511. The Future of Military Planning: Defense Tech Beyond the Bubble, AI Software as Combat Power, and the Rise of AI Wargaming (Grant Demaree)

Narrator: This episode is brought to you by Ramp, the spend management platform we use here at TFR. They're offering listeners $150 just to take a demo. We've never had an offer quite like this. Claim your $150 before this offer is gone at our partner link ramp.com partners tfr. Now onto the episode.

Grant Demaree: Welcome to the podcast about venture capital,

Brenna Kuiper: where investors and founders alike can learn how VCs make decisions and reach conv.

Grant Demaree: Your host is Nick Moran, and this is the Full Ratchet.

Brenna Kuiper: Grant Emery joins us today from the Bay Area. He's the CEO and founder of One Brief, the defense company building the operating system for military planning. One brief recently closed a, uh, 200 million Series D at over 2 billion valuation and acquired Battle Road Digital, a, ah, leader in war gaming and simulation technology. I'm especially excited to guest host today's episode. I'm Brenna Kuiper, executive producer of the Full Ratchet, and Grant's West Point classmate. Grant, welcome to the show.

Grant Demaree: Hey, thanks for having me, Brenna. It's good to see you again.

Brenna Kuiper: You too. So I'd love to jump right in with a quick backstory, um, on what you've been doing since we last saw each other and your path to tech.

Grant Demaree: Let's do it. Um, so we were, uh, West Point 2013 classmates. Uh, so I left school, went to the 101st, had a couple of army deployments where I was exposed to operational military planning. Ah, first in Liberia for Ebola, later in Iraq as part of the Counter ISIS mission. And I thought this, this military planning, it's. It's upstream of all the military decisions. And if we could fix this, we would make military staffs truly superhuman. And, well, we'd. We'd fix everything. Almost everything. So I got out of the army and, uh, set out on a number of adventures to try to fix it, which I think you'll hear about today.

Brenna Kuiper: Yeah, awesome. Um, can you talk briefly about what your mission at One Brief is?

Grant Demaree: We're here to fix military decisions. Um, and we see this as making each military headquarters superhuman so that we can deter wars and when necessary, win.

Brenna Kuiper: Amazing. Obviously, very much needed. How would you say your West Point background and your army background prepared you for becoming a founder and where you are today?

Grant Demaree: There's two sides to this, right? There's the generic founder side, and then there's the one reef in particular. And I think it helped with One reef in particular pretty well because I just wouldn't have thought of the customer and the user and the problem this way, if I hadn't spent a bunch of time with it. Spending a lot of time witnessing these large scale operational military decisions was kind of why I imagined what would a military staff look like if it were truly perfect and how much of a change in the military's combat power would, would that be? And I actually think it's huge. I, I think you could probably multiply your combat power about Eightfold by getting to perfection. And I, I think that the idea of military planning as like the, the thing that's upstream of everything else and fixing it as the prime mover to change everything else is just really, really unnatural. And it's not really where the, the, where the market was thinking at the time we started this. And I don't think you get to that conclusion as an outsider on the question of did the background prepare me for this company? Like hell yeah, I did this because of that. Um, I don't really want to construe, uh, West Point as a founder factory, though that obviously doesn't bear out in the numbers. We have just this huge number of super ambitious people. And that has not manifested in the density of unicorn startups that you see at Harvard or Stanford. And I don't think that's for lack of trying. Um, I have some theories about why that we can get into, but, um, I'll hold on that for now.

Brenna Kuiper: Yeah, we'll have to remember to come back to that. Um, you kind of touched on this briefly, but you know, you were early to war gaming, course of action development, operational planning. As a real software market, what did you see that others were missing?

Grant Demaree: So I guess there's one sense in which I was totally not really, um, because everything that we did is something that a bunch of people had already tried to do a lot of times, starting with Secretary McNamara. So like almost everything in one brief, from like automated or semi automated operational military plans to full global integration was all something that after the Cuban missile crisis, Secretary McNamara tried to put in place. And there were at least 12 efforts to do this that I, that I of and probably more that I don't know of. Um, so in that sense, like, we were totally not early at all. I know we were late, but I also want to want to answer your question fairly. Um, and I guess there's two sides of being early here. There's the customer side and the investor side. And on the customer side at the time, and actually still now, um, within defense software, I think the consensus was around current op software. And the idea was if you Build something that integrates a bunch of data sources in order to make our kill chain faster. You will win the war. And I kind of endorse that. I think that's really valuable. I think it's borne out in recent conflicts that that's a really, really good thing to have. And I also don't think that's sufficient to change the strategic outcomes. And I think fully superhuman operational military plans looking on a longer time horizon in the kill chain are the only way we're going to get really good outcomes. That's uh, that's what we set out to, to build. And I don't think the customer uh, side market was as excited about that at the time we started as they are now. And then on the investor side, I mean this was a little bit pre defense tech being exciting in the first place. So we still had to sell defense as a uh, a worthy field all and then we also had to sell this idea of the top down go to market is never going to work with what we're trying to do. But there's viral growth and this big network effect and if you start small it's going to blow up. And selling people on this bottom up go to market in defense was, was very hard.

Brenna Kuiper: Yeah, that's not shocking to me. Just with my own experiences and you know, now that we're on the investment side, um, yeah, that, that doesn't sound too surprising. Um, what's one thing about military planning that would shock a civilian tech audience the most?

Grant Demaree: I mean I guess the classical answer is how manual and not automated. This is where everything is word in PowerPoint. And you'd say well not anymore. We have all these cool tools, not really on the major operational military plans all the way up to the operational strategic decisions of the national command authorities. That's mostly word and PowerPoint still. But I actually don't think that's the most difficult conceptual leap for this civilian tech audience. Because like in civilian tech like you're like oh yeah, I found something. This is if it's super manual, that's a problem that we can be solved. And I think the bigger conceptual leap is that military plans and military decisions in general are very, very hard to evaluate in a way that enterprise uh, ones just aren't. Because the enterprise ones eventually cash out in a revenue number or cash flow number like you, you can them financially in some amount of time and make predictions about what the financial effect is going to be. And there just isn't that generally agreed on terminal value, um, against which you could Evaluate every military plan, um, which makes a much, much, uh, more difficult problem. Um, that's actually why I think, you know, large planet scale simulation is so important to getting, uh, getting the future of military planning. Right. Because if you can see the, if you can see the outcome, it's a lot easier to judge it against like a whole host of metrics instead of a single one.

Brenna Kuiper: Yeah, that makes perfect sense. What was the most humbling moment where the market surprised you while you were building? One brief.

Grant Demaree: We really struggled with top down sales. So we have not really struggled to sell directly to individual military commands. Um, and I think it kind of surprised people that we were able to walk in and say, here's the demonstration, here's the results. Uh, we've gotten in other places. Um, you can put your hands on it, try it. And people were like, oh, okay, cool, I'll give you my O M funds directly for this because I think it's going to make me a lot better. Um, and that just didn't show up at large top down sales. So when you go to the big daw, um, wide organizations that you would think would give you like big buckets of money, we have a 0%, had a 0% success rate there. Um, this is the first year we've ever gotten into the palm, the program objective memorandum, which is basically what feeds the president's budget. And we, relative to other defense startups at our scale and valuation, we were the slowest to do that by far.

Brenna Kuiper: Interesting. Switching gears slightly. Um, I want to talk about co founders. You operate in a high stakes, high pressure industry. What does a great co founder relationship look like?

Grant Demaree: We're kind of playing an easy mode here because Rafa, uh, is baseline smart and did a huge amount of work and it's like, it's a lot easier to work very well with someone who's like really smart and is contributing just, just epic amounts. But I, I think there was another thing that was very helpful to us early. Um, so like I was coming from a totally different world, like coming straight out of the military. Um, we didn't even really have a shared language and definitely not like a shared set of well, what do you do in this circumstance? What makes a good hire? What should a hiring process look like? What do you expect from a sale? How are we supposed to communicate? What, um, do we and do we not value in the outcomes of a customer pilot? Um, and there's just a million early startup questions that if you have a shared language and a shared set of assumptions, you can work really well together and if you totally don't, um, and it's really hard and like we went through an accelerator and got a lot out of it and every week after the class we would take the, the long way back to our apartment, like a good two plus mile walk and we would discuss the lesson in great, great detail. And I think like really leaning in on those wasn't um, a we will do the book move all the time. We, we actually deviated from the book move quite a lot. Um, but even having a shared default means it's a whole lot easier to communicate and a whole lot easier to get to a decision. And I think a, uh, very deliberate effort to get to, to get to a shared understanding quickly, uh, around the core questions of a startup was, was pretty helpful to us.

Brenna Kuiper: Makes total sense. And I'm sure you spent a lot of time together while you guys are figuring all of those things out. Next question. Paul Graham's Founder Mode argues that scaling advice can push founders into manager mode too early. How do you think about founder mode and a company serving the military?

Grant Demaree: Okay, so I, I kind of loved the article. Like, I think this is like one of the most, most read uh, articles across Silicon Valley. So um, Paul Graham is, is basically saying that that Founder Mode is taking direct control over some portion of your company, including in a late stage. And it's kind of the opposite of leading through one or more layers of leaders. And there's an obvious sense in which you can take this way too far. Like 400 people in our company, we're in all these time zones, we're in all these areas. We have a huge number of teams doing things right now and if I were to directly lead all 40 of those, that seems a little bit insane. Um, so I, I think there's some extent in which obviously founder mode starts to become dumb when you get really big. And I think it's only a question once you get to some scale because you know, series A, series B, you're still the, the doer in chief. And this isn't really a question. Um, but even though there's an obvious sense in, in which doing a lot of delegation and trust and selecting the best leaders and offering a great deal of autonomy is, is right. I still think Paul has a really good point. Um, where there's a host of properties that a founder is likely to have that no one else does. Um, in one sense they're like, they're super incentivized because most, they have all the stock. They generally have the power to make large, controversial decisions very quickly, especially if you've retained board control, um, and generally have really, really good intuitions on where the market's at and what customers are going to want. And if you combine the motivation with the decision power with the understanding in the market, there's a lot of hard pivots and difficult actions that only an empowered founder can do and that no one else would be able to do. Um, so in that sense, recognizing, uh, when you're in one of those and being like this is the opposite of the classical management lesson. This is a time when management principles will not work because someone who was delegated this shouldn't, uh, be expected to do it. I think you recognize one of those and then go, go do it. And I think our acquisition of Battle Road was, you, uh, know, one example of that. And then the like, hard pivot of resources into, um, you know, what we're internally calling the future of command, which is basically integrating these two products to create superhuman course of action. More on that later. Um, I think both of those were examples of uh, like founder mode things, even if 95% of the company is not that. And like, if you try to founder mode everything, you'll just die.

Brenna Kuiper: Yeah, um, I mean, this is a perfect segue into the next question. Um, you know, as mentioned, you've had a huge start to your year with your Series D and the acquisition of Battle Road Digital. What does the road ahead look like? For one brief what's your next step?

Grant Demaree: We're going after the big moonshot, um, which is truly superhuman course of action that no staff could come up with unaided. And like, I've had this theory for a really long time that you had to merge the software for operational military planning with the uh, war gaming and simulations to be able to do this. And the idea behind it is, let's say I chain a bunch of Claude or ChatGPT prompts together and do it really, really well to try to produce a plan for the defense of Taiwan. How good is it going to be? And I think the answer is today it's not going to be that bad. Like, I, I think it's going to be Better than the 50th percentile staff college student. Um, and it's also not going to be that great because there's going to be a lot of like, subtle little things around timing and sink matrix and how many munitions need to be here and like little difficulties that when you get a plan that's literally a thousand pages of detail plus a lot of that aren't written in the plan. Um, and you lack an underlying world model. You end up with some really, really big problems. And even if you didn't have them, there'd be no way to tell without simulating out the planet. And you escape that problem when you bring in a really large simulation that is planet scale physics based. You're literally showing every munition trajectory, every radar cross section and detection probability. So that you take your plan and you play it out totally with AI, um, accurately representing what each blue commander and each red commander would do. And you're able to say here's how my course of action did in this simulated war and then automatically improve the co op. Um, and we've actually been able to do that in prototypes so far. Like that's, that's literally the reason we bought Battle Road. That's been the, the dream of the company the whole time. And we, we just got that done. Um, I, I, I don't want to, I don't want to overplay that where um, we're very, very proud of getting all the pieces working. Being able to generate the COA with a number of agent chains automatically spawn an accurate war game and have blue commanders make reasonable choices and play against dread, automatically generate an after action review, improve the course of action and then keep doing that in a loop until it's better and better. Um, and the, the game now is how do you get each of those uh, command decisions to enough accuracy that you truly trust the simulation. Um, and in a matter of months I want to be at a level of accuracy to go to a combatant commander and say here's the course of action that we, we came to. You'll see it and in your intuition say I'm pretty sure this is going to be better. And then dive in to the evals around, you know, each uh, simulated decision that went into this and say yep, I checked this one and I actually do buy that's how a prc, uh, commander or an Iranian commander would have, would have behaved and ultimately trust the outcome. Um, so I, I want to see that in live operational use on super and J Wix inside the current government fiscal year.

Brenna Kuiper: Incredible. I almost wish I was still in maybe just a little bit so that I could witness that in person.

Grant Demaree: I mean I sometimes, I mean I wish I was still in like not, not too often. This is super cool and I like it but like every time I'm like,

Brenna Kuiper: but uh, right now it would be pretty amazing to see this all in action. Um, so obviously we've talked a lot about Battle Road Digital. Is there something, is there a major change or a way that we should be thinking about what one brief is becoming and how it's changing, um, towards the future given um, the simulation and live operational design decision support that you just mentioned.

Grant Demaree: Wargaming simulations are this really really hard problem. Like I think people have been working to get to really accurate simulations that commanders trust and rely on. They've been working on this for a really long time. Um, like uh, there was a pretty large company uh, that burned about $200 million to deliver nothing of great use. And like we looked at five different uh, wargaming and SIM providers to decide on Battle Road and having one that actually works is an absolute game changer. Um, there were a couple of things that got us there on this one. And like after like multiple years of like looking for this, like you get really excited like ah, you guys did it. I know you're early, I know that you don't have literally everything, but I can see the ingredients. There's.

Narrator: Yeah.

Grant Demaree: And uh, seeing a whole world physics based sim because I don't think, I don't think you can get to where you need to be if you're uh, hand waving, uh, or if you haven't built the foundation on which you can build almost any military or non military capability on top of. And I think the availability of that just epically changes what a staff can and cannot do in its decisions.

Brenna Kuiper: Amazing. With one brief now operating at real scale, how are you thinking about where to deploy capital next?

Grant Demaree: I guess the first answer is always that we gotta keep the main thing. The main thing like reaching this future of large scale command decisions via uh, integrating one brief with this sim to have the full loop from agenta course of action to AI playing AI in the war game to using that to get an after action review and make the course of action better and keep going. That's the main thing. And um, I love a side quest. Um, but we really gotta win on that. And I also think there's, I think there's not quite enough M and A going on in defense software. Um, and there's, there's some areas where, where we can do a really good job. Um, I think getting, getting uh, end to end is really important across a number of staff sections, some of which we just don't serve that well or at all right now. And I think there's some products that serve some of the staff sections we're not in that really benefit from being fully integrated into a Like your overall planning and staff collaboration tool with your wargaming and sim. Putting them all together into a single picture where you spend time on the one screen can be really, really valuable. And I think what we've shown out with the previous acquisition uh, was that we can help these scale a lot faster than they would already. Um, partly via uh, our giant infrastructure and security team to get them onto more networks faster, partly via a uh, presence in a number of customers. So you can say hey you're in this one. But we know based on looking at a number capabilities, this type that this is the one that's going to win. And you know, we want to field it to each uh, of our customers very quickly and integrate it into our product offerings. And what that means is that when it's truly strategic and when the numbers pencil out, it's a priority to me and uh, I would absolutely deploy additional capital into uh, into the right acquisitions.

Brenna Kuiper: Still talking about capital, there is a lot of money pouring into defense related tech right now. Are we in a defense bubble?

Grant Demaree: I guess we start with like what, what a bubble means. So like a uh, bubble means that the future discounted cash flows across the whole industry just don't add up to the present valuations. Um and uh, I actually think uh, that on net across the whole industry they do add up. Um, so, so let me start with the bear case and then why I, I don't really buy the bear case. Um, so, so the bear case would say that um, these defense tech companies are very unlikely to be wildly profitable because the current market leaders are not wildly profitable. Where you have these five really large primes, they do mainly cost plus contracts. Ace rtx, kind of the exception that has done a pretty good job at a fixed price but they're not, they're not crazy profitable. And I think a lot of experienced leaders in defense say that as you reach scale the government will have really huge objections to you uh, ending up with these massive cash flows that are necessary to justify the present valuations. Um, and I guess the question is, well will they? Um, I think they might be okay with it. Um, I think they might be okay with it because we, we have at least one existence proof which is uh, SpaceX. Uh, SpaceX is the most valuable defense tech company. Um and you know by building something that is just extraordinarily hard to replicate, um, they've been able to, to, to work at a, at a price point with the government. Um, that, that does work. Um, I also think that the uh, you know, expectation of Firm fixed price is probably like a historically contingent, unnatural thing. And I think that the investors that are betting on the proportion of contracts that are firm f continuing to go down, I, I think they're right. Um, so on the, on the question of like when you reach scale, can you sell commercial products at a premium? That is basically the government's trade for you taking a bunch of upfront risk to build something that may have lost huge amount of amount of money. I think that's a good trade that the government's willing to make and that there's some evidence that they've already made it. And I, I think it's in uh, you know, every American's interest to keep making that trade because the thing that incentivizes people to build things that are non consensus and better in some kind of unpredictable way. Um, and I think that portion of the, hey, it's not really a bubble if you buy firm fixed price. I think that's pretty widely debated already. I think there's a second portion of the around like how big does defense spending get? Um, that I haven't really seen in a lot of large funds model. Um, and what I love about you know, late stage is that the investors will like literally give you their, their work and you're like, okay, here, here's, here's, you did a DCF on me out a really large number of years. Do I agree or disagree? Um, and I think one of the things that's usually missing is there's some probability of real great power war or um, at least higher intensity competition than there is now. Um, and my best bet is, is about a 12% per year risk of that. Um, and that sounds like it shouldn't massively change the numbers. And I actually think it does massively change the numbers. Um, and the reason I think it changes so much is that it changes how much the government's willing to pay for combat power. And if right now you're willing to pay only about 10 billion per year for like the marginal 5% of US combat power, sounds really low, but that's actually what I think it is. Um, and your willingness to pay went up a hundredfold and there's only a couple areas where you can actually buy a bunch of marginal combat power very quickly. The areas where you can, which is munitions autonomy and software, especially AI software, you end up um, wanting to buy a huge amount of that very, very quickly. And you end up with both the revenue going way up and the cash flows going up by even more. Because it was both More revenue and more high margin revenue. And my best estimate is that the possibility of great power war, even though it's only in my estimation about 12% per year, actually multiplies, uh, the like, uh, expected value of a defense tech company in a, uh, high priority area by about 3x. So it's basically a huge increase. So on the like, hey, are we in a bubble or not? Um, it kind of depends on what kind of fund you are. Like if, if I'm a mega fund, fully diversified, um, I don't think defense tech is a bubble. I'm going to allocate 5 or 10% of my fund to that. These are massively correlated bets. But it's kind of okay because it's less correlated with AI than everything else. I would absolutely allocate there. And I think the expected value is really good. Alternatively, if I'm like a 50 to 100 million dollar defense tech specialized fund, I actually think that doesn't pencil out, um, where I think you put this massively correlated bet where most of the time it actually go badly and there's a small number of times where it goes very well. Um, and I think it's a really rough time to be a defense tech specialized small fund unless you have some special advantage. I know it was a super long answer, kind of. Any questions?

Brenna Kuiper: Well, I'll follow up to that. Where do you think the hype is most disconnected from what the military actually needs?

Grant Demaree: This is one of my less consensus takes. Um, and I love those. I, okay, I love submarines. Dad, uh, was on submarines, Little brother's on submarines now. And I understand the argument for building a whole bunch more. And I'm using this as a representation of all the things that take a really long time to build, of which, you know, new, uh, SSPNS especially is sort of the, the best test case because the argument in favor of them is so strong. Like, yeah, like our old ones are getting very old. It's a very difficult situation. And I think it's pretty easy to say we want a whole bunch more of those and we should put huge amounts of money into that. And I only kind of buy it. Um, and the reason I only kind of buy it is that I'm not sure how much like short range additional combat power, meaning next five years the marginal dollar there actually produces. Um, and it basically comes to a discount rate question where like what is the discount rate per year of the marginal combat power? Where I think if you were a Cold War planner, you might say maybe 5%, maybe less. Where like, or actually let me phrase that better. So how much better is it to have combat power right now than a year from now? And how much better is it to have combat power a year from now than two years from now? And if you're expecting like really really fast technological growth and we're expecting AI to be genuinely transformative and uh, um, already investments made five years from now to be massively more efficient, like maybe 10 to 100 times more efficient than today, then then you're probably also expecting things to obsolete relatively quickly and your discount rate should be really really high. And I think that most people, when they stop to think about like what is the correct discount rate will come to something that's very, very high relative to the Cold War. And if you don't stop and ask the question, you're probably going to stick to 5% or less because that's what our system is. Bu, um, and I think that if you are playing on a 20% uh, per year discount rate of the value of the marginal combat power, then military uh, investments that pan out in 10 years start to look really bad. And I think that's the world we're actually in. So I would look at every military investment um, around ships where the marginal dollar doesn't deliver a bunch of ships very quickly and I would say that doesn't look great. Or you know, future aircraft where the combat power materializes in some number of years. That's a lot of years. And I would say that if you expect AI to be pretty good, um, if you expect the pace of technological change to be very high, then that's a really bad bet and we should get like really rigorous around how much combat power is this? On what time horizon? And actually have to state the discount rate because I think almost all like military budget debates around between smart informed people are like uh, are on top of this like hidden variable which is what is your discount rate?

Brenna Kuiper: Mhm. Are you going to get in some trouble with your family for that answer?

Grant Demaree: Um, they do listen to these like

Brenna Kuiper: maybe, maybe you might get a text in about a month or so or whenever we publish.

Narrator: Are you still managing expenses with outdated apps and spreadsheets? If so, it's time for Ramp, the leading US spend management platform covering corporate cards, travel, AP and more. All built to save time and extend Runway. We use Ramp at TFR and it's been a game changer. As our preferred partner they're offering listeners $150 just to take a demo. No need to sign up. Head to ramp.com partners TFR to claim your $150 before the offer expires. Now back to the interview.

Brenna Kuiper: Um, switching gears after the anthropic controversy around defense work, why should the best engineers in the world choose to work on military tech?

Grant Demaree: I'm a really big believer in defense tech and in America in general. Um, and I want to make the case that if you personally are likely to be skilled in defense software, it might be the most leveraged use of your talents literally in the entire world. Let me start with, uh, American hegemony is really, really good. Like crazy good. Like since World War II, uh, American hegemony is why we have so few deaths from violent conflict. People, uh, living longer and freer and wealthier by just insane margins. And given a choice between living in the post World War II era and any other time in human history, it's, it's just not hard. It's so much better. And a lot of that is because the, the US has been the most benevolent hegemon in, uh, in all of history. Um, and that doesn't, like, that doesn't logically prove that additional US military power right now is good simply because US military power over the last 80 years has been good. Um, but I actually think it still pencils out even now. Um, one of the reasons being that American hegemony makes war, large scale war, a lot, a lot less likely. Um, and I think we've seen that when the US has lots and lots of military power, we're able to secure sea lanes, we're able to deter large conflicts. And when people doubt American military power, then these conflicts are less likely. Um, and the next objection that someone might say, oh, is well, no, no, no, like American power is declining. It's, it's too late. Like, uh, you're not going to get to a level of hegemony and adding a whole bunch of military power is not actually going to reduce the deaths from violent conflict or create a stable world in which, uh, you know, the development of AI unfolds in the most, uh, fav. Favorable geopolitical situation possible. Um, but I, I actually think you can get to, and, and in fact will get to really, really dominant combat power on the US side. And I, I think that's where my view most differs from, you know, what I hear, whether in, in Silicon Valley or, or New York. And I think, um, that the, hey, no, no, it's too late for a US Combat power argument is totally wrong. And that in fact we could build huge amounts very, very quickly in a way that no one else can match. Um, and my theory here is that there's only three ways to make a lot of combat power really really quickly. Munitions autonomy, meaning like UAVs, UGVs, um, autonomous undersea vehicles and so on. And then AI software. And of the three, um, AI software gets you the most combat power the fastest. In part because it scales really nicely. Like the cost of duplicating software is pretty close to zero. Um, and in part because um, it affects like basically every area of military decisions if done right. And my best estimate is that literally perfect military decisions, meaning you take exactly the action that gets you. Uh, the closest to whatever your terminal set of values was is we're basically reaching perfection on this is about an 8x growth in combat power, um, which is way more than you need to get to a, uh, a balance of power that is many, many, many times more favorable than even the most favorable in history, like immediately post Cold War. Um, so my basic thinking here is that US has a pretty good lead in uh, AI software in the military. Um, making that bigger, uh, creates a balance of power that is massively favorable to the world. And if you have a lot of very talented people doing it really, really well, you're likely to create an environment where everything else that's important going on in the world, such as these AI models getting better is likely to go really well. Therefore, if you can do a good job in defense software, uh, you really, really should.

Brenna Kuiper: I think you made a very good case for it. What does a military staff look like in five or ten years from now if one brief continue succeed?

Grant Demaree: I mean, I guess first of all five to 10 years is a really long time horizon. Like I think even imagining two years from now is so different that it would be like someone in ancient Egypt looking at Napoleon's staff and saying wow, this is just, this is so much better. Um, I, I think you start with extremely automated. Um, like there's a huge number of military decisions like break take something uh, fairly simple like the sync matrix, um, for non military listeners. Sync matrix is just a militarized gantt chart, like a giant chart of when everything's happening, like what depends on what. And it can be like thousands of entities and it's so human done that a lot of large headquarters will have like multiple full time sync matrix managers. Um, hey, and that's the sort of thing that like a human deciding that is insane. And I think in less than two years you end up with virtually every major military decision heavily automated. Um, and that's everything from the stuff that I think other companies are doing really well, like the kill chain where you really do integrate all the data from every sensor and the right shooter fires immediately, up to logistics where stuff moves at the right time to the right place in an anticipatory manner. Um, up to the planning where for whatever set of policy goals or terminal goals you might have, you find exactly the right campaign, uh, and exactly the, the right set of operations to, to make that happen.

Brenna Kuiper: Shout out to Pittsburgh as well. So that's where I'm located now. Um, the armies um, AI center is located here. So I know that's been a big push in obviously in the broader military, but in the army especially, um, with the new functional areas and I've run into quite a few military officers here who are still active duty and they're at the AI center and so it's incredible to see the big push that the military is making for that.

Grant Demaree: Yeah, I'm really, really encouraged about that.

Brenna Kuiper: Yeah. So to circle back on something you mentioned earlier. How do we solve the West Point military veteran founder problem? How do we get more grant demaries, uh, coming out of active duty, founding these companies, how do we solve that?

Grant Demaree: So we've got a couple already. Like my cto, Corey Andrechka is actually a Naval Academy graduate. Um, he's so brilliant, I love him. Um, and uh, he's had quite the career. Um, so I think uh, I think it's pretty clear that this does happen and as defense tech becomes more important and becomes more legitimate, I think inevitably more people are going to try to do it. Um, but I also wonder, the idea of build a startup as the thing that, that you should go do, um, has been around for a while now and there's not a crazy number of ah, academy graduates who are, are going to do it and building these like massive unicorns. Um, and I've got a couple of theories on why. Like you can explain some of it just with will people leave with the service obligation, um, which means that they're not spending their early 20s in technical roles or they're not spending their early 20s getting exposed to startups. And there's this window of time when you could have founded a company and, and you don't. I don't think that totally explains it, but I think that gets you partly there.

Brenna Kuiper: Mhm.

Grant Demaree: I think the language barrier also doesn't totally explain it, but it gets you partly there where if you went to Stanford or mit, you, you know the right things to say to signal like you know, what you need to say to an early stage investor that will make them think you get it. And if you don't even know what ARR is, let alone like how much you need to, to make a certain signal like uh, or and certainly not like the really subtle things that someone might look at to know if this is a high potential founder or a low potential founder, then I think you're going to get underestimated. And I think if you're coming out of the military transition, I kind of reject the advice of you have to do it at some other startup first. Obviously uh, I didn't do that and uh, eventually got to all the lessons, um, some of the lessons, um, but I do think you've got to at least save a bit more money to be able to survive for long enough to like, like find those. But I still don't think that totally explains it either. Um, like I think there's a, a fundamental um, personality difference there where the only path to victory in early stage startups is to be non consensus and right where like if you're non consensus and wrong like you're just doing something silly, it's going to flop. Um, and if you're consensus and right then like why you um, where if it's already the consensus then someone with greater resources, um, contrary to popular interpretation, large scale companies are actually very efficient. Um, why would they not do it? So you've got to be breaking the consensus and you've got to be right at something. And I think that in the military most of success comes from coordination games like in the game theory sense where you need everybody to be doing the same thing. Let's say you uh, and I, Brenner are going to do a react to contact uh, like at West Point and like oh we got this enemy ahead of us. Okay, everybody flank left. Team leader says flank left and you're like no, no no, I want to flank right. Um, that's, that's not going to be a good outcome. And I think that most of the military is that on a larger scale like I think 90% of the separated offenses power actually just comes from being a source of coordination. Like literally his words are a source of coordination whether they, they uh, even, even in situations where you're not making an order with legal authority. Um, and that means that uh, I think you're trained to not want to do things that are incredibly non consensus. And I also think there's a lot of selection where like doing, doing well in the military selects for someone who can play coordination games successfully, um, which might mean that you're somewhat lower in like, openness and like, as a big five personality trait, uh, than the most open Stanford MIT founder. And I don't think there's a solution, ah, to that because it's actually like a really, really desirable trait in the military. Um, but if I was trying to get, uh, founders coming out of, ah, West Point and the Naval Academy and the Air Force Academy to be really successful, um, I'd say save a whole bunch of money so you have time to figure things out, dive in really, really fast so you learn the language and you learn how to signal. And if you feel like if you know in your bones that doing something non consensus where like someone can look at this and say, this is what you're doing is not prestigious, it's not on the correct track. Like, can't you see what uh, all these other people are doing that is on the correct track? If you know in your bones that that's not what you want to live for a bunch of years, uh, then I wouldn't go into a business where that's literally what leads to victory.

Brenna Kuiper: It's great advice. Hopefully we get some listeners who are making their transition off active duty and starting to think about their future and figure out if that's going to be the right fit for them, if that's something they're interested in doing. There's no better time. Right?

Grant Demaree: Uh, I do agree with that.

Brenna Kuiper: If we could feature anyone on the show, who should we interview and what topic would you like to hear them speak about?

Grant Demaree: I think CENTCOM has made some really impressive strides in defense software and their use of AI in a way that's not really replicated anywhere else in the world. And their chief data officer is this gentleman, Cyrus Jabbari. Um, it's pretty busy right now, um, as you might imagine, given what's uh, going on in the world. But I think a firsthand look at what the frontier is at, what a combatant command can do with software and AI. Um, I think the listeners would really appreciate that.

Brenna Kuiper: Awesome. What book, article or video would you recommend to listeners?

Grant Demaree: There's a sequel to Ender's game, um, called Xenocide, and it's not really as popular as the original. It's like a little slower, a little more philosophical. And it's the best book on like, on how to think clearly about a question you're trying to investigate. Um, and it was a really foundational philosophical book to me. I read it like 20 years ago. Um, I reread it a couple times. Um, but it helped me a lot in defense tech.

Brenna Kuiper: Awesome. Do you have any habits or behaviors that are a secret weapon?

Grant Demaree: Gotta get your exercise. If you exercise and eat healthy, you'll be able to sleep. And if you sleep, you'll be able to show up for everybody.

Brenna Kuiper: And what is the best way for listeners to connect with you?

Grant Demaree: I mean, my email is pretty easy to find at this point, so, uh, sharing it publicly doesn't really hurt. Um, I'm, um, um, I'm grantonebrief. Com.

Brenna Kuiper: Awesome. Grant, thank you so much again for joining us today.

Grant Demaree: Hey, thanks, Brennan. It's great to see you again.

Narrator: All right, that'll wrap up today's interview. If you enjoyed the episode or a previous one, let the guest know about it. Share your thoughts on social or shoot them an email. Let them know what particularly resonated with you. I can't tell you how much I appreciate that some of the smartest folks in venture are willing to take the time and share their insights with us. If you feel the same, a compliment goes a long way. Okay, that's a wrap for today. Until next time, remember to over prepare, choose carefully and invest confidently. Thanks so much for listening.

Grant Demaree: It.