patience, unclear or unachievable endstates. The U.S. does not have unlimited resources and, as we have shown, does not have an unlimited ability to produce, so designs that are not tied to solving a problem taking up years of resources and time are things that are unsustainable.

This week we found out that the Pentagon invested $469 million to build 155-mm artillery rounds at the Mesquite facility. To date, that facility has produced no projectile parts. This is another sacrifice on the altar of “we will just figure it out.” We will not just figure it out. This is why deep knowledge in production and technology is important and why countries and companies have to rebuild that knowledge when they lose it.

Share

One note at the start: the Inspector General report goes to great lengths to say, “the contractor” and never acknowledges which company is managing production. It’s only odd because that was a very big storyline for the company when it was announced that this plant was starting, and they made a significant investment in the facility ($200M). The contractor is General Dynamics Ordnance and Tactical Systems, and they also run the Scranton Army Ammunition Plant. When you add in GD’s investment, the total investment exceeds $669 million to produce zero useful objects. (See: this, this, and this, as examples)

The U.S. need for increased artillery production is in response to the use of artillery rounds during the Ukraine War. The Pentagon wanted to quickly expand production to meet the new requirement by increasing output to 100,000 rounds per month across all U.S. production sites. It is worth noting that the U.S. only has three sites capable of producing these parts. The Mesquite site was supposed to be number four. At maximum capacity, which they have not yet reached, the three existing sites can only produce 71,000 rounds a month (29% short of the desired 100,000). So, the Pentagon expected the Mesquite facility, a new site that had never produced 155-mm rounds before, to produce 29% of the country’s 155-mm artillery rounds. The program office, CPE A&E, accepted the “risk” in the contractor’s plan. The contractor’s plan was to buy production equipment, not intended to produce 155-mm rounds to U.S. specifications, and adapt it. In simple terms, they were going to figure it out.

This contract started in November 2023 and ran into issues right away. By June 2024, they had received a Letter of Schedule Inquiry from the Program Manager. This letter is an administrative tool that says, “we don’t think you are going to make it.” Those letters progressed from Letters of Inquiry to Letters of Concern and eventually to stop-work orders. The report lays out the risks in the plan:

CPE A&E accepted the risk associated with the contractor’s plan to purchase and adapt M107 metal part production equipment to produce a newer variant of the 155-mm projectile metal parts at different specifications. According to CPE A&E officials, the Mesquite facility was a high-risk, high-reward opportunity, and CPE A&E is realizing some of that risk now because the facility has been unable to produce acceptable rounds. CPE A&E accepted the contractor’s proposal to acquire and adapt unique production equipment and an unproven production process because equipment was already available. According to the CPE A&E official, obtaining and adapting the available production line equipment…

I am all for taking risks, but this was clearly a poorly executed risk analysis. First, the contractor had years of experience producing at another facility using different processes and different equipment, and that facility was already struggling with communication, maintenance, and meeting schedule deadlines. There is no evidence that they could successfully adapt this technology, despite having the opportunity to attempt it, in part or in whole, at their existing facilities. Additionally, the Army did not open this contract to competition, which compounded the risks by making the solution seem pre-baked. The lack of open competition meant that any other options that could have secured production using different machines or capabilities were excluded. It is possible that no one else was better suited for this contract, but that was never tested. I believe that happened, at least in part, because the war in Ukraine made us afraid of our production fragility.

What does all of this have to do with my issue with “we will just figure it out?” My issue is that this mindset produces outcomes like the Mesquite facility. We take a high-risk solution that is unproven, rush it to scale, block out other options for solving the problem, and hope that “we will figure it out.” You will not. You will piecemeal something that might work some of the time. Some of the time, winging it works. There are times when acts of individual heroism can pull you across the finish line. Most of the time is not good enough. Real solutions require deep process knowledge. That requires accumulated knowledge. If you need an example of things you just can’t figure out look at shipbuilding and semiconductor fabs.

There are fields that take years, sometimes decades, for employees to master the necessary processes. Shipbuilding and semiconductors really highlight how fragile our accumulated knowledge and capabilities are. In 1960, the U.S. was the global leader in chips. Texas Instruments and Fairchild Semiconductor controlled nearly 100% of the global share. Ships, though, were a different story. Even in the 1960s, we produced only 2–5% of the world’s commercial ships, but we could still turn out naval vessels quickly. In just 20 years, we went from being one of the world’s leading shipbuilders to being surpassed commercially by Japan and European shipbuilders. Why did that happen? Because we stopped making commercial ships because other nations did it cheaper. We moved away from that business. It took decades for the U.S. to lose its lead in semiconductor production, and we lost that because we moved up the value chain to higher margin design and away from production.

Despite our movement away from production and into design, why can’t the U.S. simply jump back into the lead in manufacturing chips and ships? We no longer have the institutional knowledge that got us there because we moved into higher-margin work (read: design, not production). Manufacturing is hard; manufacturing at scale is harder. It takes hundreds or thousands of things going exactly right to make quality electronics and devices. The complexity only increases when we are talking about highly precise systems that must perform in high-stress environments. It takes people, machines, and processes working in coordination to produce a round that can fly more than 30 kilometers and impact within a one-meter square.

The results of the IG investigation are not surprising. They are disappointing, but not surprising. It is not a scandal, well, it is a bit of a scandal because of the money, but it is not a scandal because it should surprise no one that you cannot skip learning curves. Manufacturing expertise is not easily transferable and cannot be created on demand. A Special Operations Forces Truth states, “Competent SOF cannot be created after an emergency.” The same is true of advanced precision manufacturing. Once it is an emergency, it is too late. This has been my core argument for expanding production and reshoring defense manufacturing, from raw materials to steel. Production matters for deterrence and surge capacity. Right now, we are already seeing that, at maximum capacity, we can produce only 71,000 155-mm artillery shells per month. In 2025, Ukraine fired between 2,000 and 7,000 rounds a day, while Russia fired more than 20,000. That means Russia could expend the equivalent of one month of total U.S. production in just four days.

We need to take risks, but we are already starting from behind. We cannot believe that simply pouring money into the system will produce manufacturing excellence. Manufacturing excellence is accumulated by building an ecosystem of engineers, machinists, suppliers, and managers who solve the thousands of problems that arise every day in large-scale production. The U.S. has spent decades optimizing and moving up the value chain. We are innovative and focused on design and high margins. Those margins, however, have come at the expense of resilience and the ability to expand production during a crisis. In defense, that tradeoff is unsustainable. We can regain our expertise and industrial capability, but it will take time, and it cannot be done after the emergency begins.

Leave a comment

Subscribe now

A few weeks ago, an SVP at Shield AI posted about how their Shield AI X-BAT looks different from other autonomous planes. The premise was that their technology is different, and you can see it, and that shows that they are conquering technological challenges through design. This post is not a comment about Shield AI but about the overarching structure of competition in U.S. defense tech. The post shows that companies are creating visibly differentiated systems, but should that differentiation extend through their supply chains and components?

The U.S.’s issue in the DIB is the ability to make things at scale. We struggle with speed and scale. I am not calling for picking winners, but I am calling for some common standards that make it easier to scale production. Competitive differentiation is what the Shield AI post pointed to. They are saying their design difference is their power. This is an operations concept where you, as a producer, standardize and then specialize. In practice, this means that companies and industries would use common parts and common structures and add their special applications at the end.

Vehicle manufacturers have done this for years. It’s why Ford uses platform architectures. They can differentiate their vehicles while keeping core parts the same. It allows them to produce multiple vehicles and scale across assembly lines should they need to. The U.S., as the Arsenal of Democracy during WWII, knew this. They limited variation, and designs were meant to scale rapidly. This creates fungibility in their model, not in the products, but in areas of commonality, such as the supplier base, components, and manufacturing. That fungibility could be applied across the DIB, but it is not. How much of today’s industrial base could be redirected toward defense production without building new factories and supply chains?

During World War II, Germany produced 30 types of aircraft and hundreds of distinct vehicle types in factories that were under constant threat and suffered from supply shortages of everything from rubber to fuel. All of these were hyper-specialized, with different parts, maintenance schedules, fuel consumption rates, and rates of production. Germany compounded its disadvantages by producing vehicles and weapons that overburdened its manufacturing, logistics, and maintenance systems. They would have struggled in peacetime, but by the time of Operation Barbarossa, they were under embargo and increasing stress from the war. They took this weakness into a war not only with Britain but expanded the war, of their own volition, against the Soviet Union and, after Pearl Harbor, the United States. These bespoke systems drained finite engineering, tooling, and production capacity that Germany could not afford.

The U.S. is not Germany in 1941, but the defense tech ecosystem is starting to create a form of industrial fragmentation. Our defense acquisitions center on competition between finished systems rather than substitutability among components, production processes, and suppliers. As more companies enter this space, we are seeing more parallel supply chains, more proprietary interfaces, and more attempts to capture programs. The government can create competition at the end-item level while simultaneously creating fragmentation beneath it. The better answer looks like more involvement with industry, not just the new primes but also suppliers and the supply chain. Designing around common parts, common designs, and common tooling helps ensure that production can scale quickly. However, that does not help differentiation in product creation in the market today.

What can we standardize? We can standardize physical interfaces, power architectures, data buses, connectors, manufacturing, batteries, machine tooling, and test infrastructure. These are all at an industry level. At the government level, we must get better at defining and establishing the means to run qualification testing. These similarities streamline production downstream and expand suppliers’ markets, which, in turn, drives down total costs.

There are also things that industry should not standardize. Companies need to keep their mission software, autonomy, sensors, payloads, and any manufacturing advantages they have discovered. Differentiation is where companies create value for their shareholders and the warfighters who use their products. Differentiation and similarities, when balanced, start to make the defense industry modular in a way that it currently is not. That creates the ability to rapidly expand production when needed. This is the type of fungibility that would build surge capacity, create larger markets, and better justify and distribute capital investments.

Share Mind of Things

To fully incorporate this change, the government must incentivize defense tech companies the right way. Currently, their incentives revolve around winning individual programs that are disconnected from other programs in the department and creating proprietary intellectual property. Both create vendor lock-in with the program and help ensure that companies remain incumbents. This creates structural barriers to entry and strips modularity out of the market.

The government should standardize mature infrastructure and interfaces but preserve areas for competition and experimentation at the mission-capability layer. What would the DIB look like if it designed systems and weapons for wartime modularity rather than differentiation? Looking at small UAS, we can see several areas where standardization would be helpful. Batteries, manufacturing materials, and communication standards are all areas that could be standardized below the level at which individual companies differentiate themselves.

Silver Bulletin

America used to be exceptionally patriotic. Now we're below average.

I’m going to admit to feeling a swell o…

Read more

2 months ago · 67 likes · 39 comments · Eli McKown-Dawson and Nate Silver

Predictably, the numbers reflect fractures in our politics parties, but my pride in America is not limited to politics.

Their chart shows a 44% drop in Americans feeling proud to be Americans.

This was not the most interesting chart. Instead, they had one that tracked the share of Americans who were extremely proud since George W. Bush’s first term. At that time, 61% of Democrats and 75% of Republicans were extremely proud. Today, 74% of Republicans are still extremely proud, but only 17% of Democrats are. Those numbers have trended downward across every presidency since GWB1.

It was interesting to me because I grew up in a military family, with grandfathers who also served in the military, and we were never overtly patriotic. We didn’t listen to Lee Greenwood on repeat, and honestly, I hated Toby Keith’s post-9/11 music. That said, we were thankful that we lived in America and for the opportunity that it afforded us. Both of my grandfathers retired from the military, started small businesses working with their hands, and were able to provide for their families. Things were not always perfect, but they consistently improved, and America was a force for good in the world. It underpinned a liberal democratic order that saw the U.S. at its head and helped lift large swaths of the globe out of poverty. America was aspirational, able to change and reinvent itself in ways that other countries did not.

Today, I am thankful and proud to be American. I might not agree with the administration on much, but my pride in America is not limited to politics. I am thankful that we can change our government with ballots and not violence. I am even more thankful that I have lived across the country and found community and family in places as diverse as Alabama and Washington State. My hope is that people will start to realize that the things that make America great were never our politicians, but our communities; that our differences are not so great; and that we are in a fortunate position because we are in a country that can change.

Happy Independence Day everyone!

Leave a comment

I got to spend several hours with an Estonian commander. He was very forthright about the issues he sees with Estonian, NATO, and even U.S. acquisition systems. He offered that now, four years after the full invasion of Ukraine and 12 years since the invasion of Crimea, he recently took part in an exercise with multiple countries, and none of the teams were using drone detectors in their tactical units. He was disappointed but said that this is exactly where the alliance is currently. Every year, he attends conferences where they discuss the need to speed acquisition and learn lessons — I have written about those same lessons for over a year now — but the conferences are filled with words, and very few of the countries are actually taking the types of steps that would mean change.

Subscribe now

I pressed him a bit here. Was he just talking about buying new drones or equipment? He said no. Faster acquisition would be more akin to an outcome of the type of change that he wants to see. It starts with a mental shift that testing, validation, and experimentation need to happen much faster and with less bureaucratic oversight. His reference was to U.S. Special Forces and how long it takes to try to make a change to the organization of a Special Forces Group. He said, “How many times recently has your Special Forces tried to change their TOE (table of organization and equipment)?” The answer is several, but commanders have limited authority to make changes without asking the larger Army. Yes, there is an MTOE, or a modified TOE, but even then, the Army limits these modifications to local administrative changes.

He said, “In Ukraine, they are changing organizations several times a year. If two regiments decide a change is necessary, it’s done, no additional questions asked. They have eliminated sniper positions, expanded drone and fires capabilities, changed how teams are organized and employed, and are constantly innovating.” You can imagine the feeling this caused me, having lived through multiple attempts to define the future of Special Forces during my past career. But this is more than any one unit or its need that day. It is about building the type of testing and validation at home that allows for this type of experimentation. That is where my friend was struggling.

We get lots of weapons and toys from companies that claim to have been tested in Ukraine. What I have learned is that there are layers. There are companies that have gone to Ukraine and tried to work there themselves. Those companies are no longer in Ukraine and have tried to sell other things. There are companies that have been to Ukraine, were used by Ukrainians, and were then rejected for any number of reasons, but largely because they did not work. Then, there are companies that are in Ukraine, have been tested by Ukrainians, and are still working. The Ukrainians are very matter-of-fact in their assessments of new technologies. Many do not even make it to the field, as Ukrainian experts look over the technology and can quickly find multiple points of failure. They can do this because they live in a live testing ground.

The live testing ground and the ability to separate the wheat from the chaff is a key development the U.S. needs to make. This must be more than just adding a few quadcopters to the National Training Center or Joint Readiness Training Center. It means more units, at all levels, need experience operating in these environments. We see stories every time a Ukrainian unit comes out to train with NATO forces. The Ukrainians operate as Red Force; they use the techniques they have learned on the battlefield and are credited with destroying large numbers of NATO forces. I think the reasons for this go beyond their being well practiced and stem from the fact that NATO and the U.S. have not been training like they are going to fight for a long time. We hand-wave, “Oh, there will be jamming.” We still like to have our formations too large, too concentrated, and too controlled, and we do not build the types of training areas where we can run long-term jamming across multiple spectrums in the U.S. because of the disruption to commerce, flights, safety, and myriad other reasons.

What is the right answer here? I am not sure. It is an extremely hard problem to solve, made harder by the fact that we have limited contact in Ukraine. Our military personnel in the embassy largely work from other parts of Europe. U.S. units have only limited in-person contact with Ukrainian units, and many are reliant on periodic updates and word of mouth, which I realize is also what this is. Increasing contact there is one of the most important things we can do in the short and mid-term to address organizational issues or form the basis of ideas we think we might need to employ in a future conflict. A true lesson here is that the next conflict will look different than we expect, so I do not think we need to make large-scale changes today. But I do think each commander needs a viewpoint, ideally, a strong opinion held weakly, about what the optimal organization should be and then needs to test it repeatedly. Expanding contact needs to be everything from expanding the number of observers in Ukraine to setting up paths for direct unit-to-unit lessons learned and embeds during U.S. and NATO training exercises, not just as a red force, but also offering insights and instruction on how to best counter new and emerging capabilities.

The darkest part of our conversation was on how Ukraine handles transition. This is another area where I wish Special Forces had been more engaged from the beginning. In U.S. unconventional warfare doctrine, there is a final stage to a campaign: transition. This is how you take a population, or a portion of the population, that has been fighting for years and integrate them back into society. Shovels, construction, shop owners, etc. An issue with Ukraine is that the wartime economy has created some real incentives that make it hard for soldiers, some of whom are earning 3,000 euros a month, to go back to a 300-euro-a-month construction job at the end of the war. This is not even addressing the shock, PTSD, and general trauma these soldiers have. U.S. SF doctrine defines transition as:

Transition is the final, most difficult, and most sensitive phase of UW operations. The planning for transition begins when the U.S. Government decides to sponsor an irregular organization and ends in the UW operating area upon cessation of operations. Transition does not necessarily mean demobilization or the commencement of FID operations. Successful transition from combat operations requires a comprehensive approach to stability tasks where the drawdown of irregular indigenous forces support lines of efforts in security, justice and reconciliation, humanitarian assistance, governance and participation, and economic stabilization and infrastructure. Achieving the desirable end state (stable governance, sustainable economy, and safe and secure environment) will involve diverse objectives.” – FM 3-18, 28 May 2014

Transition is so hard that it is difficult to find a good example of it in U.S. history outside of the American Revolution. On a small scale, there was the transition of the French Underground back to civilian life, but that was on a different scale than demobilizing the Ukrainian military. Post-conflict transition was clearly on my friend’s mind. One final thing he said was, “Some of the Ukrainians are afraid of what would happen if Russia began to look very unstable. The only thing the West fears more than a strong Russia is a destabilized, weak Russia.” He said there are worries that the West could abandon Ukraine if it looked like Russia was going to collapse. “That would make the Ukrainians feel like everything they had lost was for nothing.”

The conversation was sobering. It was good to get a better understanding of how some in the states closest to Russia and the war see the world. He asked some extremely tough questions, questions that he and his country are going through right now and that the U.S. is also struggling with. Those tough questions do not have quick or easy answers.

Share Mind of Things

In February, the Pentagon first threatened Anthropic with a supply chain risk designation because Anthropic believed they should be able to limit the use of their model and how it is used by the Pentagon. This was already a contentious debate. It is one that we have had on the show and that I was of two minds on until about a week ago. At the time I sided with Anthropic. I thought, and still think, discussions of the limits of AI’s use in surveillance and fully autonomous weapons is a good debate. Also, like Amodei, I believed the Pentagon was making two inherently contradictory threats. One was the threat to label Claude a security risk; the other was claiming Claude as essential to national security.

Share

There was, however, another claim in the same piece by Amodei, “We have never raised objections to particular military operations nor attempted to limit use of our technology in an ad hoc manner.” This is the issue. Anthropic did post-deployment limits and they did so without alerting paying customers they had or even could, unless you read the fine print. Claude made the argument in court that once their model was deployed, they could not alter, update, or recall it. Then they altered the model, effectively recalling it and forcing in a less advanced model, because of a concern they did not share with the user. From the May appellate court case:

Judge Neomi Rao, who was nominated by Republican President Donald Trump, questioned what basis the court could have for second-guessing Hegseth’s judgment. The Pentagon’s dispute with Anthropic centers on how AI technology can be used in fully autonomous weapons and potential surveillance of Americans.

“I take the secretary to be making more general points than the ones that you’ve identified,” Rao told Anthropic attorney Kelly Dunbar. “It’s about risk, and they say, ‘Well, based on what we know, we can’t trust that the (AI) model may not have something embedded within it that is going to create a problem for military capabilities.”

Judge Rao seems to have called it. The fear was that Anthropic could restrict models without telling the user. They said they did not have the ability to do that. They did. I still do not agree with the Pentagon’s approach to this whole case, but Anthropic lost a lot of credit with me on this one. It was an unforced error they did not need to make and could have announced in advance of launching the new model. Instead, they showed they could do exactly what the Pentagon was afraid of. Based on this alone, I expect the Anthropic and Pentagon fight to get worse before it gets better, because neither side trusts the other.

The author sees the problem. U.S. Marine Corps Littoral Teams are outranged and outgunned by potential adversaries. This is a risk to his forces today. Marines are using 81mm mortars, with a max range of 6 km, against potential enemies that could deny beaches from 12 km away. His solution is commercial FPVs. He describes the “defense establishment’s” resistance as arguing:

that relying on commercial-off-the-shelf drones would introduce unacceptable vulnerabilities. Skeptics note that "drones bought off store shelves" lack military-grade encryption and hardening, making them susceptible to advanced electronic warfare.⁸ However, this perceived fragility is offset by the disposable nature and low cost of these systems. The Marine Corps' modernization roadmap emphasizes the need to rapidly field a "low-cost, air-launched family of loitering, swarming munitions.” Instead of relying on a handful of highly protected assets, units could use "simple and relatively inexpensive systems" to overwhelm adversary EW through mass, turning unhardened platforms into a highly adaptable swarm capability.

There are several issues with this characterization of the discussion. One reason I want to highlight this piece is that I want to see young officers and NCOs writing. This captain is writing and thinking about future force structures, which is great. However, those ideas should be pressure-tested because captains eventually become colonels and generals. If the misconception of why we have the NDAA for drones is allowed to fester, it makes it hard to engage with partners, including the defense establishment. Whatever the defense establishment is, it is a group of people trying to build for warfighters. The issue here is that the solution to this range problem is not as simple as “buy COTS drones.” The cheap COTS drones are, by and large, built with parts sourced from vulnerable supply chains.

Supply chain security is the first reason the U.S. military should and will pay more for drones in the short run. This makes sense when we think about the economics. U.S. workers make more than Chinese workers; therefore, U.S. and allied secure supply chains cost more than outsourcing to Shenzhen. The Pentagon should not simply be buying COTS solutions for drones because the majority of drone manufacturers are not U.S.-based and do not have secure supply chains. This is why the U.S. required the purchase of critical components through NDAA compliance. Could the PRC disable all of the cameras or take over drones if we used non-NDAA-compliant parts and systems? Maybe this is a real vulnerability, maybe it is not. Basing doctrine and fires on devices that are reliant on PRC-controlled supply chains is absolutely a vulnerability. If we do not think it is, we have not been paying attention to the PRC’s willingness to use supply chains to its advantage, from critical minerals to magnets. The U.S. should not base any of its ability to generate combat power on an adversary’s production capacity.

The other issue is that mass alone does not overcome electronic warfare capabilities. That actually requires increased cost. Whether you fly one, 100, or 1,000 FPVs at an emitter that is blocking their current comms, you are likely to lose them. Getting past that requires directional antennas, frequency hopping, AI, computer vision, fiber optic drones, and a system that can coordinate all of this to ensure the right mix of capabilities is available. This is one advantage that the U.S. combined arms doctrine supports. Unlike other forces that are solely reliant on their FPV capabilities, with limited support from adjacent fires, Marine littoral doctrine was, and is, based on a high volume of joint fires supporting their movement. That calculus may be changing, and may have already changed, but that is one reason it is not on the Littoral Team alone to neutralize threats 12+ km from the shore. In fact, being a sensor for other shooters is a core part of the design of the Stand-in Force.

Share

Some people, those that are willing to put up with our pedantic takes, will say, “no, what I mean is when will a weapon system be fully autonomous and decide when to use lethal force?” I do not know. I imagine soon. It is not a long distance from the Counter-Rocket, Artillery, and Mortar (C-RAM) system that can quickly identify and destroy mortars, to an automatic weapon that can detect movement in a no-man’s land and fire on it.

I plan to write and talk more about autonomous weapons, but before I dive too deeply into that, I want to spend time thinking about the adoption of AI in our decision making. From a military perspective, I believe that thinking about how we think, and what we surrender to AI and what we retain for human commanders is important as it will greatly change how military officers are trained, raised, and what is expected of them.

What prompted this discussion was actually sitting in an ethics class recently. Ed Freeman, the “father of stakeholder theory,” was discussing how managers and business leaders should think about ethics. I would love to say that I gleaned some deep insight from Ed about how to think about AI and decision making. I did have many insights. I think he is a truly brilliant man and great in a classroom. Something that stuck with me, and honestly distracted me during the course, was a portion of the group was actively putting his questions into Claude to get responses. This actually triggered one of my worries about the adoption of AI in the military.

If in business, people are willing to hand of ethical decisions and how to think about them to Claude, when money is on the line, what are commanders going to do when faced with hard problems and lives are on the line? Does AI start to serve as a tool to allow a commander to abdicate risks rather than a tool to help reduce risks?

The idea that it could be a tool for abdicating risks worries me. Abdication breaks the chain of accountability and creates power vacuums that increase institutional instability. The military, as a hierarchy and by virtue of what it exists to do, requires leaders to own risks. That ownership extends beyond personal responsibility and extends to a moral justification for their position. Commanders are not in charge because they have served for a sufficient time or because they have a certain rank. In fact, the U.S. Army has something called the command select list specifically because of this. This is a list of officers that have reached the rank of Lieutenant Colonel and above that have been trusted to lead organizations. The Army also promotes other officers to Lieutenant Colonel. These are non-CSL and are often staff officers. They are in charge of themselves and their small shops. The Army selects officer for command because they believe these officers will exercise the type of judgement that is necessary to lead people and organizations into harm’s way. How fundamentally different is that process if decisions start to move toward AI?

I do not have great answers to any of these questions. But over the summer I plan to write and think more on the adoption of AI, the risks, and the benefits. We also plan to continue having these discussions on WarTalk with leaders and thinkers from across the world.

Leave a comment

Soldiers in the U.S. have not always been festooned with garlands and honors. I am not even talking just about the post-Vietnam soldiers. Those familiar with conflict see war, and, by proximity, military service, as both horrible and necessary, while those at a distance see only an abstraction. In my lifetime people have described the bulk of the men and women who serve as little more than the great unwashed masses. (See: The New Republic, Sojourners, Inequality.org) The people who would join would have to be those without options. I do not agree with this. While economics play a role, there are plenty of people who could join the military and choose not to despite their clear economic incentives. Many want adventure and to be on their own, and some, and this is a smaller few, hope that they get to go into a war. This is what the poems, songs, and stories are about.

Share

One of the defining features of patriotic writing is the idea that military service and fighting for one’s country is a virtue. You see ideas around the positive nature of service, and often the need for compulsory service, resurface in countries as distinct as Russia, Britain, the U.S., and China over time. Starship Troopers makes a core argument that service should be core to citizenship. These ideas have a deep and lasting impact on our society. As wrote in his recent Progressives Don’t Understand Masculinity, war and military service have an essential male quality that offer men status in society, event today. We are attracted to stories of the warrior though we often revile the atrocities committed during fighting.

I think that war is seen in a positive light because of both its risks and the fact that it is by and large fought by those who cannot and do not make the decision to go to war. Why we fight, not just who fights, should be the core focus of any analysis of war. The why is what is important. It is also what can change the most dramatically due to personal circumstances. A classic example is the change in Rudyard Kipling during World War I.

In 1919, Rudyard Kipling wrote Epitaphs of the War. Contained among them is perhaps the most scathing take on World War I written at the time.

If any question why we died,

Tell them, because our fathers lied.

It’s important to understand what led Kipling to write this and who he was in British popular culture prior to writing it.

Kipling spent a long portion of his literary career writing about the Empire as a force for good in the world, how military service was a patriotic duty that was necessary to sustain the empire and core to British identity. His novels and poems often engaged with the age of empire. In Kim, he follows a young Irish-Indian boy who becomes a spy for the British Secret Service. Soldiers Three is a collection of short stories that follow soldiers in colonial India, and in Barrack-Room Ballads, he captures the dialect of the common soldier in ways that are still used today.

Kipling, like some of today’s commentators, saw military service as core to teaching and testing the range of emotional control, endurance, and integrity under extreme pressure that are necessary in society. In fact, at the outbreak of the war Kipling celebrated the idea of the “iron sacrifice” that Britain was required to make to defend Europe and the world from the threat of the Germans. Yet, when we turn back to “Common Form” and Kipling’s other post-WWI writing, we see a man who is changed.

In 1914, Kipling had a 17-year-old son called John “Jack” Kipling. Jack was every bit his father’s son. He wanted to join the Army or the Navy and serve Britain. He attempted to go to Sandhurst and become an officer, but the British military barred him from commission because of severe near-sightedness. Kipling—both Kiplings—would not accept that outcome. Rudyard Kipling used his considerable name and friendship with Field Marshal Frederick Roberts to get Jack a commission in the Irish Guards. This was August 1914.

Jack and the Irish Guards arrived in France in August 1915, on Jack’s 18th birthday. In September, at the Battle of Loos, he was wounded and went missing following an attack on German positions. Jack died in France, fighting a war of attrition, trenches, steel, and barbed wire. As a father, I cannot imagine the impact that this would have on me. I would be beside myself.

Kipling may have been as well. He clearly grieved, but he also turned against the war. It is easy to dismiss Kipling’s early work as that of a man removed, and to say that only after the horrors of war touched him personally did he see the war as a terrible thing. I think this is overly simplistic. There are plenty of men who see the horrors of war and still think it worth paying. However, Kipling came to believe that the war had been one more of choice than of necessity. This is a narrative that has grown in modern military histories. Christopher Clark’s The Sleepwalkers: How Europe Went to War in 1914 highlights the many missed opportunities that could have prevented the war.

Still, Kipling thought the war was worth fighting. More importantly, he believed that any war that is being fought should be fought with a view on winning. Kipling came to write and make the public aware of the sacrifices British soldiers and sailors made while simultaneously holding politicians that failed to prepare the nation for the trials of combat to blame for the length of the war and loss of a generation. These were politicians that dreamed up easy success without understanding the realities of conflict. Kipling also attacked generals that left soldiers to “die in their own dung.”

It is often the people who are the most directly impacted by war who see both its utility and its horrors as two sides of the same coin. Too often, war is thought of in the abstract. It is something that can wash a person or a country in glory. The truth is more complex. As Sherman said:

“I am tired and sick of war. Its glory is all moonshine. It is only those who have neither fired a shot nor heard the shrieks and groans of the wounded who cry aloud for blood, for vengeance, for desolation. War is hell.”

Despite the hell that he lived through, Sherman never regrated his actions in the Civil War. He, like many, felt that the most just war was a short war, and the best way to make a war short was to be brutal. This is the hard reality that many face. War is brutal, there are still things that are worth going to war over, but the use of war should be measured and a last resort. This is not to say that you must be brutal. Just that once fighting is started it needs to be carried out swiftly and against the aggressor, not civilians.

My family has served in uniform in some form or fashion for generations. They endured the post-Vietnam stigma and continued to hold their heads up. The reality is that service in the armed forces is part of our family. It’s from that position that I want a society that thinks critically about war and warfare. Like all human endeavors war is both simple and hard. It tests people, will, resolve, and reasons. It also requires deep thought an analysis. What is most pernicious about warfare is that it is the fathers who decide and the sons who die. Something Kipling came to see only too late. We should look critically on war and why we fight, because it is something that is worth our attention as a citizenry and we should not accept uncritically the pronouncements of anyone.

Leave a comment

“The bar fight is the PhD, “says Dino Gardner. I submit that is an incorrect understanding of what the OSS did and what violence generally is. The bar fight is actually failing the PhD. It is better to recognize a fight is coming and build a coalition of allies strong enough to stop any fight. Second to that, being able to convince another person to fight for you is better than fighting yourself.

Let’s remember that bar fights are at least as much chance as they are skill. When you come to the point of throwing hands, you admit that you are accepting unknown danger. Does the other person have a knife? Or a gun? Are they in a bar that is their local and surrounded by friends? All of those are important questions if you are about to offer up combat.

The PhD is knowing when and where to apply force and by what means. Seldom did the OSS carry out their missions alone. They worked with partisans. Those partisans were willing to carry the bar fight to the enemy while being orchestrated by the OSS. To believe otherwise is to believe a lie. The OSS in France did not want to fight. They wanted to complete their mission. If they did that and never fired a shot in anger, it made no real difference.

The OSS were glorious amateurs. Georg Olden, Ralph Bunche, and a cast of Americans from many backgrounds led the OSS to success upon success. Clearly, it takes all kinds. What I worry about with the idea that “the bar fight is the PhD” is that we start to misconstrue willingness to do violence to others with capacity to lead resistance.

When I was out with my Australian mates, I found we were talking about when we were young. When I was very young, I was in a fight with my brother that resulted in bloody noses, black eyes, and a lesson that I have never forgotten. My father, after telling me off for hitting my brother, tied a cloth around my eyes and made me try to walk down a hall. Predictably, I knocked into a table. He told me I could have blinded my brother and that is what it would be like for him. Pretty jarring as a young kid. Jarring enough to stay with me until today. What I learned was not, “don’t hit your brother,” which was the likely lesson. I learned to pull my punches, identify and avoid confrontation when possible, and, most importantly, that every use of force comes with risk. My brother had bloodied my nose as much as I had bloodied his. My lesson was not do not fight, but do not be caught fighting.

The same is true of the OSS. At the core of many arguments around the bar fighter surrounds both honor and the knowledge of that ability to deal out pain. As to honor, Falstaff said, “‘tis no matter; honour pricks me on. Yea, but how if honour prick me off when I come on? how then? Can honour set to a leg? No. Or an arm? No. Or take away the grief of a wound? No. Honour hath no skill in surgery, then? No. What is honour? A word. What is in that word honour? What is that honour? Air. A trim reckoning! Who hath it? He that died o’ Wednesday. Doth he feel it? No. Doth he hear it? No. ’Tis insensible, then? Yea, to the dead. But will it not live with the living? No. Why? Detraction will not suffer it. Therefore I’ll none of it. Honour is a mere scutcheon. And so ends my catechism.”

“Discretion is the better part of valor.” The core purpose of the OSS was not to engage in bar fights; it was to win the war. They did that by living and continuing the fight. Not by chasing honor like Lt. Dan looking for a glorious death, they scrap, survive, avoid the direct fight, and win in the shadows.

A final observation: the idea that a FAANG engineer who shoots on the weekend is the solution to this particular problem, or the one that Admiral Bradley was addressing, is wrong. We need people like that, sure. We also need the person who joins the Peace Corps, learns another language, and lives in another culture, or the SEAL pup who does well in school but needs adventure. As the SOE and OSS taught us, what all of them need is training. Having had the privilege of hearing Major General Singlaub speak on multiple occasions, I can tell you that OSS training was at least as much about teaching you what you were incapable of as it was about teaching Ivy League PhDs what they could do. Per MG Singlaub:

“Well, I think the way in which individual training was administered to people who were already militarily qualified. The training taught you skills that you didn’t already have. Most importantly, it enabled the instructors and others to evaluate your ability to absorb information and instructions, perform skills, and demonstrate your ability to handle problems.

In fact, if there weren’t problems, they’d create them. Occasionally one of the psychologists or psychiatrists would be dressed and introduced as one of the students. And that person would invariably screw things up. It was then up to the guy who was the acting leader at that time to exercise leadership and solve the particular exercise problem without this guy’s help or to reduce his participation to that of an automaton. This deliberate effort to use this training as a means of psychological assessment was unusual for me.”

Many of those who joined MG Singlaub quit. That’s normal. There is a belief that capability in one field, say FAANG engineering, means that you naturally will have capacity and capability in another. It’s true that a high-level education in any field could be a key indicator of aptitude, but equally important are pain tolerance and willingness to persist. That is where many struggle. Persistence in the face of failure is core to winning the imbalanced fights the OSS undertook. Setbacks were regular.

What this means, in the end, is that SOCOM has to find smart people and teach them to fight. They also need to find people who can fight and teach them to think. It’s only when you have a mixture that an organization charged with a mission as wide as irregular warfare can ensure it has access to the best of what America has to offer.

The current one started with the U.S. Army not including Special Operations Forces in its multi-domain operations doctrine and has since spiraled. The Irregular Warfare Initiative at West Point recently published a good example of this with Ned Marsh’s The Last A-Team: Special Forces Aren’t Special Anymore. Ned, a retired colonel and former deputy commander at Special Operations Command Europe, makes an argument that hinges on this assertion: “A twelve-person A-Team with American equipment and American biometrics will not survive infiltration into China, Russia, North Korea, or Iran.” He goes on to compare the size of U.S. Special Forces to the British Army—there is a lot to discuss there on whether that is a good metric, given the UK’s struggles in fielding a force—and to claim that Special Forces is constantly deployed and therefore unable to experiment. This means they are dying, and we should help them by “burying the beret” and building what is next.

I will start my critique at the end. Ned claims, “The joint force does not recognize SF as a critical component to operational success because it never did.” I am sympathetic to this argument. I think it is, by and large, correct that parts of the Joint Force wanted to dismiss Special Forces and SOCOM more generally. However, when I look at history, I see that while the Pentagon may have felt this way, political decision-makers did not. A quick roll-up of the last several decades would indicate that Special Forces was given core roles from the invasion of Afghanistan to Plan Colombia, and from MAC-V SOG to Det-A and the 46th SF Company. They maintained roles and positions that were important despite being overwhelmingly dismissed by military planners.

Despite the long history of never being recognized as important to operational success, Special Forces continues to be used for exactly that. It was Special Forces and SOCOM that, along with Kurdish and Arab partners, defeated ISIS in Syria. The politics shifted and the U.S. largely left Syria, but that is not a sign of operational defeat, just a reality of war as politics by other means. In the end, it is dangerous to confuse decision-making in the Pentagon with national policy, as they are often focused on different things. The Pentagon can be tribal in ways that throw up artificial walls politicians do not see and will not abide by.

The other core conflict in the piece is the claim that Special Forces needed to transform to Charlie Cleveland’s vision of Unconventional Warfare, with a focus on “the ability to conduct UW in denied areas for extended periods,” but was not “optimized for the dispersed operational footprint future operations would require.” This is the great indictment of the article: that Special Forces never truly went through the transformation that was demanded. Instead, they focused on presence, and presence is not enough. However, that is where you build the “genuine cultural fluency” necessary to work with partners. The issue was not that Special Forces were deployed too much; it is that they were deployed, rather than living in key countries.

It is mistaken to believe that the purpose of Special Forces, or the OSS before them, was to infiltrate hostile countries and perform UW. That could have been a byproduct in rare occurrences, but they were designed to work in areas that were permissible and semi-permissible and to build capabilities that could remain during an invasion. Or, in the case of the Jedburgh teams in Vichy France, to work with partisans to weaken an occupying power. The idea that these teams worked in completely denied areas during this time is incorrect. Because they had the French partisan networks, they could move, though with risk. They were not invading Germany and Japan and attempting to build resistance from scratch. That would have been suicidal.

The natural outgrowth of the Jedburgh teams came during the Cold War in the form of Det A Berlin and the 46th Special Forces Company in Thailand. These were units that lived and worked every day in Berlin and Thailand to prepare either for an eventual Soviet invasion or for deep operations against Communist militaries in Laos, Cambodia, and Vietnam. More, not less, time in the country was necessary for their missions. The core issue of UW is one of time: time to know a country, develop capabilities, and gain cultural fluency. The issue with approaches that argue SF needs to deploy less to “conduct deliberate experimentation, honest assessment, and radical redesign transformation requires” is that they deny where transformation actually happens. For an example, look at Ukraine.

Do we think the Ukrainians needed less time in contact for transformation? Is it more likely that we would see and get better transformation lessons if we were more closely deployed with Ukrainians, watching how they are tackling the issues they face in combat, with ubiquitous surveillance, jamming, and the emergence of new technologies and equipment? I believe transformation in the abstract, without close contact with the problems currently facing fighters, is a recipe for disaster. Rather, if, after 2014, Special Forces units had truly lived in Ukraine, embedded with their military, and prepared for the next steps of conflict, they would likely have a better understanding of current conflict today.

At the core, these types of decisions, to bring all SF back for a reset, to cut deployments, or to increase forces overseas and embed with partners, are not Special Forces’ or SOCOM’s decisions to make. These decisions are made in the Pentagon and with political leaders. That is a core issue here. It was a president, FDR, with the championing of Bill Donovan, who started the OSS. It was a president, JFK, who started the Green Berets. Both of these presidents had political reasons to do so. What is the political impetus we should look to today?

Subscribe now

Share Mind of Things

Leave a comment