In episode one of Building the Base, hosts Lauren Bedula and Hondo Geurts chat with Balan Ayyar, retired US Air Force General Officer and now CEO of Percipient AI, a Silicon Valley-based artificial intelligence and machine learning firm. Balan’s unique background, spanning military service, government IT, and startup leadership, provides valuable insights into the challenges and opportunities at the intersection of national security and technology innovation.
Topics discussed in this podcast include:
- Balan Ayyar’s Personal Journey: Balan shares his immigrant background, family inspiration, and how he transitioned from a distinguished military career to becoming a tech entrepreneur.
- Fundraising Challenges: We hear the early challenges of securing investment in the national security and intelligence technology sector and the skepticism faced by startups like Percipient AI.
- The Fast-Moving Nature of Technology: Balan highlights the rapid evolution of AI and machine learning technologies, making government requirements quickly outdated and challenging for primes to keep up.
- Incentivizing Primes and Government Agencies: The conversation emphasizes the need to align incentives for defense primes to adopt commercial technologies and for government agencies to prioritize solutions that improve user experiences and operational effectiveness.
- Mission-Tuned Software: The importance of customizing or “mission tuning” software to meet the specific needs of national security and intelligence organizations is discussed, emphasizing the need for dedicated efforts from technology companies in this space.
Katherine Boyle is a General Partner at Andreessen Horowitz where she invests in companies that promote American dynamism, including national security, aerospace and defense, public safety, housing, education, and industrials.
She was previously a partner at General Catalyst. Prior to GC, she was a general assignment reporter at The Washington Post. Katherine holds a BA in Government from Georgetown University, an MBA from Stanford, and a Masters of Public Advocacy from the National University of Ireland, Galway
BENS Intro 00:02
Business executives for national security welcomes you to building the base here thought leaders and practitioners discuss how we can ensure our shared security and prosperity. They’re shaping the future of the national security industrial base. Your hosts or Silicon Valley defense expert Lauren Bedula, along with Ben’s Distinguished Fellow and former head of acquisition for the Navy Marines and special operators, Hondo Geurts.
Lauren Bedula 00:26
Welcome back to building the base, Lauren Bedula and Hondo Geurts. and we are so excited to have Balan Ayyer with us today, a seasoned CEO and retired US Air Force General Officer, who has seen a lot of the issues we talked about on our show, both from the military perspective and as the founder of a startup. And prior to founding Percipient AI, which is a Silicon Valley based artificial intelligence machine learning and computer vision firm based on the West Coast. Balan was the CEO of a multimillion dollar government IT firm so can talk to both that larger company perspective and startup perspective to some of these issues. So Balan, thank you so much for joining us. We’re excited to dig in.
Balan Ayyar 01:12
Oh, well, it’s a joy to see you both congrats on this incredible podcast, I think you guys should do exactly the right thing by bringing luminaries like Mike Brown and others that you’ve interviewed on to share their thoughts. And hopefully I can advance the conversation with whatever insights I can offer you on this important set of topics in such an important time to be discussing it with incredible innovation happening in the commercial economy and the speed of change of these technologies couldn’t be a more important time.
Hondo Geurts 01:39
So Balan one of the things we really love here is the unique stories of all of our guests. And you know, many of them intersect multiple communities. You’re unique amongst unique people as a General officer, as an operator as a working in policy in the Sec Defs office, now multiple times CEO so give us a little sense of your background and how you got to being on this podcast from your early days.
Balan Ayyar 02:09
Yeah, thanks, Hondo. First, thank you for your leadership and service I’ve watched you just do extraordinary things for the nation, both in the Air Force, of course, with your incredible leadership, but also it’s SOCOMM, where it was such an incredible time to see the way that you thought about commercial technologies kind of live the all the challenges that I’m sure you’ve talked to your guests about. So congrats to you and Lauren, listen to Ayyars have had just an amazing American life. I mean, I’m, I’m totally blessed. My parents immigrated to this country. My dad came in 1954. You know, at that time, only 100 Indian families a year could come to America until 1965 When they reorganized immigration and started a merit based system. So a very small number of families that came that early from South Asia, and we were watching the European float make sense, just based on our relationships and nations and how nations grow. So my dad, my dad, you know, he’s left, he’s left us now but he just bled red, white and blue. I mean, this just incredible young engineer from Kerala, he just lived to write 200 companies to respond to graduating and 52 with his engineering degree to respond. And in America wanted to come to America so badly, took a young engineering position and, you know, culminated 40 years in Boeing as a technical fellow senior Senior Principal Scientist and technical fellow there but but working on military programs, which is you recall Boeing great company, of course, still great American company. But in those days, great days for the Air Force, you know, hundreds of 1000s of people working across the incredible variety of weapons systems mind that. So a great inspiration for me when I was young to see my dad’s hard work ethic, and then my mom, you know, just everything you can imagine in a mom caring for others. Really CPA, gifted education teacher, but her primary purpose was serving less fortunate, which set a good example for all the little, little Indians as I say, and when I graduated high school, my mom said, Hey, I’d like for you to serve in the Air Force by her brothers had served in the Indian Air Force, and she felt so happy about what America had given us. And so I didn’t really get a choice to go to college. Unfortunately, I missed those four years that, you know, you and Lauren enjoyed. No, I’m just teasing. But then I had just wonderful 25 years, just like you Hondo amazing opportunities. My beautiful bride I met in college and she’s at University Wisconsin, but I had a good friend at the Air Force Academy. So we started courting my senior year and we’ve been married, you know, for whatever it is now. 3435 years and beautiful children. We have 25 years of service, wonderful experiences like you, like you described flew jets all over the world, that these just incredible opportunities from the Air Force to strengthen our character and our leadership and to take take seriously the responsibility culminated with commanding forces in combat in Afghanistan as a commanding general there in 2013 2012 2013. Very difficult time. But really just completely, everything you could hope for as a young person to have responsibility for men and women who are in uniform. In fact, I had begun, I’ve never had a brigade of soldiers under my command before, although as you know, hundo, as you grow in leadership, you get greater and greater levels of responsibility. And as a young officer, General, almost 14 years ago, I took command of the recruiting services and brand new Brigadier General 2010. I had a great two years of wonderful experience there and then over to Afghanistan, and never had a brigade of soldiers before but I’ll tell you a hundo, I just fell in love with them. I fell in love with the Joint Force, and fell in love with these young men and women who are serving in faraway places in a way that I had never really appreciated flying jets. And I think you’ve probably experienced it in the Navy too. These young sailors are so impressive. You know, at their heart, they care deeply about each other. And they’re just such great American citizens. And so, as a firm, we still support the US Army because Mark Milley was there as my fellow commander and became the chief. And it was just a complete capstone experience for me. And then I came out of the service in 2013. Take care of my parents and my children who were growing up largely in three years as a general officer without their dad. So I decided I needed to be more relevant to my family and have been in the business for the last 10 years. The first few years, as Lauren mentioned, I led a great small IT service that grew rapidly, I got a lot of credit, it wasn’t really me, I just had put the right leaders in the right positions, and the company grew. So all of a sudden, I looked really good on paper, because the company grew so fast, but it was really just leadership, anybody, anybody who saw that environment could have made those decisions. But then I went out and did what I wanted to do, which was pitch in Silicon Valley to build this company. And I thought, as you and Lauren had pointed out that it was so incredibly important for us to have the best technologies, you know, we had a number of attacks in Afghanistan, I thought might have been prevented, or that we could have protected against or potentially predicted if we had some kind of human mission team going on human machine going on, you know, Hondo, because your team was supporting, I was partnered with, obviously, special operations forces there for my test course, we had Air and Space supremacy, if I can use those terms. Nobody was threatening our forces in the air or in space. And so as a result almost every big time, but certainly every brigade had their own overhead. And we had so much data flowing Hondo and Lauren, you just couldn’t believe it. But we still couldn’t process it and analyze it quickly enough, in order for us to take action when an attack was unfolding or to prevent an attack. And so I had a couple of seminal, you know, attacks there while I was in command, and I came out of there praying that someday we’d have the compute, and the algorithms and the AI processing capability to help our viewers or brilliant analysts and operators accelerate their understanding of the environment so they could protect lives in any environment. And I’m so pleased to tell you as I sit here, it’s 2023 10 years later, we have all of the capabilities I dreamed about in our software. And it took us four years of coding 100 million in venture capital, and incredible customers in the intelligence community that know how to take risks. And now we have capabilities in Visual Intelligence and correlating ground air and space based visual sensors that I dreamed about took us a dozen years. It didn’t take us 10 years to get the computer and the AI aligned with the missions and the brilliant scientists to develop mission tune software for these missions. But we’re there now. So anyway, that’s a brief summary. But it’s wonderful to be with you guys.
Lauren Bedula 08:30
What a story. I love it. And I always want to start by asking what drove you to do business with the defense and national security community? But I think you so clearly answered that with your story. So I’m going to tweak it a bit to ask about the experience as a founder fundraising, when you started with it, 2016 17 having those conversations a little bit on the earlier side? What was it like? Were you met with skepticism or?
Balan Ayyar 08:57
Yeah, well, I don’t want to listen. First of all, let me just give a shout out to my investors. Venrock Rockefeller, unbelievable. Nick Biem, our series, A investor understood AI was unfolding and could have a dramatic effect in security intelligence, believed in the team, the incredible technical team that we have. They’re led by one of my co-founders, Raj Shah led the delivery of Google Maps and, and then Vasu Parameswaran, Jérôme Berclaz these are the top scientists in vision in the United States. You know, lead Microsoft’s maps and Nokia and Siemens and other great companies’ efforts, including Uber in autonomy and autonomous vehicles. Really, the constellation of brilliant scientists envisioned and Nick Beim believed in us, James Murdoch and Ben Fitzgerald from lupa systems and Lee Fixel From Addition. These are investors that have really had a long view. I wish everybody had the opportunity to pitch in front of them with the critical kinds of work that we’re doing to support net security and intelligence but let me tell you in 2016 Even In-Q-tel was skeptical. What to see In-Q-tel, beginning of 2017. And I brought a couple of the scientists that we’re going to hire. And I said, we’re going to fix this problem, person recognition, correlation vehicles, all the things that I thought we were attacked by and facing attacks from, I was so happy that you tell took the meetings and said, Hey, we’re very interested at the end of those meetings, both of these partners that I think you tell network and point 72 another investment firm, but they, they’re like, you know, we really love what you’re trying to do. But this just doesn’t fit our model. You know, we’d like to find commercial companies. And so we’re not really interested in companies that are building a venture backed story around solving for National Intelligence. And I was puzzled, I thought, hey, you guys are making a mistake. In my experience. On the other side, if you don’t really focus hard on those use cases, you won’t get adoption, you won’t get these yawns, Naval Special Warfare, if I can go back to Hondo’s background, you won’t get these young, Special Warfare operators to adopt unless you have a great user experience. And they understand and can trust what the software is telling them to help them make decisions. So it was tough, Lauren, but I found a constellation of wonderful investors that believed in us, Michael Marques and others from W RV, that our seed round, and then we are off to the races with demos. And interestingly, you know, years later, Cattell came back to us and said, Can we be part of the toy company? We are at that time already ready, you know, ready to go to scale the Intelligence Committee. So we really, really didn’t need and could tell. But, but it was, it was hard to learn. I have to tell you, I don’t want to say it was harder when I was young, it’s still difficult to pitch if you’re focused on that security intelligence. And the reason is, it sounds very good. People are really altruistic. But these investors, many of them are on a timetable for their investments. And we still haven’t synced up the ability to deliver great outcomes from nest security and intelligence focused companies. Frankly, we need a couple of companies, including mine, that have spectacular exits. Forgive me for saying that’s the motivation for some of these venture capital firms. But, you know, that’s the mythology they work on, does that make sense? Learn when they see companies that go to scale, with great products, they’re motivated to find other investments in that arena that could potentially deliver that outcome. But when you have a company that slogs its way through an enterprise, ATO, my goodness, two years before, you can be used in an operational sense, in our case, 18 months, but, you know, goodness gracious, that’s a big heavy lift for venture capital firm, they take the pitch, and they go, that’s really great. We really love that we love serving Americans. And then quietly in the background, they’re like, these idiots don’t know what they’re doing. They’re gonna burn through a lot of cash and not deliver any outcomes. And so a lot of venture firms are still very hesitant about companies. But we’re gonna break that we’re gonna have a few great stories, I think our firm will be among them. But it was tough, it was a tough go in 2016.
Hondo Geurts 12:48
So we’ve, you know, you’ve been at this for 10 years. So you’ve kind of seen at least like one generation cycle of this dual use or national security venture cycle. And I laugh when I hear, you know, me being one of them used to talk to Congress all the time and think that was stuff. They haven’t gotten into briefing Investment Groups and trying to get somebody to give them money. So I’m sure that you know, I’m sure it’s a great experience. And so I do think things are trending in a positive way. Can you kind of give your perspective on dealing with the government as a customer, maybe over these 10 years? You know, what’s gotten better? And you mentioned, you know, what are some of the things you know, we can do in the next year or two to really get this moment, you know, capture the momentum?
Balan Ayyar 13:33
Well, I’m so grateful Hondo First, there’s a lot that’s been done because of your leadership and others, really great policy recommendations and boards and commissions. I’m all suggested to investors for you, I’d be looking at firms that are working on dual use technologies, because it’s epic, the impact these technologies can have in when values or intention, life and death missions, our ability, in our case, if I can use Mirage, our software and percipient AI as an example, the implicate implications of having the best technologies at scale, in an escrow to intelligence environment are literally life and death. The decisions that you’re special were for folks in the Navy, in the Air Force, in the army have to make against malign actors, a whole host of them and your peer potential adversaries, our freedom of action hinges or depends on our ability to accelerate. So we think the work, a lot of work has been done on kind of the early side of this national security commission on AI, the CSIS reports, a lot of body of work that says we should be looking for the innovation ecosystem. But if I could to the second part of your question, and so a lot of good work thinking about this, but very little, I think, substantive and meaningful change on where the incentives are. And I just wish, you know, I really do think you’ve got the leadership and reputation and Lauren and your team there. You know, have the influence, the convening power, if I can say to bring leaders together and talk about why we have to get the incentives right, for example, our primes to do wonderful work for our nation they build great as they build great weapons systems, they’re not oriented toward the 10s of millions of dollars in investment it takes to build great software products ahead of the technology’s impact on these missions, you really do need Silicon Valley product companies to build in areas where the technology is moving so fast. Let me just give you one example, Lauren, and Hondo, we’re doing things right now and 2023, with our AI with our software that we didn’t think was possible in 2021. Can you imagine, we’re releasing software right now with zero shot and one shot learning for vision. We thought in 2021 2020 2020 2018, as all the big services requirements suggest, that the most current AI, that we had to build models like foundational models, like you see for LLM in natural language processing, wouldn’t need any of that. Right. And so all of the government requirements are completely outdated. Now, for the big services, prime, prime driven services projects to build vision solutions. They’re completely outdated. This is the problem with fast moving technologies and building Government Solutions. They almost always are yesterday’s technology tomorrow for the Prime’s that are building them that don’t have a background in those disciplines, with fast moving technologies that are highly disciplined. So the risks are even higher against China, who’s operationalizing AI, and Russia who’s a malign actor, as a near peer, but I think dying or decaying power still still dangerous, as we know from the war in Europe, but I would say the incentives aren’t yet right in the right place, Hondo, we’ve got to find a way on multiple levels, we got to incentivize great primes like lighthouse, and booze and others to really adopt commercial capabilities, there’s a great law faster that’s out there, it hasn’t really been enforced, because companies haven’t spent like we have $100 million building capabilities that that otherwise the government would have to build themselves. So they don’t have a lot of commercial options for high end AI tools. But where there are, there ought to be incentives for them to evaluate those tools, as opposed to building hundreds of millions of dollars in government development, which in four years, results in really dated technology hitting our operators. And so I’m I got a knot in my stomach, I feel like it’s almost like a COVID-19 moment where all of us should be talking about what the implications are, if we don’t get this right against China and potential other adversaries, we need the best firms with the best technologists providing commercial options for your forces and Naval Special Warfare and others to evaluate. And I would like you to think not just about the incentives for the primes to adopt, but also the incentives for these organizations to adopt, you know, this, if you’re the director of one of these agencies, and you get awarded a $400 million development contract. To develop a computer vision platform, say instead of looking at Mirage as a platform, commercial off the shelf solution, boy, you don’t want to lose that 400 million, you might get three SI promoted with 200 people in seats, developing that software over the next four years, none of whom will be there when we deliver the software that no one likes, but it’s a seduction. You know, I just jokingly say, Hondo, I’m giving you a hypothetical, which is actually true. These agencies, you know, their incentives are not to turn down that money from Congress, but actually with the prime to develop that software. And you know what the prime Hondo is, it’s beautiful. It’s a seduction, we will build a perfect software for you, you’ll have every requirement you’ve asked for. And of course, in many cases, they do deliver on those requirements. It’s just terrible software to use. None of your children who are native to digital applications, would put that software on their phone. But we ask our young soldiers, sailors, airmen and Marines, because we develop it as a government solution, we force them to adopt it. And then in AI, the key is voluntary adoption, because they’re contributing their knowledge to the enterprise intelligence the software is building. So you don’t want them angry when you’re using the software, you want them really excited about using the software, building trust with the software and contributing their knowledge to the base. So it’s a tough situation. Hondo, I’m sorry to filibuster here. But I don’t think all the incentives are yet right. Our experience, as excited as I am about the intelligence community, and their willingness to take risks, they’re contracting. Some parts of the agency that know how to contract software are very exquisite, other parts that are more operations focus, and don’t know how to really adopt enterprise software, it’s still very much kind of hit and miss whether you you end up in the right lane and have the kind of conviction necessary to sustain a venture backed firm, early on to give a signal that we liked this technology, it’s dual use, and as potential applications at scale. That’s what your investors need. So we’ve had a great relationship. Would I have liked it to move quickly? Yes. I’m thankful to my investors for letting us have the tenacity and technical ability to hire the technical team to deliver these outcomes, but we still haven’t got it. completely right Hondo, I don’t see a lot of firms really focused, there’s a lot of phones that let me just say you’ve talked to Ted Schlein. Ted is a great example. I think Ted would tell you, if I could channel him that it’s not enough just to toss your commercial software over the fence, and hope for the best. With our Special Warfare and the kind of parts of our forces that are willing to take risks, you’ve got to do what I call mission tune software. So you have to have a dedicated effort from your company. If you want to win in this space, you can’t just hope for the best, you got to have a team that understands these missions, you got to have engineers that are willing to tune you can call it customization if you want, but really tune it to make it easy to adopt integrate deal with friction points work the processing. And so you got to be all in. And that means conviction from these customers is very important to sustain that all in mentality. And because the incentives aren’t in the right place, yet Hondo is still hit or miss. Does that make sense?
Lauren Bedula 20:59
Absolutely. And you gave some great advice. I love the point about focusing hard on use cases for achieving adoption. I think that’s an important piece of advice for founders. And I want to dig into your experience with primes a little bit more to our partnerships part of your strategy, or how are you thinking about partnerships for going? Yeah,
Balan Ayyar 21:22
Of course, it’s kind of a, I could have kind of an intellectual discussion about this. And it’s so dreamy, right, I’m the perfect partner for a prime. Because we’re not a services firm. I’ve led services firms, and we grew radically. But the incentives were never there for me to introduce commercial innovation to lessen the number of people I have in seats, the incentives are always there for me to secure my position for the re-compete. So I had this kind of dreamy notion that we’re perfect for a prime, small Silicon Valley firm that doesn’t want to take their business and wants them to present capabilities that’s going to help their customers transform and build trust. But the reality of it is Lauren, I still haven’t found it yet. I think the primes because there’s a lot of discussion of commercial artificial intelligence, and in our firm’s case, are interested in learning about what our capabilities are. But let me just tell you a story that I think is sometimes all too indicative of how primes Look at this. I can’t use these names, but I’ll share them with you. There was a period where we thought our capabilities met a large intelligence agency’s requirements, and we’re very excited. And that intelligence agency had already licensed the version in beta. And it was so exciting. Their analysts love the software, and were thinking that this potentially could offset their desire to build. But here comes this big service opportunity that they’ve been planning coming down the pike. And we told them, Listen, whoever wins this, because there’s a lot of work that has to be done that we can do, please share with them that you’ve got a commercial alternative on the software piece. And we thought at one time, we knew who that Prime was that was going to win this, we went to see them in advance and said, Hey, take a look at our software, we’ll go under NDA, you really don’t have to build this because we spent at that time 50 or 60 million, we’ve got, you know, just the top five scientists have 70. At that time, 68 patents envisioned from geospatial intelligence, you won’t find a better constellation of brilliant coding scientists trying to impress the prime, because they don’t have that kind of people on the bench. They’re all in programs. And we thought we were very impressive. We gave him a demo, they loved it. Can you believe it? Lauren, we found out a few months later that they went to the customer. And, and kind of copied from our website and information, they had given them different portions of the Mirage platform, and overlaid it with what they were going to do. And they told the Prime they told the customer, hey, we’re going to build this capability. And the customer has seen a rush, and hey, that looks a lot like what we’ve seen from Raj, are you working with a percipient? And of course, they were embarrassed because they didn’t know the customer knew us so well. And they had to back off and say, Oh, no, we’re planning to have them on our team, we were just kind of showing you what a combined, they were planning to build, you know, kind of learning from us hoping that they could get the money to build in that service’s prime. So there’s a little bit of that which I call immaturity. Thankfully, I’m confident they could build, they have to have the same banana heads that we have to build the software we built. But there’s still a little bit of the, hey, we’re gonna we’re gonna build this for you in their mindset. Now, lately. So that’s back in 2018. If I fast forward to 2023 because we have an enterprise ATO for our software. And because we are going to be the enterprise solution in the intelligent US intelligence community for a number of missions central to intelligence, all of a sudden we have some street. You know, we have a reputation on the street. And now the key primes that are interested in doing work in this space. When they talk to our customers, they go, yeah, if you’re going to do integrative work, you should talk to the solution provider that’s mirage. And so I feel like we have earned respect after a number of years of slogging it out and going through the mission accreditation process with our customers. Now the prime is probably We are what take us with, you know, with a little bit more to do with us with a little bit more respect. But it is hard Lauren, if you’re a young entrepreneur, you have to be careful about exposing your technology too soon to these Prime’s. You should not underestimate how powerful and deep and competitive these Prime’s are when they want to build capabilities for the nation, because that’s where the incentives are for them. So I wish I could give you a better picture there. But I do think that it’s still a very exciting time. I do think there are some enlightened primes that know they can’t be everything, on all technologies to all customers. And it’s really a smart strategy to partner with best in class technology providers, I think, but I wouldn’t say it’s a cautionary tale. That’s kind of a quote from mean girls. It’s a cautionary tale.
Hondo Geurts 25:47
Yeah, I think that’s a good perspective, both of the challenges and maybe opportunities. You know, I do think if we’re going to compete, if you look at what’s happened in Ukraine, they have built a great network. If you think back to your time in Afghanistan, a great network of teams. You know, you talked about incentives, are there other things that could happen that would allow this network to form? You know, I’ve used the term network versus industrial base? Because I think we need something much more dynamic. What else do you think are other things either the customer could be doing or the industry could be doing to facilitate that?
Balan Ayyar 26:26
Yeah, I think so I’ll give you a couple of ideas. First of all, I think you’re right to say network, it is true, you guys are right to say there’s an industrial base, there’s nothing wrong with that terminology. But the truth is that these kinds of combinations of companies come together around markets, and so you create a new market of different technologies coming together to solve a certain mission or set of missions. Right, we might need the latest technology in hybrid cloud infrastructure to process the latest emission applications which require that infrastructure. So it’s a combination of companies that come together to present a use case solution at scale for the industrial I’m, excuse me, for the, for the, for the services and for the armed forces in for the next security and intelligence space. And we need these kinds of networks to form around great mission use cases. So it’s a wonderful way to think about a 21st century industrial base, as a group of innovative companies coming together, you know, a little bit like a wave rises, different forces in play, you know, there may be one eight, there may be one wave in AI, there may be another wave in Compute, there may be another wave of, unfortunately, instability in the world that forms, those, those forces coming together form a really towering position. If we get the right attitudes, and an alignment, in the effort to solve those problems, we can have really great combinations of companies come together. I think that’s what you’re hinting at when you say network. And the question is, how do we incentivize that I keep going back to that Nobel laureate, 1977 paper from Stanford, which says, you know, a lot of problems you can cite if you go back to where organizations and individuals are incentivized to find the reasons for their actions. And so when we don’t get the outcomes we’re looking for, look to the incentives, and I would say, but separately from, from where I think our agencies are incentivized to build government software solutions, you know, that used to work, in many cases, Hondo, I think we’re you and Lauren are on the right trail here is with the advent of these new technologies, including incredible compute, but also artificial intelligence, the cost of us continuing to pursue exquisite Government Solutions against our near peer adversaries, is so potentially so high, and so devastating to our ability to influence that, we almost have to have kind of a groundbreaking reset on, on really both for the primes. So I would say, what I would recommend a for example, is with the primes, I would say, hey, if you’ve got a $400 million contract, to build computer vision software for an intelligence agency, and you find a commercial software, that you could largely tailor to those requirements, will give you up to 50% of what you’re supposed to earn, without putting people in seats. So the margins for your profit go way up, if you can adopt that commercial software, because we know in time, the government’s gonna save much more than that, by having a commercial software at scale that’s constantly updated with the latest technology. That’s not a legacy that is going to keep the government ahead of the latest threats, as opposed to building a government software that doesn’t have a product owner and doesn’t have all the things in Silicon Valley, we know are necessary to continue to adapt. So I think we could change the relationship of the primes to commercial adoption. That’s one reason that they have a reason besides just doing what’s right for the country, which is nice but doesn’t move the needle on their delivery every quarter of their revenue. You should give them an option if they adopt to have higher margins on their work to gain some of the benefit of the adoption of that commercial. Software. That’s why I think on the other side of that 100 the extent that you meet with lawmakers and you and Lorna Lorna influencing them, if you can find a way for the agencies to, to to be able to keep some of the money that was otherwise obligated to them for the purpose of developing a software, if they adopt a commercial software, some portion of that maybe it’s just 10, or 15%, can stay with him for all of the other integration that has to take place, or adoption of requirements or changing of requirements has to take place when you adopt a commercial solution. And they have to integrate all these kinds of unseen costs in the way that you have to integrate a commercial solution that’s not exquisite and built for your legacy system. That is hard to anticipate. So I think recognizing that, you know, the faster law says three tier, right, the third tier is, even if the commercial software or commercial product doesn’t meet your requirements, it directs the government to adopt their excuse to change the requirements to adopt the commercial software. So you’ve got to give them a little bit of a slush fund. If they, if they, adopt commercial software to meet their requirements and have to change their requirements in order to make that fit. You’ve got to give them a bit of a slush fund. So I think has to work to be done on helping the agencies have more authorities and work to be done to incentivize the primes to be hunting for, you know, the differences when I’ve ventured backing onto I have a group 75 scientists engineers in Silicon Valley, those venture capitalists are paying for those scientists to be on the bench to work on the mission use cases that volume prioritizes. No other services firm has that kind of latitude to put 50-60 million into a product development, they have to wait to be paid by the government to develop and hire and build the team against that problem set. So if we’re going to stay ahead, you have to use Silicon Valley, I say Silicon Valley, it could be Austin. I know you have a great base in a number of different places in Boston and Austin. But the idea that the innovation ecosystem doesn’t build the great companies of America has to be harnessed by us. So we’ve got to move to a more proactive position for our agencies in our primes to be able to adopt commercial software. And I would say the third part of that Hondo on software, if I can, is that, if you when you’re talking to the legislative and, and and the executive branch, my recommendation would be to make sure they’ve got the will, that both the resources and the will to, to evaluate, to constitute their mindset to be constantly evaluating the latest capabilities. So they know how far you know, you’ve got to have resources, and a training environment separate from the operational and production environment. As you know, because of your time with software, it’s you got to have a big sandbox that you can test these buddy technologies in individual agencies may or may not have the bandwidth in their network to create a sandbox, I’d like to see DOD, I’d like to see, you know, the Navy, the Air Force, the army have these kind of technology sandboxes to test new technologies in and have it be like red flag or haven’t be like the big exercises or impact that you were on. And I liked this idea that software is starting to dominate our ability, even to exercise hardware. So we’ve got to have software oriented environments that are kind of free of charge. If you’re trying to displace a mission with software that’s commercially available, we’re going to give you the environment to test that baby and and we’re not going to hold your program accountable to the field software, you know, we’re gonna let you test and evaluate and and then make a decision to adopt in a positive construct. Does that make sense Hondo?
Hondo Geurts 33:28
You got it, I think it’s a great suggestion.
Lauren Bedula 33:29
You both alluded to the point. If this network is going to be successful, it’s not just going to be the US and so on to get your thoughts on how you approach international partnerships or business as well how you think about that.
Balan Ayyar 33:43
Okay, so let me just make an incentive upset, like I’m complaining a lot. But let me just tell you why you want to do business with the US national security intelligence base, everybody buys your software, if you win with the United States Navy, use you if the US Navy is using your software, US Army’s using your software, everybody in the free world wants your software for similar missions. And it’s true in the intelligence community too. For example, our first international customer is the United Kingdom. You can kind of imagine who in the United Kingdom, I don’t really talk about our customers. But let’s just say they’re interested in the security of the United Kingdom in the way that our intelligence group is interested in the security of the United States. And the only reason that the United Kingdom is our first international customer is they had colleagues that understood that our software had been taken to scale by the US Intelligence Committee. Can you imagine it’s the best business development? If you have one of these go with our customers in US national security and intelligence for your software, the world comes to your doorstep, you have to think about the risk, you have to think about how you’re going to manage the deployment, you have to think about all the implications of operating internationally. But it’s so exciting. So our first international customer is starting at the end of this month, working on all the issues that the world is facing with counterterrorism and other issues for Western liberal democracies. They want the Mirage software. And it’s not because we have great advertising, we don’t advertise the spy I guess is the first I’ve done in six months or a year, you won’t find me on panels, different conferences. I just don’t do it right, we got to let our software do the talking, our software has to produce outcomes. And we have to let our customers do the talking. And that might be the second point I’d mentioned, Lauren, I do feel like some of these technology companies, they get in trouble because they promise ahead of what they can deliver. And, boy, my recommendation with these net security intelligence customers is not to do that. I mean, I don’t know how else to say it. It’s to be very humble, and have a lot of humility, about what the latest capabilities can bring to bear. It turns out these well trained operators, their minds are like the cosmos, compared to what we can offer, you know, all of the vision capabilities, we can offer, maybe 10, to the fourth or 10, to the fifth neural processing, the human brain is 10 to the 35th. And so we have the mental we have it, we have the processing power of a small lizard in Mirage, compared to compared to one Naval Special Warfare operator, what are we doing we what we are is essentially, we can do something, it turns out the lizards do some things that are quite impressive, and can be quite helpful. And so our software can look at 1000s of hours of visual information, and triage for these incredible operators and analysts. The elements of the data that they want to that they’re interested in. And that’s not a thing, that’s not something the brain likes to do, right look at 1000s of hours of data to find where Hondo and Bollin have been seen before. But boy, once I give him the top 10 hits at Starbucks and at Jacksons, they’ll go to town figuring out who Honda was seen with what car he drove, who he came in with, who he left with, they’re operating at a higher level of their cognitive training. And it’s exciting for them to have an apprentice like Rush. But what you don’t want to do is over promise and under deliver right is what we see is why we get AI winters because we say we can do all these things we can turns out the frontier of artificial intelligence is completely nascent. The ability to contextualize still belongs to the providence of our amazing men and women in intelligence and in uniform. And don’t believe, you know, what, you’re what you’re experiencing with chat GPT 4.0, with these foundational models, you know, chat up 4.0 Doesn’t never experience gravity doesn’t know what the color blue is, it knows what the relationship between words are, and all the things that it’s been trained on with transformers and, and these large language models. But it can easily create things that aren’t true that you can’t do with intelligence, right? If you’re operating in estrogen intelligence, your AI can’t make stuff up. It just can’t, right at the risk of trust, which is the currency of the realm. So you have to be very careful about over promising and under delivering, I think Hondo may have a thought on that. But my recommendation is, we can help you accelerate your understanding. And in every mission, that’s a little different, some missions that may be higher than others. So that makes sense, learn. It’s not one size fits all, when you have an AI enabled software, it depends on the mission, how far forward we can take you and your understanding. So I think you have to start with a lot of humility about the use cases, and how incredible our men and women are that are balancing these eyes’ intentions and making decisions and where your software can accelerate. You have to let them test, evaluate and confirm before you talk.
Hondo Geurts 38:13
So great points here. I’m gonna shift a little bit. Again, you had unique experience running recruiting for the Air Force. You’ve now been, you know, you’ve talked about some of the talent you’ve been able to get on the technical side. But what’s your sense of bringing talent into national security? You know, 4567 years ago, that was a struggle, right? Is that changing in earnest? And are you seeing, you know, young talent being attracted to these missions? And what more could be done there?
Balan Ayyar 38:44
Yeah, well, you’re dealing with really long term challenges Hondo which I love. Of course, my heart is broken and saddened by the all volunteer force and the challenges we’re facing. I don’t think that the Navy and the seals do it the best, but the Navy has done better than the Air Force and telling our stories. And the army too. I think, you know, if I could be king for a day, I’d like to take overall recording for the army, all recruiting for the armed services, because I think we have great stories to tell, which resonate with young men and women as they did when you were leading in Special Forces and a navy but, and certainly even in the Air Force. We have great stories. And so I use the stories of what we’re doing and why it’s important to keep the best technologists in our firm and just the grace of God, we’ve been able to do what we’ve, we’re recruiting against our scientists, Amazon, Google, meta, everybody. And you know, we’ve kept our top scientists for five plus years, they could have million dollar a year opportunities. In fairness, we did lose one for exactly that million dollar a year salary, plus all kinds of things from Amazon. They were in trouble. And he’s the top scientist in the world in this field, and they needed him and so he’s there now, but everybody else has stayed with us, including the rest that they tried to recruit. And the reason is because of what you said. For the first time in their lives, not building a dance app, not working on things that you know make it easier for us to collect tolls. wills, and to collect taxes and you know, all this software that’s making the world more efficient, none of which really moves your heart or moves your soul. And so, you know, I don’t want to, you know, be too confident here. But I do think the work in the national security intelligence space is completely undervalued. And I do think there are brilliant young men and women in this next generation that really care about what and how they’re spending their time, what they’re doing and how they’re spending their time. So for us, if you came to one of our company meetings, I spent the first few minutes talking about how important our software is. Let me just give you one vignette. So come just license our software. You know, the story is such a beautiful story, I mean, to Special Operations Forces members lost in a Middle Eastern nation, it turns out just a couple of years ago, turns out there were surveilled the entire time, they were there on a mission to build foreign internal defense and to help strengthen that nation was a partner nation for us their wonderful relationships. They didn’t realize they were being surveilled before they were killed, murdered really, by a third party actor that was also operating in that foreign nation. And of course, in the analysis, after the fact we realized these guys had been surveilled from the time they landed, and Mirage, the vehicle recognition module in mirage that you can use tells you whether you’re being surveilled, you know, you put that camera, the back of your car, you’re driving on a route in whatever country you’re in Syria or otherwise, and Mirage, you take a little SD card out of your GoPro, whatever camera you’re using, you put it in the Mirage laptop, it’s a native application, immediately up comes your route. And here are the vehicles of interest that we’re seeing across that route that may be potentially surveilling you seeing across multiple segments, different directions, time, and it’s so beautiful, it’s very hard in a busy city environment to know whether other intelligence services are following you, and if they are for what purpose, and you know, so conscious license that and that’s gonna go across the world, with their operators. And I tell my firm, you know, if we can, in some small way, make our forces safer, every one of you ought to go home and have dinner tonight knowing that you contribute to the safety of our men and women in uniform who are protecting our values, protecting our freedom. And now we know because of war in Europe, those are things we can take for granted. So I spent a little time talking about the story of our software and why it’s important. And of course, at a higher level visual intelligence Hunter, I think, you know, and Lauren, you may know from your time at Ben’s, everybody around the world is gathering data, it’s kind of what we call a ubiquitous technical surveillance environment right now. And so our adversaries are trying to take apart I would say, reveal what our Special Forces, what our intelligence officers and what our our diplomats are doing, and to preserve our freedom and to preserve our values and our influence in the world. And they’re trying to conceal the way they’re operating in these global environments. And so our software helps, in one small way, our forces in our intelligence community deal with the threat of continued surveillance in every global environment. And these are the kinds of technologies that we need in order to make intelligence work. And so I just use those stories with my company to try and keep the talent and then when we recruit Hondo, we’re very upfront and forthright that we believe in America, we believe in American values. I don’t think we should be mealy mouthed about that. I think some people kind of skip around it in Silicon Valley, and I think it’s getting better. But what we share with them is listen, we love Silicon Valley, we think it exists because America believes in, in, you know, a market based, meritorious, you know, competitive economy that goes after the brightest talent for the most important reasons. And we think what builds Silicon Valley is what’s going to guide us through this next century of challenges to the human condition. And whether that’s for climate change, or whether that’s for keeping, keeping free and open societies. We think technology has to contribute. So anyway, it’s a story that we tell Hondo. It’s the same story that you were telling when you were in the service in terms of how important this work is. And I think the right people respond to it. That’s my, it’s not easy, though. I gotta tell you, there’s some folks that would rather not be involved in work that, you know, potentially, two sides of the coin potentially can be used offensively and defensively. But we believe that we’re on the right side of history with our values.
Lauren Bedula 44:05
It’s exciting to hear more and more talent drawn to things like force protection or mission and the realization that you can contribute even from the private sector or thinking through those public private partnerships. So on that note, Balaam, I just want to say thank you so much for joining us today. And for all the advice to our listeners, I think your points about not over promising when it comes to the tech and humility are great ones and really focusing on use cases and thinking through incentives for more partnerships. So thank you again, Milan for your time.
Balan Ayyar 44:35
Oh, thank you, Lauren. Thank you Hondo. Great to see you both congrats on a wonderful podcast that you’re on and enjoyed it. joy to be with you and I look forward to working on these big issues together. Alright. Thanks, buddy.
You’ve been listening to building the bass, a podcast from the business executives for national security. Join hundreds of senior leaders and executives dedicated to the mission of keeping our nation safe. Check out our projects we’re currently working with important Coming events and the many ways you can get email@example.com
Building the Base Episode 40: Shelly O’Neill Stoneman
In this episode of Building the Base, Lauren Bedula and Hondo Geurts welcome Shelly O’Neill Stoneman, Senior Vice President of Government Affairs at Lockheed Martin. In the discussion, Shelly shares…
Building the Base Episode 39: Joe Anderson
In this episode of Building the Base, hosts Lauren Bedula and Hondo Geurts welcome General Joe Anderson, former Deputy Chief of Staff of the Army, G-3/5/7, and current President and…