“AI is not Magical Pixie Dust”
A conversation with the CTO of Lockheed Martin and the Director of Project Maven on the Future of War
The architect of the Pentagon’s AI strategy—and now the Chief Technology Officer for the world’s largest defense contractor—meets the leader of the Pentagon’s most famous AI program. For the first time, hear their unfiltered insights in a masterclass on modern warfare from two of the most influential minds in defense technology.
I’m thrilled to bring you this conversation with two of the most influential voices in national security. It’s one of the most compelling (and unfiltered) discussions I’ve had on how AI is actually being used in the real world of combat.
Dr. Craig Martell, CTO of Lockheed Martin & former Chief Digital and AI Officer at the Pentagon.
Colonel Arnel David, U.S. Army officer & the man deploying the Pentagon’s most ambitious AI initiative, Project Maven, within NATO.
Together, they dismantle some of the biggest myths around AI and warfare, what’s happening on the front lines of Ukraine today, and where we need to go from here.
We discuss how:
NATO adopted Palantir’s Project Maven’s Smart System in 6 months, instead of 20 years.
Why solving capability gaps matters more than chasing shiny tech.
What “human-AI teaming” really looks like in live combat environments.
Why AI isn’t “magical pixie dust,” and what that means for leaders, innovators, and strategists.
This is the kind of conversation I wish more policymakers and founders were paying attention to — because it is so fundamentally tied to the future of war and peace.
🎙️ Watch the full conversation here:
This conversation was hosted by the University of North Texas on September 13, 2025, in Dallas. With special thanks to William David Hamilton, Director of National Security & Economic Strategy, for convening the panel.
Transcript
Nina Schick
Ladies and gentlemen, this is going to be a great panel, so you don’t want to miss this. It’s my honor and pleasure to be moderating. I’m going to invite up my two panelists. I don’t think 45 minutes will do it justice.
Perhaps I could preface the panel by saying both of the leaders beside me are architects of an exponentially changing environment. Not only are they strategic leaders; they’re actually building things that are having effects in real theaters.
To my left, Colonel Arnel David. He’s the Director of Strategic Initiatives at the NATO Supreme HQ, and he leads Task Force Maven. He’s achieved the impossible: he got NATO to adopt the Maven Smart System in just six months, versus a procurement cycle that usually takes 20 years. His “digital insurgency” concept is revolutionizing how the Alliance fields capabilities.
To my right, very honored to have Dr. Craig Martell. He’s an academic, he’s worked in industry, and he has worked in Silicon Valley. He brings unique perspectives as the Pentagon’s first Chief Digital and AI Officer, where he helped rebuild how we think about AI and exponential technologies from first principles, and he is now leading Lockheed Martin’s CTO office—recently in post. I don’t want to take away from my panelists because they’re the perfect combination of innovative military thinking and industrial base excellence.
So perhaps we could start with you, Arnel. How did you do the impossible? How did you take what usually takes 20 years and distill it down to six months?
Colonel Arnel David
Thank you, Nina. Happy to be here. It’s an honor to be here with you. It’s a great question. The impossible—it feels like a series of miracles. If anyone’s familiar with NATO, the North Atlantic Treaty Organization—you know the jokes: “No Action, Talk Only,” “Nothing After Three O’Clock.” Those are the things you hear about NATO. The difference we made as a small team is that we were absolutely ruthless in execution, and you have to be relentless.
There’s no easy button. It doesn’t matter if the generals want this or the secretary wants that. You can’t move the permafrost—the frozen middle between senior leadership and the bottom. That permafrost of bureaucrats willing to say no—they exist everywhere. I would hear all the time: “NATO doesn’t move this fast. Slow down. We don’t do it this way.” I never accepted that as an answer. I’d tell them that’s unacceptable, because if you want to win a war, that’s not how you behave. You have to take a higher degree of risk.
That’s what we did. We moved out fast, we didn’t take no for an answer, and we got this done rapidly—and we didn’t break process. There’s a lot of process—TPS cover sheets—in NATO to get something done. We didn’t allow drift to happen—no “after the long weekend,” no “next week.” We said, “Do it by Friday.” There’s no substitute for being ruthless in execution, and I think that’s why we got something done so fast.
Nina Schick
We were just speaking backstage, Craig, and you said AI is neither a panacea nor a Pandora’s box. Perhaps you could elaborate what you mean by those comments?
Dr. Craig Martell
I’m fairly well known for being counter the AI hype. I don’t think it’s a panacea that solves everything, nor do I think it’s the one dangerous thing such that if we have it, we win, and if they have it, we lose. You hear people demand AI when what they really mean is: “give me magical pixie dust to solve my hardest problems.” Or you hear vendors say, “If you buy this, you’ll get magical pixie dust that solves all your hard problems.”
It’s important not to treat “artificial intelligence” as a monolithic technology. In fact, I think we should get rid of the term. “Artificial intelligence” was coined in the ’50s as a fundraising buzzword—John McCarthy used it that way—and it often still functions that way. Because what is AI? Actually, let me ask that question. What’s AI—if it’s a singular thing, what is it? [Audience: “Math.”] Well, of course it’s math, thank you. What’s AI? [Audience: “A means of gathering data.”] A means of gathering data—okay. And what’s the promise of AI? [Audience: “Solves everything.”] Sorry, sir…. Compresses time to solve really hard cognitive problems, I think, right? I’m gonna add that piece to it.
All AI is statistics at scale. That’s it. It’s statistics at scale. We gather data from the past to build a model to predict the future. That’s it. That’s all it is. And that means there’s not one single way of gathering that data. There’s not one single way to build that model. Some data and some models solve some problems. Other data and other models don’t solve those problems. There is no monolithic technology called AI.
How many people remember linear regression? I’m sorry. I used to be a professor, so I have professor-itis—you can tell. How many people remember linear regression? Linear regression is AI. That is AI. I’m not kidding you. Many AI systems are just linear regression with a lot of variables, right? Deep learning is logistic regression, a variant of it—just billions and billions of variables. So all you’re doing is gathering data from the past to build a model to predict the future. Uh, I’m sorry, Nina, I’m—
Nina Schick
I told you you don’t want to miss this panel….
Dr Craig Martell
So let me just add one more thing. Where can that go awry? That goes awry under two conditions. You gather the wrong data—great, I gathered a bunch of data, I sampled wrong, I looked at the wrong things; it doesn’t solve my problem because it doesn’t predict the future that I wanted to predict. Or—here’s the really important one—I gather the right data, I build a model, and then when I go to deploy the model, the world changed.
So how many people believe in warfare the world changes? It changes all the time—very fast. Let’s say I built a model—Project Maven does a lot of the vision stuff—that determines that this is a roadway and this is a cattle path. So I built a roadway detector. Now there’s a bunch of bombs that get dropped like that. Do you think that roadway is still detectable by the same old data? No. And there is no magic that suddenly goes, “Oh, that used to be a roadway; now it’s a roadway with big craters in it.”
So it’s really important to ask the following question: What problem do you want to solve? What capability gap do you want to address? Not “What technology should I buy that’s magic?” Stop asking that question and ask what capability gap do I want to address? And then answer it by saying, What data do I need to solve the problem and what algorithms are gonna work? And then try it and see if it works. So really my message is: there’s no easy button. It’s still hard work. Okay. Sorry, Nina, that was just the first question.
Nina Schick
No, that’s terrific. First-principles thinking—start with the problem. Perhaps I could ask [Arnel] a little bit more about that in the context of Project Maven. How have you taken that approach? This is where the rubber hits the road, right? We’re really seeing it now in Ukraine. What can you share with the audience about that — and try to talk about it without using “AI”?
Dr Craig Martell
You can say “use case” or “statistics at scale.” That’s what you’re allowed to say.
Colonel Arnel David
I agree wholeheartedly with Dr. Martell. What we’re doing with the Maven Smart System that we bought from Palantir for the Alliance is, for the first time ever, we’re taking all the data they have on these different networks and these different systems that are decades old, and we’re starting to clean that data to make ’em smart objects and machine-readable so we can exploit them with different applications on the platform.
For example, targeting is one of them. We want to prosecute 1,000 targets in an hour. That was the guidance I got from my boss, the Commander for Europe. That’s the standard that we’re at. He wants that for NATO. That’s a problem to solve—how will we do that? We need to get the sensors linked into the Maven Smart System. So we’re gonna pull all that data from these different sensors. So it could be satellite-type sensors, it could be, you know, radars.We need that data.
And then things like logistics. You know, we want to know how long it takes to move a brigade combat team across the transatlantic and then across the continent of Europe—through rail or through HETs (the trucks) to move these tanks. We need to know how long it takes. And so these types of narrow applications solve the different use cases— we’re starting to see that we can do this rather rapidly, and I think that’s a game changer. For the first time we’re now fixing our data ontology, our data environment, that we can now exploit and interrogate with different applications.
Nina Schick
We know data is the foundational layer. How difficult was it— and you can’t skip the hard work — especially in the context of NATO, to actually bring that data together?
Colonel Arnel David
There are a lot of haters out there. There are people who, for whatever reason, might be jealous that we bought this American system versus another. We have 32 Allies; I live and work in a multinational environment. It isn’t enough to be a salesman and say, “Hey, we really need to use this.” You have to be ruthless and ferocious and say: if you don’t make this work, we’re not going to win the next battles—and in the next war, we’ll have an unnecessary loss.
So we said—the stakes are too high—and issued orders from the command: you will integrate these data sources—these authoritative data sources—onto this primary platform for war fighting. People say no; we smash through the resistance. But it’s happening. More and more, seeing is believing. We’re not salesmen; we’re warfighters.
Once you start standing it up—and you’re not using PowerPoint to fight the war—you actually have systems and sensors integrated for a common operational picture: to see ourselves (so we don’t shoot each other) and where we think the enemy is. That’s really helpful if you’re going to fight a war. Seeing is believing, and that’s what’s starting to change the mindset. Because you can buy an expensive technology, but if the mindset and culture aren’t right in your organization, you’re wasting your money.
You have to change culture and behavior—which is arguably the hardest part.
Dr Craig Martell
Can I just comment on that? I want to echo what you said: you got everybody to put the data in the right format, in the right place—and that’s what makes the magic happen. It’s not that you bought some app or foundation model or “magic in a box.” You did the hard work of putting the data where it needs to be. That’s extremely impressive.
Nina Schick
Both of you have talked about the human–AI teaming models. What do you mean by that?
Dr Craig Martell
The right way to think about how these statistical methods—I’m trying really hard not to say “AI,” because I think it’s a terrible term, even though my whole career has been in AI… I think the right way to think about these statistical methods is: they’re statistical, which means they’re guaranteed to be sometimes wrong.
So let’s talk about a classic human–machine teaming that you do all the time. How many people have adaptive cruise control? You turn it on and it’s a game changer to not have to worry about the car in front of you. That’s really nice. That’s the machine doing its job for you. Sometimes it gets it wrong, it gets a little too close to the car in front of you. What do you do when the system gets it wrong? What you do is override the system. You give it some input— you put on the brake, you turn the steering wheel in a car. I used to have adaptive cruise that was awesome, but the lane-keep assist was terrible. And so I would have to intercede with the lane-keep assist all the time, but I didn’t have to with the adaptive cruise.
So human–machine teaming really is the right way to think about these statistical tools, because they’re guaranteed to be sometimes wrong. I’ve seen Project Maven in action, and the system will make some determination. And that determination might be 60% correct? Well, it still did—it got to that 60% solution way faster than a human could. And then a human uses their cognitive load to look at that and then correct the system. “Well, no, it got it right that time. Oh, that’s one of the 40% [when it didn’t ]—here’s the right answer.”
And then that gets fed into another system, which makes another judgment. And then there’s another soldier on the other end who’s an expert at this job who corrects the system. That human–machine teaming gets you to the right answer where neither of them by themselves would. Now, the humans will probably get you there—but slower. And the machine probably wouldn’t get you there often enough alone.
Colonel Arnel David
There’s not much more to add on that. I would say that there’s always a human in the loop with everything we’re doing— and that the machine, and this teaming concept, it’s helping us think smarter, make better decisions, and make them faster. This human–machine concept is something that translates to the commercial sector too – I mean, why would you not want to do something faster if you can? So if you can get the data right, it’s informing you to make a decision. It has a confidence level. You make a decision—you still have the art of making that decision as a human.
Dr Craig Martell
It’s probably the case that the machine alone is not gonna get it right enough, right? You need humans actually guiding the machine. And that’s the right way, I think, to think about these tools: they’re being guided by the humans that are using them.
Nina Schick
So thinking better, thinking faster—the human–machine loop. Obviously this raises the spectre — it underlies this entire event— of strategic competition. So how are you thinking about asymmetrical advantage? And you’ve recently taken the post at Lockheed Martin, both in the context of broader geopolitics, but also what are your strategic priorities for the company?
Craig Martell
That’s a great question. I’m a little bit new. I’ve only been the CTO for three months, so I am not going speak with authority about what I think our strategic priorities should be. Next year, I’ll have a stronger opinion. But I will tell you what I think my job is. My job is to first sort of develop three distributions—you can tell I used to be an academic because I talk in math-speak— but three distributions. One is a distribution over the kinds of gaps. Some of those gaps are immediate weapon gaps— we gotta think about those. Some of those gaps are longer-term gaps.
[To audience:] So how many people used to be a warfighter, or are a warfighter? Okay… So every one of you knows that you had to use some system that just didn’t work the way you wanted it to work, and you were forced to use it. And if it were the case that you could make a phone call into the past to the person that designed that system and said, “Please just change these one or two things,” that system would be a hundred percent better, right?
We see our job as taking that phone call from the future. So some of our job has to be solving the fight-tonight problems. Some of our job has to be thinking about: what does future warfare look like, and what does it mean for it to be right then? And so trying to figure out what the balance there is—is the job. It’s gonna be pretty exciting.
Nina Schick
And a follow-up to both of you on that—we were just discussing it offstage— what’s the process of figuring out what future warfare is going to look like? What’s the next five years going to look like? What’s the next 10 years going to look like? And how do your engineers and designers actually start implementing that?
Craig Martell
I’m actually gonna defer to the warfighter on that. From the technology side, it’s to think about desired capabilities and not technologies, and then think about the technology second.
Colonel Arnel David
That’s a great question. So just to build on that: when you put engineers and operators together, magic happens— because then they’re talking to each other about the actual problem they’re trying to solve with this technology. And so that’s always goodness: scientists, engineers, operators—anytime you can do that, what I’ve learned over my time, that’s a good thing.
So how do you think about the future of warfare? I mean, I read a lot of science fiction, and there’s a book called Ghost Fleet by Peter Singer. That book influenced the U.S. Navy in more ways than many—like the, you know, massive departments of people studying war—did. That one book. And so if you’re trying to imagine a future, I mean, science fiction actually isn’t too bad.
Science fiction and history—read some books—those of you who are students and academics: there’s goodness there and it’ll help you down the road. But the other thing is wargaming. And it’s one of the things I do— with Fight Club – is that if you want to manufacture conditions for stirring up some anomalies that you might want to explore or extrapolate further, wargaming and simulations can help with that.
To get you thinking — even in industry, even in the commercial sector — how often having competitive play can help you think differently about a problem. And it can be fun. You can use gamification to just kind of incentivize your workforce a little bit to have some fun and then to think differently about a problem.
Thomas Schelling used to say it’s hard to imagine— no one human, as heroic as they are, can think of a bunch of things that they would never think of. And sometimes you have to create the environment that can emerge, and gaming allows you to do that.
Nina Schick
I find this so fascinating. Perhaps you could elaborate a little bit more to the audience with what you are doing with Fight Club – and creating those conditions for ‘creative thinking.’
Colonel Arnel David
Absolutely. I’m happy to make a plug for this organization. I started in the British Army—I was an exchange officer several years ago—and we were kind of dissatisfied with the wargaming community because there were a lot of rolling dice and maps. And that’s all good because there’s still a time for some of that, and people should make decisions and role-play.
But there are a lot of commercial off-the-shelf (COTS) solutions available that are pretty damn close — physics-based engines that can tell you whether it’s an F-16 firing a JASSM at a certain distance with radars – you could test these things. And why would you not want to do that a thousand times before you fight a war?
So what we’re doing at Fight Club is that I have a bunch of civilians, academics, students— all kinds of leaders. One of the first people that joined was the Secretary of State for Defence equivalent in the UK — joined Fight Club — and that just spiked up the interest. We’re experimenting with different types of war games and, you know, simulation software that’s available off the shelf. And then we’re testing it.
We did a fight in the South China Sea with the Philippine Marines and some Australians and people online. And then within a month—which usually would take a year and a million-plus dollars—we assessed what that would look like as one good round, a good repetition. Then you can do more and more, often to help understand what are those changes that might happen. Like I said, it’ll emerge from just fighting each other regularly.
So we’re fighting all the time—and not just physical kinetic warfighting. We’re looking at information warfare; we’re looking at humanitarian aid – how to do that better. You can have games for all kinds of stuff. And I think, everyone should think of that in your organization—are you doing those types of things to help stir some new thought..
Dr Craig Martell
Although, Colonel—you broke the first rule of Fight Club….
Nina Schick
Don’t talk about “Fight Club” ! … Backstage we were talking about one of the things that you think really distinguishes the U.S., and that has to do with the innovation culture. So perhaps both my panelists could comment on the innovation ecosystem. How are you thinking about the importance of the innovation ecosystem, and do you really think this is a decisive edge for the U.S. as compared to opponents around the world?
Dr Craig Martell
I think our major asymmetric advantage is our academic–industrial innovation ecosystem and the ways in which we tap into that. There’s a lot of talk about some of our adversaries—particularly our pacing adversary—having most of their work be dual-use. But I think the real question is: to what degree does that technology integrate into warfighting? To what degree does that technology integrate into systems?
And I’m always super impressed by the way that we tap our academic capabilities in the United States. And, by the way, our startup infrastructure and ecosystem taps into those capabilities—and the way that those new technologies—and not just startups, even “old crusty” companies like Microsoft, Google, or Lockheed—tap into this innovation ecosystem, and the ways in which that new technology makes it to the warfighter. It’s extremely impressive. And I actually think that that’s our major asymmetric strategic advantage.
Nina Schick
Talking about just that innovation ecosystem and how quickly you managed to deliver the impossible by taking risks, as you said—how do you think you could replicate that model across other critical capability areas?
Colonel Arnel David
That’s a good question. So the Chief Technology Officer for the NATO Communications and Information Agency (NCIA—NATO’s tech and comms agency) came out of my office a couple of weeks ago. He said, “Hey, we need to do this more often.” And we know we “broke glass” here when people like him—and other three-star-equivalent civilians at NATO Headquarters (very senior civil servants)—are coming down and asking for demos. They’re trying to figure it out and saying, “We need this one.”
So what we’re going to do next is we’re going to take—I don’t think it’s a hundred-person problem. I think you pick 10 to 12 people who understand the acquisition process. And I did this before. You heard General Mike Murray get his award for Futures Command. I— also as a lieutenant colonel —was part of the nucleus for the planning with a friend over here called Paul Norwood. In 2015–2016 we were blueprinting that command. It was a small group of us who had the right mix of skills and understanding of acquisitions and innovation—and how to build a command and re-blueprint and break up the Army.
And so we’re looking at NATO—how do we do that? You can’t do a massive reform and be so disruptive that you paralyze the Alliance and it can’t act. So you refine on the margins and find ways to do things like an OTA (Other Transaction Authority—fast-path contracting). NATO has enough policy—and we did an urgent requirement to buy this. There are ways to do it. So we’re studying: what can you do to go faster?
And the one thing we’re doing right now that’s been the success is this task force. I only have seven people on my books who belong to the task force, but it feels like there are a hundred people working for me because, out of nowhere, people are showing up with this revolution—this urgency—to say, “I want to work for you.” And it’s not “for me” per se—they believe in the idea of change. They’re tired of the status quo and they want to plug in and help. So I have a load of people, and I’m like, “What can you do?” “I’m good at social media.” Excellent. You need to beef up our social media—join the team. So we have loads of people.
But this task-force idea—and in the military you might understand what a task force is—you task-organize; you assemble yourselves against an emergent problem. In this case, it’s deploying this AI-enabled system across the Alliance. And the idea is: our current structures may not be fit for purpose when a problem arises. The question is: are you going to have the courage and leadership to assemble properly to attack that problem? And that’s hard. We did that in this case—and we’re still learning a lot.
Nina Schick
And so much of that comes down to what you hear all the time in industry—and what you also mentioned is our asymmetric advantage—and that’s people, right? And talent. So how are you thinking about developing the next generation? What’s critical?
Dr Craig Martell
Yes, I think that’s great. Let me say that I’ve heard a lot about AI framed as a Manhattan Project, and I think that’s fundamentally flawed. I think we need to think about it as an Apollo Project. And what do I mean? The difference there is: the Manhattan Project was a targeted weapon for a particular use case. It was secret, and it had some broad technological side effects, but it was still a pretty narrow band. The Apollo Project was a whole-of-society project that included massive funding for science and technology across educational systems—across high schools and across universities.
That’s the kind of energy I believe we need to put back in. My mother—who’s long retired now and has a PhD in math—but her bachelor’s in math was paid for by the Apollo Project. And what did she have to pay back? She got a scholarship—what did she have to pay back? She had to do two years at a contractor. She didn’t even have to do two years in the government—she just had to do two years in a contractor that was working on space technology. This is the kind of broad investment I think we need to make.
We at Lockheed think about it all the time this way. We have Code Quest, we have AI Quest, we have Cyber Quest. The side effect of those is we get high-school interns. Think about it—you’re a high-schooler and you intern at Lockheed Martin—well, that’s gonna help you go to college, I promise. And then when you’re in college, you intern for Lockheed Martin—and then we hire you. These sorts of early investments are exactly the right kind of investments to get the people who need to have the right skills to be able to build what the future’s gonna look like.
Nina Schick
I love how you referenced that whole-of-society mobilization. You’re an engineer. We have the brightest generation of engineers. How do we make them want to work for something beyond just building a food-delivery app—or five AI friends?
Dr Craig Martell
Part of the problem is the food-delivery app pays them four times the amount of money the defense industry can, right? So we have to think about ways to solve that. And it’s not just finance.
One of the things that I want the CTO of Lockheed to think about is: how do we make the work we’re doing the coolest? Because what engineers want even more than money— because they’re always gonna make enough money—what they want more than money is: I’m working on the coolest thing —
Nina Schick
— Mission?
Dr Craig Martell
Well, mission, yes—but also the coolest technology. Some people are very mission-driven, but engineers on the whole are cool-technology-driven. And they’ll take that over money. So part of what we have to do is just show that the technology that’s necessary for the mission is some of the hardest, coolest, most exciting technology on the planet.
Nina Schick
Okay—going back to future warfare. Colonel, your concept of Prototype Warfare could revolutionize how we think about fielding capabilities. Could you elaborate a little bit more on that?
Colonel Arnel David
For my PhD I’m doing right now at King’s College London, I’ve been studying this intensely over the last four years. I believe in taking on a prototyping mindset and approach—whether that’s using gaming and simulations and just building things and/or testing concepts out.
If you look back in history, we used to do this a lot more often and we had a higher degree of risk. With the CORONA satellite—it’s all declassified several years ago—we wouldn’t have put a man on the moon if we weren’t firing rockets and taking risks and prototyping them, and then learning— taking the data from the telemetry and learning to improve those rockets.
I mean, Elon Musk right now—he’s just firing rockets, firing ’em. He doesn’t mind if they blow up—because that’s the mentality you need. You need to prototype. And maybe it doesn’t have to be rockets.
I’m a strategist, so I have to think about—now, you know, I spent 12 years in special operations, but also now looking at combined arms formations moving across the European continent—and that’s really hard. That’s complex warfare. So, I don’t believe our plans are going to survive contact with the enemy—that’s an old saying from Moltke the Elder: ‘no plan survives contact.’ And I still believe that. So what do you need? Prototypes of plans. And the systems we have now can help us think through different options—so you’re like the quarterback: “I don’t like the conditions—I’m gonna call an audible.”
Right now the theory that undergirds all military planning is 1950s–60s rational-choice theory. Since then, new theory has emerged in academia; we have new technology—but we have not changed our fundamental principles of planning in the military. And so I’m an advocate for changing that. And I think prototyping—this Prototype Warfare concept—would be a game changer if we were to adopt it.
And just one last point— back not too far from here, we had the 1st Cavalry Division, a world-class, powerful division in the United States Army. If you remember the movie We Were Soldiers Once, that division was built pre-Vietnam. They were, I think, the 11th—or 10th—Airborne/Airmobile Experimental Division at Fort Benning, Georgia. And we experimented—we had experimental units. We need to bring that back. And if anyone needs it, our allies need that, because their land formations have been decimated over the past two decades—because we’ve been playing tracer tag in the desert, you know, fighting insurgents and terrorists. So we’ve kind of atrophied in our land-combat formations.
Where we need to start building experimentation is—we might be integrating one-way attack drones and electronic warfare and starting to see what kind of combinations of packages would work. But I think prototyping in that mindset, like General Thurgood was alluding to, would be really useful for us as warfighters—just to test stuff out. And then you’ll have that engineer–operator–scientist convergence, talking to each other—working to do hard problems.
There’s also the Mike Tyson version: ‘Everybody has a plan till they get hit in the face.’ I like that.
Nina Schick
On that future-of-warfare point, perhaps I could throw it to you, Dr. Martell, and say—well, the paradigm of warfare itself is changing, and if I were to be playing devil’s advocate, I’d say: why would one need such expensive aircraft carriers when the future of warfare is obviously gonna be about cheap drones? How would you respond to that?
Dr Craig Martell
I think my answer to that really is: maybe—and that’s an empirical question, and we should think about ways to empirically validate that. If you look at the story of startups in the defense space, it goes something like this: “The old primes—they’re crusty, they’re old, they’re slow.” (It’s actually not true, by the way.) “They’re crusty, old, slow.” And the new startups come in — they’re fast and they’re agile and they’re quick. Sure—they’re fast and agile and quick about those things for which a 70 or 80% solution is okay.
Now, that’s where I think Lockheed could do a lot better. There’s a lot of things where a 70–80% solution might be a good starting point, and then you can be agile and move quickly with your customer. Love that idea—the division as a whole could do better at that. But there’s a lot in the defense space where a 70–80% solution isn’t acceptable, and you don’t want to start at 70% and then iterate. You want—as much as you can, with good systems engineering and technology—to get to a 99.999% solution. You can’t just wish away the requirement for things to be correct in warfare.
So let me give you an example. A really, really old tech— a mortar. We can go back to World War I. A mortar—let’s say what you know about that mortar is it’s gonna land in an oval with 99% accuracy—it’ll be within that oval. Okay—well then you feel comfortable adjusting the windage, adjusting the height, because you have a good sense of where it’s gonna land. But now what if I tell you: here’s a mortar, and it has a 70% chance of landing there and a 30% chance of landing somewhere else—and you’re using this between a hospital and a school. Okay. That’s not a tool—it’s really low-tech, it’s a mortar (yeah, you can build it in your backyard)—that’s not a tool that’s amenable to a 70–80% solution.
I just think it’s really important: absolutely, we all have to get more agile; absolutely, we all have to get fast. Let’s spend our energy figuring out what domains that applies to versus domains where—yes, we want to be as fast as possible—but we also have to be correct. And when you hear the story about startups in this space, nobody talks about the ‘correctness part’. They only talk about the fast and agile part. And I think it’s important to differentiate which capabilities you can learn in an iterative way and which capabilities you just gotta get right.
Nina Schick
Are you seeing that in the field as well?
Colonel Arnel David:
Yeah, I agree. I mean, there’s always this debate about precision versus mass. Do you need mass? Precision? In the U.S., we don’t have the luxury of choice—we need both. And everywhere I’ve been in our world, it’s kind of nice to have both. The Europeans—where I live—they don’t have mass and they don’t have precision. They don’t have either. So we have to help them figure out where to invest. I mean, they just can’t scale.
There’s a chart in my office I use because I always hear, “We keep buying—you’re making this ‘Buy American.’” I’m like: this is stuff that’s good and works; it’s not that I’m just trying to—if it’s Swedish and it’s good, we’ll buy it. And there are some things out there—there’s tech—you know, Gripen is not bad, Kongsberg is not bad—I’m joking, obviously.
But there’s a chart a friend shared with me in D.C. that shows the last 50 years of tech startups and companies and their market-cap positions over time. You know—media, Apple—the United States is like everywhere, and then Europe is like a dot—a couple dots. Home Depot is bigger than the entirety of the last 50 years of European tech—because of the over-regulation. They can’t scale. So they complain—but I’m like: well, then you have to change your behavior. You get to change some of your policies. The GDPR, all those things they were talking about earlier are causing them not to be able to scale.
So I’m seeing that—if you wanna look at the innovation ecosystem—it’s challenged. If you were there, why would you wanna work with NATO, like the Secretary said, or defense? That’s probably a bad bet. You can’t afford it; you’ll go out of business. So yeah—that’s what I’m seeing across the transatlantic.
Nina Schick
Yeah—I mean, I used to live in Europe and I’ve moved to the U.S. for a reason…. So what do you think: We’ve heard so much today about how we need to rebuild our industrial base. We need to get the right talent; we need to repatriate supply chains. Perhaps this is kind of an end to the era of globalization as well. What are the most critical areas, in your view, for investment for the United States going forward?
Dr Craig Martell
Oh, I don’t have an answer to that yet. And I think that’s a great question. I would just argue that we focus on capabilities and not technology. I think it’s a fundamental mistake to say, “Let’s throw our weight behind this tech sector or that tech sector, or this buzzword.” We should make those choices because these are the capabilities that we need. And to deliver those capabilities, let’s do some very boring, good old-fashioned engineering and science and figure out what’s gonna solve those capabilities—what’s gonna deliver those capabilities—and then invest in those technologies. Let’s make sure the horse is before the cart.
Nina Schick
I know we’re coming to the end of our conversation now—perhaps I could ask both of the panelists: what gives you reason to be optimistic, or pause for thought, as you look at how quickly the landscape is evolving—how quickly geopolitics is evolving? What are your thoughts on this? …. Deep Philosophy!
Dr Craig Martell
She didn’t tell us about this one backstage, so we’re making it up as we go!
Colonel Arnel David
I’m rather optimistic. I see our country continue to do great things. And again, this is not a political answer—for me as a warfighter, no. I’ve been all over the world in different places, and I’m still proud to wear this flag and be cut from this cloth—because our nation can do things that no other nation can do. My Army can do things no other army on this planet can do. And what I would offer you is that while it’s very hard to work with our allies—I do it every day, and it’s painful at times (I won’t name which countries, but some of them are very difficult)—we gotta do it.
And so what I would appeal to you is that our country’s greatness and our strength still remains in having an alliance of partners and networks—no adversary on this planet stands a chance against us when we’re united with our allies and partners. We are 10 to 12 times more powerful by any measure—military, economic, what have you. But that means you—in industry, students, academia—we have to court partners, and we have to find points of convergence. If there are tech startups over there—get invested. They can’t scale, like I just said. So they always come to America to scale and expand—we have to do that because we don’t want to fight alone. And—like I tell the Europeans—and my brother and son are in the Army, been to war—I don’t want them to pay for my sins on some future battlefield defending Europe.
So we gotta get our stuff together and be ready to fight so that we don’t have to fight, because that’s the only way deterrence works. Working with our allies—help them out even though it’s painful—we gotta do it, because we need them. Thank you.
Dr Craig Martell
Mine’s a little bit more boring. I think we need acquisition reform—first and foremost. And the reason I think that is: we have the tech. This country delivers some of the best tech on the planet—always, always, always. Inside Lockheed—I’ve been here for three months—inside Lockheed: unbelievable brilliance, unbelievable science, unbelievable engineering—just go walk around outside—unbelievable engineering. But the difficulty sometimes is getting that past the valley of death, and being part of FAR Part 12 versus FAR Part 15—the acquisition battles often lock that technology from getting to the warfighter. So I would start with acquisition reform.
Nina Schick
Please join me in giving it up for our panelists. They’re incredible—so proud of them for their work, and for giving us their time and their insight today. Thank you so much.
.
My MinisȚRUctyrE
https://open.substack.com/pub/originalmuman/p/42-years-later?r=699zlt&utm_medium=ios