“I think we have to go and take big risks in order to get big rewards. And startups are a really good ecosystem for that. And so if you see, and if you can foster that sort of high risk, high reward innovation, then that’s important for everything that our country needs.” – Daeil Kim
In this episode about startup innovation, Daeil Kim from AI.Reverie and Daniela Perdomo from goTenna join host Ken Harbaugh for a conversation about partnering with the Department of Defense as a startup.
Daniela Perdomo is the co-founder and CEO of goTenna, a startup that designs and develops technologies for off-grid and decentralized communications. Find Daniela on Twitter at @danielaperdomo and goTenna at @goTenna.
ACME General: Welcome to Accelerate Defense, a podcast from ACME General Corp. I’m Ken Harbaugh, Principal at ACME, and host of this month’s episode. On Accelerate Defense, we’ll hear from political figures, military professionals, and other thought leaders about how innovation shapes our national security landscape.
This month on Accelerate Defense, I talked to two CEOs of tech start-ups about what it’s like to work with the DoD. My first guest is Daeil Kim, founder and CEO of AI.Reverie, a simulation platform that trains artificial intelligence to understand the world. We’ll talk about how AI.Reverie does that and how it has partnered with the DoD on some really interesting work.
Daeil, I want to start with a bit of your backstory because it is fascinating. First of all, welcome to Accelerate Defense.
DK: Thank you. Happy to be here.
ACME General: Your path to becoming a CEO of a tech startup wasn’t quite traditional, you studiedLit (Literature) at Sarah Lawrence College. And if my notes are right, you never took a math class. Can you give us the 90 second version of how you go from that to running an ML and an AI company in the tech space?
DK: Yeah, so the 90 second version of this is I got a liberal arts education at Sarah Lawrence focusing on literature. After that, I wanted to study the mind, mostly through psychology so I started volunteering at labs at Columbia. That led to doing research in schizophrenia and studying MRI and studying how those brain images can actually be used as a biomarker to understand that. That led to more computational neuroscience, which then led to a desire to then study computer science and artificial intelligence, which then led to doing a Ph.D. at Brown focused on understanding how to basically do machine learning there, scaling machine learning and all of that. It was the transition that occurred over several years so it wasn’t like I just graduated from college. I spent about three to four years honing up on things and just learning more about the brain and really at the Ph.D., that’s when I really got to understand how things worked.
ACME General: I’ll see if we can tease this out over the course of the interview, but I think there’s a metaphor in your pathway for this trope of finding innovation in the unlikeliest places, which is something that we are really into at ACME General Corp and trying to illuminate for government clients and the primes that sometimes the most creative people, the most unconventional approaches to problem solving are found in the unlikeliest places like the Lit program at Sarah Lawrence College.
DK: Right. Exactly. Yeah. I think creativity is a function of, to some degree, of being able to expand your mind, open your mind to new ideas and new approaches, not being always stuck with a conventional doctrine. Certainly something that Sarah Lawrence has taught me was the importance of narrative and understanding stories and understanding how things evolve and appreciating subjects such as history, for example. Being able to just really get a sense for how the world works.
In that sense, that helped me sort of keep in mind innovations and the importance of innovation. I also want to stress though, the importance of an education that I got at Brown and doing a Ph.D. there. One of the nice things about doing a Ph.D. is that it really gives you a very deep understanding of the landscape of things. And I do think that creativity is tied to that. To some degree, one has to understand the borders of where things are before innovation can happen and understand where things are lacking. And so I would say that Sarah Lawrence certainly gave me that sort of sense of creativity, but also the Ph.D. gave me a grounding to know what problems really needed to be solved in the world.
ACME General: Speaking of problems to solve, you are now leading a company that is solving a massive problem for the future of ML and AI, which is training algorithms. Can you talk to me about what AI.Reverie does and what’s the problem with real world data? Why don’t we just stick with that?
DK: Yeah. Well, I think one of the biggest bottlenecks that actually is happening in computer vision is the lack of data. And there’s a lot of reasons for that, but data is just really hard to come by. It’s fickle, it’s really noisy. And of course there’s the problem of annotating that data in order to train deep learning algorithms. The main dominant paradigm of how to train the computer vision algorithm is to get images paired with metadata, such as bounding boxes or masks that represent interesting things within the image.
Obviously that part of the metadata, that has to be done by a human and that represents a huge bottleneck in terms of being able to accelerate the field. That, and as a company that started to solve that through the use of synthetic data and the way we define synthetic data in the context of computer vision is data that comes from images that might be generated from software like a game engine or through algorithms themselves. And so the power you get from that is that you don’t need humans anymore in the loop because you’re generating it yourself. You know where everything sort of exists and lies and it allows you to create any kind of metadata you want, that you can train an algorithm for. That really accelerates the whole process.
ACME General: Give us the use case. And if I can impose on you to talk about the RarePlanes use case in particular, that’ll scratch my Navy pilot itch. Tell us what you have done there and why synthetic data is so much more manipulable and saves so much time.
DK: Exactly. This project RarePlanes, we worked with a lab based off of In-Q-Tel, called CosmiQ Works, and the collaboration there was CosmiQ Works would be responsible for actually getting real world labeled satellite images of particular planes and we would be generating the synthetic counterpart. Now the big problem there is that: A) satellite images are expensive to both acquire and annotate. The annotation itself is really complicated for the things that they wanted specifically around the planes, like the number of engines, the kind of wings that the plane has, the various advanced labels that one can generate from those things, with the problem of detecting.
ACME General: That annotation is all human in the loop, right?
DK: Yes, for the real world data, yes. Humans have to go in there and look at the airplane and decide what kind of airplane it is, understand the kinds of features associated with that airplane and label all that properly. It turns out that result, not the best labeling performance as well because of the complications there. There had to be another layer of going in there and fixing those labels by domain experts. As a result, there’s a lot of issues of course, with that process.
But the challenge is that there are certain planes that rarely appear in the satellite images. If you’re looking for a particular airplane that is really, really hard to find, you might have to wait days, weeks, months for that airplane to even pop up or even pop up at all. What you can do with synthetic data is you can actually just create a 3D model of that and be able to then place it in a simulated world that we created. We actually replicated real world airports around the world and we were able to place them at any configuration, orientation. We can change the environment, the weather, all of these things. And so we can create thousands or tens of thousands of examples of this plane that you might only have one single image of in the last six months you were looking for it.
ACME General: I’m glad we’re talking about this because I think it is piquing the interest of our listeners with ties to the defense industry. And I want to talk about how you have engaged DoD, but you didn’t start there. You are very much a dual use company and you started in the non-DoD space, right?
DK: Sure. Yeah. We have a lot of Fortune 500 clients that we work on. But I will say this in terms of the government, I think in many ways the government has historically as well, been one of the biggest proponents and supporters of research into areas like this. They’re sometimes one of the first people to take the risk of saying, “Let’s try to make this work.” Because they’ll get the idea, even though it might move a little bit slower, given the sort of levels of things we have to get through in order to confirm a contract, they are some of the biggest supporters of this kind of work.
ACME General: For context, can you describe your current engagements with DoD? I understand you’ve got a SBIR phase two, you’ve got an IDIQ. Let’s describe the landscape so people know that you’re speaking from a position of authority here. And then I want to talk about the pros and cons of going down this path.
DK: Yeah, the stuff that we can talk about publicly is definitely the SBIR phase two award. That one’s really interesting where we’re actually trying to create an algorithm that can work in GPS denied environments. If you’re a plane that’s flying and you don’t know where you are because all the GPS satellites or towers in some hypothetical scenario are gone. Can you just use a computer vision, look at the image outside of the window and be able to tell you where you are in the world? And so one way of doing that, is to sort of tie and simulate a synthetic replica of the world and then have perfect metadata to GPS for that and strain something. It’s a bit more complicated than that, but that’s the high level idea.
And the IDIQ contract is definitely something that’s really exciting. We’re really glad to be part of that. And there’s going to be a lot of, we haven’t defined a project exactly just yet. What that contract allows us to do is that we’re one of many vendors to take part in being able to modernize the things when it comes to data, at least, for the Air Force as well.
ACME General: Given all of your civilian clients and that that is the bulk of your business, what about these opportunities within DoD drew you in? Was it a financial imperative? Was it the opportunity to be at the cutting edge and to have the chance to take risks with these programs?
DK: Yeah, well certainly the financial thing is important. The government is willing to pay for the work that we do. And also they’re willing to take risks as well. Often when you think about Fortune 500 companies, the risk taking there might not be as high because they have to account for every dollar in a way that leads to some kind of profit at some point down the line. And another thing is that these are some of the most important and difficult problems to solve if you think about it. If you are thinking about how AI is going to be so fundamental to the functioning of government and being able to do important things in the world, you want to be a part of that, I think.
ACME General: Was this path always envisioned when you started the company? Did you know at some point that you would be leaning on SBIR-like contracts to help the business along and to incent the type of risk taking you wanted to do? Or was it an opportunistic thing later on?
DK: Yeah, I think it sort of creeped up on us a little bit in terms of where. When you’re doing a startup, especially in the space that’s deep tech and that’s when we started about four years ago, nobody really understood synthetic data in the commercial space and its possibilities. We were just sort of looking to see who would be interested in that. And so it wasn’t like we were going to say, “We have an entire government plan. We’re going to go ahead.” We have to do a lot of learning on our own over the past several years. And I certainly wish we had sort of a manual for if you’re a tech startup starting off, how do you engage in work with the government? That was just sort of trial and error for us, to be honest. Yeah.
ACME General: Well, part of this show is to help flesh out that manual, if you will. What are some of the biggest challenges that you had to overcome or are continuing to overcome in this partnership?
DK: Understanding the right decision makers and getting a better sense of time in terms of how things will sort of turn out. The nice thing about working with corporate clients is that it tends to be faster. You meet with them, you can sign a contract within a week or two and then just things start getting moving and then you can start working. Oftentimes with the government, you sometimes have to wait and you don’t know when the funding will appear or when something will go. I think they’re doing a much better job in the past year or two, for sure. I think they’re innovating in that space, but before it was always a question. And there were certain projects we worked on where there definitely wasn’t an alignment of what was actually in the academic world and what the latest, best software tools and platforms were that should have been implemented. There was a lot of older ideas that people were trying to still work with in terms of creating the best environment for AI development. I think that’s changing now, but I think that was also a surprise that we saw in the beginning.
ACME General: Is the funding lag time that you identified the biggest issue? We had a conversation recently with David McCormick, who really identified that as one of the biggest challenges in onboarding companies like yours.
DK: Yeah. I think so because startups, sometimes they have a lifeline of six months. Never a runway for six months to a year or so. And then you’re raising money and it’s that cycle. If government wants to work and be able to support startups, just having a much more predictable sense of when the funding will come in and when things are going to be guaranteed, can go a long way in allowing startups to feel more comfortable working with the government. That’s a big part of it.
ACME General: Help us understand how the process begins. I’m sure some of it is undergoing a little bit of that mental shift. Four years ago, you’re working in the proverbial garage, you’re getting this thing off the ground. At some point you realize, you know what? A partnership with government makes sense. How does that happen? Does somebody come to you? Do you reach out to the SBIR office? And then who’s holding your hand through that process?
DK: Yeah. What we had to do in our end and it might be different for other companies was, we started off I think, just trying to see who knew who in the government. Do you have connections here? Are you connected to here? A lot of conversations. We would follow BAAs, broad agency announcements. Search for that. Look for things related to computer vision and data. See if there was a BAA announcement around that and then just follow up from there. And what would be really cool is let’s say you have just some kind of search engine, where do you type in what you work on, what you’re trying to create and it just immediately lists out all the possible government BAAs out there and things like that, that you can get access to. And it’s easy for a startup to connect to the right people to show their work in that sense. That’s something that we kind of had to do through conversations, talking to people, getting consultants to help guide us through this process as well.
ACME General: What has been the best part? And hopefully you can share something you didn’t expect. A surprise from working with AFWERX on this. Has anything stood out as a real positive feature that you didn’t anticipate?
DK: I personally like working with people in the government. I think there’s something about the work that they do and the sort of integrity they bring to it and the willingness that they’re giving their sort of their hours, their labor into the service of something outside of themselves. I think that’s one thing that I think I appreciated getting to know more people in that space. Folks in the military, things like that. And how much they care about the problem and how the problem is not just about profit. The problems are about real challenges that are affecting people all over the world. And just also the cause of just being able to sort of work on something that might help protect a soldier from something like an IED or something like that. There’s real value in that, that I think is important.
ACME General: We hear that a lot from people in your position, bringing new thinking and new technology to bear on these problems. There’s just a real affinity for the mission first service orientation of the folks you end up working with through these programs, but at the outset in your company, was there an internal debate? Because we’re hearing that as well, less from people who’ve actually begun the engagements than just in the news media about trepidation within tech companies working with government.
DK: Yeah. of course, there’s certain comp tech companies out there with outside voices that have in the media that show that there’s a lot of discontent working with the government. And as a result, there were definitely internal conversations at the company about how we move forward. But I think you have to take a position on this at some point and stick with it. And that’s kind of what I did. I said, “Okay, I get where they’re coming from and I think I respectfully disagree with their perspective on some of these things and we’ll go our own way.” And I don’t think I’ve regretted that decision.
ACME General: Am I reading your answer correctly? That the decision wasn’t just a business one, but a moral one?
DK: Yeah, I think so. Yeah. I think it’s important to work on problems that are well, it’s both. But I would say that the morality cannot be disentangled in some ways. In terms of the company’s culture and DNA. And I had to do some explaining to the company as to why we’re going to work with the government. When I bring in new people, I do the thing of letting them know that we work with government projects. How do you feel about that? Are we on the same page? If not then this company might not be the best fit for you. I have to go through that conversation with also potential employees. But yeah, I couch it in terms of, I take a moral stance on it as well, personally.
ACME General: How do VCs in your orbit feel broadly about this? I’m sure there are nuances, but is there an overriding sentiment among the VCs that you work with about this aspect of your business?
DK: Yeah. They would much prefer that all our clients were in the Fortune 500 and we have a very easy to sell SaaS model that allows them to get predictable revenue. A lot of VCs will flock to that. But, the thing I will say is that, this is deep tech. This is not like just a web app that you’re building that’s similar to 20 or 30 other web apps that are out there. This is fundamental research that is being productized to service customers. And so as a result to say no to government revenue that is willing to take risks in terms of improving this kind of research and improving this kind of technology is I think the wrong move. In fact, you want their support and you want to be able to help solve their problems and that’s only going to strengthen the company. But the conventional sort of thing with VCs, a lot of times, they will be sort of wary of if too much of your revenue comes from the government, but that’s more around the predictability issue, I think.
ACME General: Sure. Sure. Have you been able to convince your funders to think more openly about working with government?
DK: I do the best I can. It’s hard to know exactly how they fall into these things. If you’re focused primarily on profit and you’re thinking always about how do you get to this 100X kind of thing, you want to minimize as much uncertainty as possible about getting to that unicorn status, which is what a lot of VCs care about. It’s hard. People have very distinct philosophies about these things and they’re not willing to budge because it’s worked for them in the past. But I think I’ve been able to maybe push some minds a little bit more towards where my head’s at and hopefully it’s going to be more in that direction. Because I do think that public private partnership between the government and tech companies is crucial for this country.
ACME General: I definitely want to get your read on that. But I want to address the third leg of the triad before we do, which are the primes, the big defense industrial based companies. Are you at the point yet where the primes are looking at you, where you’re having conversations with the big defense contractors about how to truly scale your business?
DK: Yeah. We definitely have conversations with them. We’re always open to working with them in the right way. But I think because there’s other venues like SBIR and things like that we’ve been able to go on. I think we’ve also been really trying to leverage that. The primes oftentimes, they’re big organizations, so sometimes integrating what we do takes more time as well. If we can, I think to be honest, we do try to lean a little bit more on being able to go and just do it ourselves. But if we can work with a prime and we’re able to engage effectively, then we’re happy to do so.
ACME General: Well, you’re answering the question without me even asking it. Do you see that sclerosis within the bureaucracy of these primes as a barrier to innovation? Or do they take your call?
DK: No, they’ll always take the call. They’ll always be interested in what we’re doing, but I think what happens in really large organizations is that there’s too many decision makers. There’s too many people who have to say yes or put the check box to move things forward. While in a startup, it might be one or maybe two people at most. And so things just generally move faster in a startup like ours and sometimes we’re just pushing and pushing and pushing to get ahead. And that’s the nice thing about working. Of course the other side is that it’s much easier to sort of say that the primes are going to be there 20 years from now than a young startup, but it’s the decision makers, I think. There’s too many sometimes at a large organization to allow for efficient processes to happen.
ACME General: Does it also affect the innovative spirit and not just the speed with which processes can occur?
DK: Yeah. I think it affects everything to be honest. I think it affects everything and I and of course there’s always exceptions and I’m not saying that’s the universal case with any large organization. But even places like Google and Facebook and things like that, innovation just inevitably slows down, in some ways because of how many people there are. And so that’s always going to be a really tough challenge for a large organization to solve. And this is where startups have a real advantage.
ACME General: Startups supported and empowered by programs like SBIR and AFWERX and the other initiatives, right?
DK: Definitely for us, yeah.
ACME General: Now those have been described as positive as they are, as innovation archipelagos. This idea that they’re really kind of islands unto themselves and good things might come out of them, but they don’t really stand a chance at challenging, to keep with the metaphor, of that continental momentum of the primes and the super big contracts. Does it feel that way where you are? Do you feel like you’re stuck on an innovation island? Or are you more hopeful than that?
DK: No, not at all. I think we have a pretty clear vision of what we want to build. I think it’s going to be pretty amazing. And once we get there, I think we’ll know how to actually make it work within the government. If it takes hiring a bunch of extra people who can specially tailor a system so that it works in the right places and is deployed correctly, that’s an engineering thing that we can solve. But I have no doubt in my mind that if we build a superior product than the government, we can make it work there.
ACME General: If there’s one thing that you would tweak about your partnership or the contract vehicle that you’re on, what would it be?
DK: Any particular contract vehicle or just in general?
ACME General: SBIR, let’s say SBIR. We’re talking about how to engage small companies with the government to add them to the mix.
DK: That’s hard to say because that SBIR, that’s been a very positive experience for us to be honest. It’s been great working with the people we are working with at this point on it. In terms of tweaking anything, I would just say that just being better about these contracts, just knowing exactly when you’ll get a response. Just being able to time these things and knowing how that structure is going to go a long way to getting startups, to feel good about those things. But in general, it’s always trying to get the certainty around knowing when fundings will happen and who will sort of be reading your contract and who are the sort of eyes engaged there. Are you just sending something to the ether? And will there be somebody really responsive and taking a serious look at these things? And do you have to make the right connections in order for that to happen? That’s the part, the murkiness that I think really could be improved.
ACME General: Are there cultural challenges as well? Not just programmatic or streamlining the system, things like that, but are there times where you feel like two different languages are being spoken? We hear that from folks on the military side, when they engage the tech side, do you feel the same way ever?
DK: Sometimes. It depends on who you have on the military side. If you have people who are doing similar things and are aware, then the conversation flows pretty easily. I think the challenge for us, as a company, and I’ll speak only for ourselves is sometimes the deployment of technologies because of clearance and because of trying to make sure that those things, information is secure, that’s always a challenge for us. Because as a startup, if you don’t invest and most startups don’t go in thinking, we’re only going to work with the government. And this is the infrastructure that they have and we have to make sure our software is automatically working there.
I think there has to be a middle ground actually. The government has to figure out a way to be able to work with startups that are leveraging new technologies and to be able to sort of say, “Okay, we’ll sort of meet you halfway here while making sure the most important things are secure on our end. And we’ll be able to give you some flexibility of being able to use your technologies in the space that allows us to see the benefit of it.”
ACME General: You talked about public private partnerships and their importance to the security of the nation. Can you expound on that a little bit?
DK: Yeah. A lot of it is tied to innovation. If you think about how young startups can innovate really quickly and introduce new ideas and try things, because when it comes to this AI race, it’s a pretty intense one. Anybody who studies this subject a little bit knows that there are foreign nations that are trying to actually win this race. And whoever wins the AI race is going to really dominate. If you’re playing that game, then what you want as a country is to sort of take as much, be able to also take high risk choices. I think synthetic data is one of those things, where it’s become less high risk, but early on, four years ago, people might have thought, oh, this is too risky and that. No, I think we have to go and take big risks in order to get big rewards. And startups are a really good ecosystem for that. And so if you see, and if you can foster that while also fostering other things, but really fostering that sort of high risk, high reward innovation, then that’s important for everything that our country needs.
ACME General: Lastly, Daeil, what do you think the biggest challenges in the national security arena are facing our country in the next three years and in the next 30 years?
DK: Well, of course I’m biased towards synthetic data. I think we need to solve the data problem in machine learning. I think we needed to solve the ability to train algorithms the way we want to, synthetic data and simulation environments are one way of doing that. And the way I look at the progression of things is I think you have to solve perception as a problem first and then once you solve perception, then you can solve robotics, in terms of getting autonomous robots to navigate the world. Because without having a perception layer, it’s really hard to know, to tell a robot, to get a beer from the fridge if it doesn’t know what a fridge looks like. Need to solve that. We need to solve that as a country or however we want to do it. We need to be ahead of that game.
And then once we solve perception, then we can solve robotics. Once we solve robotics, then we can solve a lot of other problems that exist in the world. I look at that. And so in 30 years, I would say one of the things that we have to be very careful about is, what are the things we create? How much control do we have over the algorithms we create? Over the AI systems we create? And I think this is also another reason why simulation environments will be the perfect place to train AI because there you actually have control. And that the decisions we make about how we train the algorithm, we will at least be able to be held accountable for it.
ACME General: Well, Daeil, it’s been great having you on the show. We always pull a clip out to serve as a topper for the show. And I think we’ve got ours that the biggest national security problem is training robots to get beers from the fridge. If we can solve that one, I think we’ll achieve world peace.
DK: Okay. Let’s hope so, Ken. Let’s hope so.
ACME General: My next guest is Daniela Perdomo, co-founder and CEO of goTenna, a startup that creates technology for off-grid and decentralized communications – an idea that was inspired in part by the cell tower outages that were caused by Hurricane Sandy in 2012. I wanted to talk to goTenna as an example of a startup that has successfully navigated the government procurement system.
Daniela, welcome to Accelerate Defense.
Daniela Perdomo: Thank you for having me.
ACME General: I love the founding story of goTenna. I have a background in disaster relief with Team Rubicon. But I also understand from doing a little digging that there’s a little more to the founding story that might or might not involve your brother attending raves in New York. What’s the deal there?
DP: Yeah. So in the summer of 2012, my brother was going to a lot of music festivals, I believe still does and noted the fact that communications tends to fail in very congested environments where a lot of people are using the same infrastructure. And so he had an idea to solve that problem. He asked me a little bit of advice I had experienced in Tech Startups and I said, “All right. Good luck to you.” However, then about a month or two later, Hurricane Sandy decimated, not just cell towers, but power stations, which meant the Internet was down as well in New York City and in the 10 state area affected by the storm. And for the first time ever, I started to think about the ways in which our communications infrastructure tends to fail us when we need it most.
It’s not just in congested environments, it’s of course in disaster situations. And so it was a confluence of different use cases that we saw both firsthand. And the more we thought about it, the more it seemed like no one had thought of creating a communications network that was in its very nature, decentralized and bottom up. It really runs in the face of what we think of traditional communication systems as, and I don’t know that we would’ve started thinking of it this way had we not had absolutely no communications experience prior. It’s one of those naive questions that led to a novel area in terms of IP and the technology itself. The company was very much inspired by consumers, by civilians. But from the very beginning, we did identify that there was probably a big market in the government and as it turns out there is.
ACME General: Well, I wanted to ask you about something you just alluded to, which was the personal lack of experience in this arena prior to jumping into it. In some ways, that was an asset, because you didn’t realize what you were up against, right?
DP: Absolutely. I mean, if you had told me eight-and-a-half years ago, when we started working on this, that it would take as long as it did to develop this novel technology, I wouldn’t have believed it. I think that some of the lack of expertise actually probably made it easier to dive into something had I known it was going to take as long as it would to get first-generation consumer product, which was about two-and-a-half, three years, second generation, another two years over that, the professional product thereafter, I might’ve been a little more daunted.
ACME General: And this wasn’t just a coding challenge. You’re building something. You’re sourcing hardware and components and beginning with the consumer market, right?
DP: Yes. So we develop devices both for the consumer, civilian market and the professional and government market that pair regular phones and computers and enable you to send, to communicate on a peer-to-peer basis, over really great distances with others who have these devices. And not only can you communicate on a peer-to-peer basis, you can communicate peer-to-peer, to peer-to-peer. So you’re able to relay hop, Daisy chain, use your metaphor of choice to transmit messages, to people who maybe aren’t within range of you, but who are in range of other people who aren’t in range of you and so forth. And so goTenna the consumer product, the goTenna mesh product operates on the 900 megahertz ISM civilian band.
So this is a public set of frequencies that all kinds of consumer technology works on. That one, I think has… it emits one Watt. So that gets a certain amount of range in the UHF spectrum and then our professional product works through a licensed spectrum in the UHF and VHF range and up to five Watts of admitted power. So can get a lot more range and that is solely for professional use. So these devices are essentially really smart radios that pair to data radios that pair to again, your smartphone or computer or sensors and allow you to communicate on a completely peer-to-peer basis without depending on any centralized infrastructure whatsoever. So goTenna works even if you have no cell service, no wifi, no satellite signal.
So we extend the edge of connectivity to our normal communication systems don’t reach. It’s what is known as the proverbial last mile. That’s what we specialize in. But in addition to enabling you to communicate when everything else either is failing, is unreliable or simply cannot work. We can also plug into those systems. So a lot of our, both civilian and government customers do all kinds of back hauls into the normal LT network into wifi, into satellite, into wave relay, you name it. It can interconnect, and that’s what we mean by extending the edge of connectivity, both to where others cannot reach and to patch it back into centralized connectivity.
We’re talking even just point-to-point without meshing. I think our range record is about 69 miles. Typically you’re getting more like up to four miles if you’re ground-to-ground, that’s the laws of physics. That’s how far the horizon is, but you can get a lot more obviously, and essentially we’ve created this device mesh network that lets people send asynchronous data over really great distances.
That’s things like text, position and location information, all kinds of GPS use cases without any cell or Internet or satellite connectivity. The current users of goTenna Pro devices range from search and rescue teams to Wildland firefighters, to special operations teams all over the world. We even have news organizations that use us on the ground in places that are sensitive or where they’re sensitive to potentially security or privacy concerns. A few examples of how customers have used our technology, the LA Honda Fire Department in California mapped almost 23,000 acres over a couple of days with about 14 teams and ATAK, the Android Team Awareness Kit paired with goTenna Pros, awkward mesh network during Wildfire, during Hurricane Dorian, The HARP Rescue Non-profit Humanitarian Organization used goTenna Pros for coordinating delivery of medical supplies from the tarmac to the medical stations and into the hands of doctors. And their feedback was that if they were to normally save 10 lives, they thought that goTenna Pro enabled them to save closer to 60. So 500% increase.
ACME General: Am I crazy in thinking that starting with the regular civilian customer is the tougher way to go instead of with some niche, say emergency services provider or some forward-thinking DoD unit that wanted to experiment. I mean, you went for the big market right away.
DP: Yes and no. I mean the consumer market is really a great way to build a brand at the top of the funnel. Nobody writes or gets that excited about niche products other than niche audiences, but you can find the niche audiences through mainstream press and other consumer avenues. We also, frankly really believe that the business was going to end up solely focusing on consumers. We didn’t realize how big the opportunity was again, because we’re not veterans. We have not worked ourselves in the government. I mean, now of course our company is filled with experts of that sort. But that’s probably another way in which we were somewhat naive. So that being said, I think that launching consumer helped assess exactly what the addressable markets were really quickly because when we launched our pre-order campaign per first-generation product, within minutes of launching, we were contacted by a variety of different federal and state and local government agencies, not just in the United States, but across the whole world.
So they found us really quickly.
ACME General: Good. Good. Because that’s what I want to focus on, the partnerships that you have built with government, the contracts that you have secured, how did that start out and what were the conversations internal to goTenna like around pursuing those opportunities?
DP: So we’ve always been very interested in pursuing partnerships, engagements, and relationships with customers and partners who are willing to put the time in and the money in. And we were very lucky, very early on immediately following the announcement of our first-generation consumer product being on pre-order to have been contacted by a variety of different government agencies, many of which were willing to provide non-recurring engineering NRE which is a non-dilutive source of funding to accelerate product development. And indeed that NRE that we received, I believe in 2014 or 2015, paved the way for our first-generation professional product, which we started shipping in 2018.
And it was excellent because we got to work with a variety of government agencies that want to be the customer of that product. And that was really a huge acceleration. We knew we wanted to create a professional version of the product. We probably would not have moved to develop the professional product alongside our second generation consumer product had we not received NRE early on which confirmed and validated the market need and interest, and also made us feel a lot better about entering into the government space because we had our specs validated by at least some of the potential end-users from the get-go.
ACME General: Did the team have an internal debate about taking on these projects and working with the government? Because one of the things we’re exploring is that resistance that you sometimes hear about within the tech culture to engage in government.
DP: That’s a great question. We create connectivity where it otherwise doesn’t exist, which is a fundamentally enabling technology for any user. And from the very beginning, we were really excited to explore exactly who would need it. And we knew as civilian/more consumer experts, that there were all kinds of use cases from hiking to travel, to events, to emergencies on the consumer side. And even from the very beginning, my very first pitch deck for goTenna always had this one slide with a bunch of different potential use cases. And I remember there was a firefighter and there was a soldier. And so from the very beginning, we knew that there was this potential massive use case on the government side. And we just didn’t know how to access it. We didn’t know when we would, if we’d get the opportunity to et cetera. So for us, the biggest question when the opportunity to work with the governments early came about was NRE great.
Let’s make sure that… that doesn’t mean the government has any ownership over our intellectual property. And once that was figured out, then it was like, great. This is just yet one other amazing market that can use our fundamentally enabling technology. And I think that one of the things that makes goTenna such a mission-based company is that our technology is used by people who put their lives on the lines for others and it makes their jobs much more efficient and much more safe. It makes it likely they get back home at the end of the day, if they have a lifeline, both to each other and back to central command.
And that is something that we’re really proud of is the fact that lives have been saved using our technology. In many cases, we will never know the full extent because many government missions are classified. But the stories that we hear from the civilian public safety and law enforcement side of search and rescue missions of disaster recovery, where one non-profit said that they believed that they had a 500% increase in lives that they were able to save because they had connectivity at all times, which is not traditionally the case post-hurricane which, in the case of the specific non-profits.
So we feel very mission oriented. One of the corporate values at goTenna is impact, the idea that what we do has a real world impact in the highest stakes situations, the situations where people’s lives could be on the line. And that’s something that we as a team are very proud of.
ACME General: Well, I think you just answered it. But I’ll ask anyway, this is an element of cultural pride that permeates the whole organization. That’s something you talk about around the proverbial water cooler?
DP: Yeah. I mean we certainly pivoted from being entirely focused to consumer where consumer now is more of like a side business to our much bigger government business. But I think that the mix of experiences in our company, we have both people who have worked in the government, been in military service and many who have not and we have people who’ve worked at government contractors before, others have completely been on the consumer side. But we’re very proud of the fact that yeah, our technology has a real world impact on people’s lives. And not many technology companies can say that.
If you’re moving bits of data here and there, or you have an app that people enjoy using, but it doesn’t actually impact their real-world lives, that to me is a little less motivating. And I think that the people that have joined the company find it motivating to be really connected to practical uses. Whether that’s tactical or useful or not, it’s something you can hold in your hand and where the impact is really felt by the end user. We’re really tied to the mission sets of the core users. Even if we are not on the front lines ourselves, I think that is very exciting to our team.
ACME General: So you’ve talked about the upsides of working with government. Surely there have been a few downsides. Can you talk about the challenges of DoD in particular as a partner?
DP: Well, there are certainly a lot of things I’ve had to learn. We are a real relatively small startup, bureaucracy is not really part of our day-to-day lives. I think over the years now that we’ve sold into over 300 government agencies and into every branch of the DoD, we were part of a program of record with the army. We’ve had a lot of different experiences. I will say that government contracting is a black box. It’s hard to know how to forecast it. It’s hard to know in particular how to forecast timing, right? All those things can move and you might never understand exactly why something moved more quickly or a lot more slowly than you previously expected. But I will say actually that I think the DoD among government agencies is better to work with than others and perhaps that’s because of urgent operational needs, it’s because the contracting vehicles have a lot more experience working with smaller companies.
The DoD is an early adopter, right? So obviously has experience bringing new new partners into the fold over time. I have generally found the DoD to be a little less frustrating and less confusing than other government agencies, but that doesn’t mean we haven’t had moments like that as well. There have been moments, you could say there it’s even upside good surprise that can turn very quickly to a bad surprise if you get a rated order, which is an order that has been deemed so important to national security, that you cannot ship anything else that’s technically in front of it until you get that order out. And for a small organization like us, that can be very stressful and constrain relationships with other customers who all of a sudden, you have to say, sorry, you have to wait more time because a big federal rated order came in that we legally must fulfill before yours.
So, yes. Those are some of the kinds of problems we’ve run into. The other thing too, is that I think that it’s very hard to do business with the government, unless you have experts who’ve been on the other side internally. And I would say it probably took me a little too long to hire the senior experts who’ve done this before. So we probably, and I think a lot of tech startups who sell into government have been through this. I don’t think this is unique to goTenna, but we’ve learned a lot of things that probably learning on the job slows you down. And we should have had more experts in the organization sooner, which would have probably resulted in a few less headaches or lessons on the job than we’ve had. But now that we have more on a rhythm, it’s a bit less stressful and it all just makes a little more sense than it did prior to having some experts on hand.
ACME General: I hear you about learning on the job. Are there particular learnings or lessons that you wish you had had at the outset, especially when it comes to balancing a commercial business against national security business?
DP: Well, I should be really clear. Our business is now about 95% government. So we are now much more B to G business than we are a B to C business. The B2C business is more of a footnote, honestly. So just to clarify that, let me think we did in RE early on, as I mentioned. The genesis of goTenna Pro really came thanks to those early partnerships and NRE with the DoD and others. We then spent a couple of years not that focused on NRE. And I wish that the reality of how sibers or SBIRs worked had been more clearly explained to me at some point, because we would have applied for them a lot sooner. Now, we’re deep in that, including we recently got a phase one with AFWERX, the Air Force.
I wish they had been better explained to me exactly why SBIRs are so great. It’s not just to do product development alongside the government, but also provides really flexible contract vehicles that allow you to do business, not just with the agency you get the SBIR from, but any other agency in the federal government. And I wish someone had explained that to me a lot sooner, certainly because we’ve had all kinds of much slower situations because of sole source or having to go through primes and we’re a relatively small company and not a priority for a lot of primes yet. So being in control of your own destiny means being in control of your own contract vehicles and SBIRs are a great way for small businesses to do that. So I wish I’d understood that earlier. The other thing that actually is unique to the DoD and does make work with the DoD a little difficult is of course there’s a lot of turnover in program offices and in all kinds of roles. And so for instance, well, I don’t want to name the program office, but in certain program offices, there’s just nobody that was there originally. And very often, unless you have the kind of relationship with the program officer, you’re talking to them on a weekly basis, people move on without saying anything. And then you’re, in some ways starting to rebuild from scratch, even though this customer has been very successful, has reordered many times and has been happy with the product. If the people in the office change, you need to start all over again in terms of relationship building. So that is something unique and complicated about the DoD. But again, once you one have more resources and experienced people on the team is something that’s much easier to manage.
ACME General: You alluded to working with the primes, and I understand there may be a variety of reasons why you were not a priority for them. Do you have a theory as to why?
DP: I don’t want to call myself an expert in the political economy of major prime contractors to the DoD or the US government at large. I really am not. But I will say when I look at any incumbent business or entity, generally it’s just hard to change. Organizations become less flexible, things move a little slower. And yeah, so I’m not going to suggest anything there in particular, but I will say that not all plans are created equal. There are some primes we work with who move quickly, who see the advantage of bringing in nimble novel technology per small organizations and who really do very, very substantial integration work. They’re bringing in like 30 different technologies to provide a common operating picture, situational awareness or some entirely new offering by mixing a variety of commercially available technologies. So that’s impressive, what’s less impressive is when primes aren’t looking to innovate and are satisfied with the status quo.
So we have as many prime partners, current and past, are amazing as others who seem to be just moving at a different pace and focused on different things. But I think ultimately when you get to partner with a prime or even other fellow small business who are really, really engaged with the customer’s mission sets and use cases, great things happen. Truly great things. And I’m just going back to your question about challenges. I wanted to go back to probably our biggest challenge doing business with the government has been around education. So educating them about a totally new product that they haven’t thought to ask for, because it hasn’t existed prior. And specifically in the space that goTenna operates, we think that the government needs to rethink the way we empower hyper enabled operators with situational awareness technology at the last mile.
And it’s always a challenge educating the government on new mesh networking technology that fits into their traditional comstack. And it seems that individual teams are constantly interested in learning and trying new technology like global mesh networks, but getting the government as a broad institution to propagate a technology throughout the entire DoD is challenging. And it’s also conversely been very educational for me to learn how the government does business and just because you have a good product doesn’t mean you’re guaranteed success in a B to G environment. So learning who the end users are and helping empower them through the government requirements and acquisition process is equally important and a lesson that every company is seeking to do business with the government must learn. The government ultimately I think are great customers. And it’s really clear that everyone wants to support frontline operators, whether a Wildland firefighter or a special operations operator.
However, one challenge for a small company like goTenna can be helping to connect the dots. And there’s just so many disparate stakeholders, and they need to align end users with HQ programs of record and of course, procurement professionals. And that can be a tall task for a smaller tech company compared to my major prime contractor. And when it comes to procurement, we need to think about how to speed up the adoption of new technology versus sticking to programs of record, which is what I was trying to allude to before. This usually lends itself to the government sticking with some of the larger companies and not exploring what new emerging companies and providers are developing. Small startups are constantly coming out with new and emerging tech with no or few paths to take it to market without teaming with a big prime or sacrificing some of their IP to one of those big primes.
So I think the government has gotten much better at this thanks to the work of innovation cells, such as DIU, AFWERKS, CDP-invent, In-Q-Tel, etc. A challenge for small businesses is to learn how to springboard from these opportunities and understand how to partner with bigger primes and provide a holistic solution to the government. The government’s busy, they don’t want piece-parts. They want the whole solution, and it’s on us to help them find that. So a big lesson that we learned is, yes, you can sell, “onesies or twosies and onesies or twosies.” It can be hundreds or thousands, but goTenna on its own is not a point solution. It’s part of general systems that at least connects, pairs to a phone. Very often, it’s pairing to a whole other slew of communications technology ranging from satellite to sensors. And we’ve had a lot more success when we can partner with customers and potentially other technology providers to provide the end user a full solution, a system.
ACME General: Well, this has been incredibly illuminating, Daniela. So thank you so much for joining us today.
DP: Great. Thank you.
ACME General: Thanks again to Daeil and Daniela for joining us on this month’s episode of Accelerate Defense. Next month we’re talking to Dr. Melissa Flagg about R&D in the US and efficiency in the DoD.
If you enjoyed today’s episode, please rate and review Accelerate Defense on Apple Podcasts – it really helps other listeners find the show.
And follow the series today wherever you get your podcasts, so you get each episode in your feed when they come out every month.
Accelerate Defense is a monthly podcast from ACME General Corp. Our producer is Isabel Robertson. Audio engineer is Sean Rule-Hoffman. Special thanks to the team at ACME. I’m Ken Harbaugh, and this is Accelerate Defense. Thanks for listening.
Find us in your favorite podcast app.
Find us in your favorite podcast app.