Ep 55: Nathan Mintz on Directed Energy, a Driverless Future, and How He Attracts the Top Engineering Talent

0:00I wouldn't be surprised if Elon makes it to Mars by 2026 or so I that's fast yeah I mean they're going to go really aggressively. And I think once they get off the launch pad, it's going to go very fast. He's a long time to tweet from ours, Nathan.

0:22I got to know Nathan Mintz when we were at Stanford together. He had strong opinions then. He has strong opinions now and he's doing amazing things. The other day I was with a friend who works for Elon Musk. I was asking him where he's seeing the top talent in mechanical engineering and hardware. And he mentioned a couple companies and he said, you know, there's one where there's just amazing people. It's really small though. You probably haven't heard of it. It's called Spartan Radar. Spartan is actually the name of Nathan's company. Let's learn why Nathan's attracting such top talent and how he's pushing forward the possibility frontier in terms of radar and autonomous vehicles.

0:51Nathan Mintz. Nathan, thank you for joining us. Pleased to be here, Joe. Thank you. Nathan and I knew each other at Stanford. We started a nonprofit together called California of Common Sense in 2010, which led to some great things coming out of that, which we figured out. You're the founder of Epris, which we have talked about before on the podcast. You were early CEO there. You're now the founder and CEO of Spartan. Yep. Yeah. It's been a heck of a journey. And I just want to say thank you for talking me into taking this entrepreneurial journey with you. Well, it's great to get to spend time with you in front of everyone.

1:20We are currently in a state that I will not admit that I am in. You are still based here outside of Los Angeles. Yeah. For the time being, yes. I've been visiting for a couple days and getting out before they can get me. Nathan, you've built some amazing teams here. One of my favorite stories about you, as I was catching up with one of Elon Musk's talent people, and they mentioned, of course, Epris, which they knew I was involved with. And I said, where else are you seeing really great talent going? And they mentioned this little new company called Spartan. You probably haven't heard of it, but that's another one on our radar, which is ironic because you've started both of those.

1:50It's a radar company, yeah. It is a radar company. And, Nathan, but first let's go back a little bit. You actually ran for California Assembly in the past. Is that right? Yeah, about 13 years ago, and then I was a glutton for punishment, so I tried it again in 2012. And you got crushed twice. The first time, actually, we thought I was going to win until about 3 a.m., and then Venice reported. Really? So it was closer than I realized. Yeah, yeah. No, I mean, I lost by 7%. That's not too bad. Yeah, and the funny part of it was, is the first five hours of the night, I was winning by about 4 ,500 votes, and then it went to 3 ,000 and 1 ,500.

2:25The suckers in Venice Beach just finished you off. Yeah, and I was getting calls from everybody in Sacramento, I can't believe you pulled this off. And it just turned out it was where they were counting first. So everything south of the airport, I won by like 5 ,000 votes. Everything north of the airport, I lost by 15 ,000. What does an assemblyman make, by the way? How much would you have been paid? A heck of a lot. Probably comparable to a startup CEO. I don't know, about$95 ,000 a year or something. I'm guessing you're making a little more than that. I hope so. Nathan, as a successful entrepreneur, because you would have been making$95 ,000 in Sacramento, probably still failing to get much done, given how Sacramento works.

2:56But what did you learn from that? What did you learn from running? I'll tell you. Raising money in an uphill race was great practice for fundraising in the current environment for VCN startups. Do you learn how to raise money? What else do you learn about politics? Give us some insights on that. Well, one thing I learned is that Sacramento is basically locked into a number of special interests who pretty much call the shots in the state. And the assemblymen, mostly on the majority side, but some on the minority side, show up and they get their voting card. And it's, you're going to vote this way on these bills.

3:29And to a certain extent, a certain amount of dissent is allowed. But if they get too vocal, they'll find themselves kicked off of the right committees. And they won't get the war chest that gives them the influence they want. So you're not an optimist on Sacramento. You're not going to have big changes there soon? I just don't see it. I think that until the voters in the state demand change and are willing to be open minded about who brings that change rather than just sticking to certain political labels of, you know, Democrat or Republican, I think you're going to have real trouble making any reforms.

4:03I'll just throw out one thing that I read on Twitter the other day. The state has spent 20 billion dollars on homeless services in the last five years. Twenty billion dollars. The number of homeless people in the state's gone up 20 percent in that time. Well, it makes sense. You spend more money, you encourage something. But anyway, before we go into the optimism that I want to ask, you did also in your career work for Raytheon, work for Boeing. You did some amazing things in these places. I think you probably had some patents for things you created there. Yeah, I had about six there and then a couple from Epirus along the way.

4:31It was a great place to get my feet wet and learn the ropes. I was a radar systems engineer, and I worked on electronic warfare systems for about 15 years. Got to fly around in a test plane, got to go to all sorts of places that you generally don't see that I still can't talk about. So there were some real positives working there. You learned a lot. A lot of people complain these companies are inefficient. They're doing things cost plus. But there's also some real talent there as well in certain areas. I would say it's a great place to start your career. You spend the first five years there, especially if you don't just stay on one program.

5:02You get into a development environment. You work on a few different things and see a few different disciplines. That frictionless ability to kind of move from program to program in the matrix organizations really helps you build up a multidisciplinary approach, which is great for systems engineers. That's a more positive take than I've heard. There's not too many people I know who are great systems engineers who've went to these companies. You're obviously an exception there. I like to think so, but I've met better. There are busy good people there. Some of them I hired for our companies, actually.

5:29What are some of the structural issues these companies face? Why aren't they solving all the problems that some of the new startups are solving? Well, that's a good question. I think that the biggest issue they face is that they kind of wait around for the government to tell them what to build, rather than going out there and trying to invent new solutions for the problems. And the government does things along a structured approach. They used to call it the Klinger-Cohen approach. Sorry, the Klinger-Cohen approach. Klinger-Cohen was the name of the act that set it out with the DOD architectural framework and various milestone A, milestone B, analysis of alternatives.

6:03all these different contracts that go for 10 or 15 years. And by the time they arrive at a solution, it's obsolete already because they've taken so long thinking about it and sort of gradually incrementally funding it, as opposed to what we did at Epirus and others have done at other startups where they've come in and said, you know what, we have a way to solve this problem. We're going to run as fast as we can in sprints and iterate rapidly to get to what we feel is the right solution. And that requires taking a risk. You're basically building something before they've said they're willing to pay for it because you're so confident that it's just so much better that they will ultimately pay for it and you'll make money.

6:36That's exactly right. And the big companies just don't want to do that. And the stakes are high. If you're a middle manager or a junior executive at one of these companies, and you go blow$50 million on something no customer is going to buy, you're going to get cashiered. And next thing you know, you're going to be doing sustainability support in Alabama or something. Not that there's anything wrong with Alabama. There is something wrong with sustainability support, though, in many cases. Well, I mean, yeah.

7:07I struggle to find examples of punitive assignments. But yeah, that's the underlying problem, is that in big companies like that, you own none of the success, and you pay for all of the downside risk. So, there's really no incentive for you to take a risk, whereas in startups, it's the inverse. No, and it makes sense. At Palantir, we also were spending tens of millions of dollars on something we were very confident in well before there was any suggestion they were going to pay for it. So, let's go to Epris first. I've been lucky to be involved with you in various things. I'm an investor in your current company.

7:40We built Epris together. You were one of the key early guys at the very beginning there, putting it together. How did you put Epris together? Where did that come from? So part of it was I was prompted by an email from you that said, hey, we just brought in a new partner. John, he's interested in starting a defense vertical. Do you have any ideas? And it happened to be around the same time that Gatwick, which is an airport in London, was shut down for 25 hours by drones. And nobody could figure out why. So I was working in business development at Boeing at the time, competitive intel. They used to bring me in to tell them what the other guys were going to bid.

8:16And I looked at it and said, well, you know, what's going on here? Why is this so hard? Like we can shoot down a missile going Mach 7 over the pole, but we can't, you know, find a drone at an airport. And it turns out that our sensors during the Cold War, we were all focused on stuff flying in high and fast, not slow and low, which is where drones come in. Right. And so there had been a study published that showed there were 200 different systems for counter UAS. And 90 percent of them were crappy jammers that somebody built in their garage for like forty thousand dollars and didn't work. I saw one system that did work and it was this huge electromagnetic pulse system that was, you know, the size of a shipping of like multiple shipping containers.

8:56And I said, well, that's perfect because it doesn't care what the drone's doing or what comm channel it's on or whatever. It just fries the electronics. The trouble is it's$25 million. And so it got me thinking. And I said, you know, I bet you with solid state, with gallium nitride, we could get those same field levels. Yeah, because gallium nitride has only been possible to use for 10 or 15 years. And then you and Bo and the team also figured out you can use AI on chips to get the power to hit it faster, which seems to be a big part of it. Exactly. And I called Bo and he said, you know, I had the exact same thought.

9:26And so we set out with how do we do this in a white glove environment and really recreate this in a way where we own it entirely. And, you know, we actually went to Walmart and found a bunch of drones and tested on that. I remember. Yeah. What were some of the initial challenges you didn't expect? Well, let's see. About two months into us getting the company going, we got a notice from our landlord that we were subleasing that we had to move in less than 30 days. You were doing some scary defense things in his building. Yeah, he probably didn't appreciate that. I think we were replaced with people making yoga mats, actually.

9:58Reasonable. Yeah. And, you know, they ran the whole building. We were just running a corner of it. So, we had to find a place in the first two weeks to put the company in. And we ended up, we're sitting there doing all this drone testing in like a co-working space. And it was sort of, had a lot of funny looks from the bank and the... Co-working is maybe not a good place to build new weapons technology. Did you get any trouble from, for example, Raytheon for having had worked there before? I remember they sent you some things. Was it ever a big, scary thing or not? You know, I think we did a good job of kind of keeping our meat and milk separate and making sure that we clearly delineated what was our IP along the way.

10:35And I think, you know, Raytheon, to their credit, they respect that. Right. They respect that, you know, if it's on our dime and we develop it entirely on our own, you know, they don't have a claim on what's in engineers' heads. Right. I mean, our know-how is ours to keep. Right. And, Nathan, what do you think about the future of Epris? You're not running anymore, but you have a lot of really great talent you've hired there. What's it look like? I mean, everything I hear is absolutely amazing and great things. I mean, they brought in, it's kind of funny, they brought in my old boss, Andy Lowery, who Bo and I started working for all the way 13 years ago as COO and chief product officer.

11:09And Andy is really good at driving these things to scale. And his first company, RealWear, actually just did a meaningful SPAC deal, not like a vaporware SPAC like we've seen. I think they're coming out for$800 million on the exit in the augmented reality space. So the way I see it, directed energy weapons are a game changer on the battlefield. They don't have any of these magazine depth problems. The headlines right now are plagued with ammunition shortages and missile shortages and how are we going to replenish all this? The easy solution, don't use ammunition at all. And Epirus is key to kind of breaking that wheel, if you will.

11:45And then I think on top of that, the commercial applications of high power management for telecommunications, for grid infrastructure management, even medical applications open up hundreds of billions of dollars in total available market that otherwise would be unavailable. So, I mean, I'm really honored and thrilled that I got to be part of it at the beginning. And I'm really excited to see it grow over time. I think that the sky's the limit for this company. Agreed. I think EPR has changed the future of warfare, so you should be really proud of it. Yeah, no, it's the most exciting thing I certainly got to work on in my 17 years in defense.

12:21What surprised you, Nathan, about the first time, this is the first time you were starting a company, right? You helped me before on things. We'd worked on stuff together. You'd been in a big company yourself for a while. What surprised you about doing a startup? Was there anything you didn't expect? I think it's the totality of a startup CEO's job in that you touch everything. And there's really no guidebook for the situations that you encounter. A lot of it, you're going to improvise on the fly. You're going to get it wrong. And you've got to be willing to basically take a punch and get up and keep going.

12:51And that took a little bit of adaptability. I mean, I'd been a middle manager before in aerospace. But it's a little bit different from being a middle manager in a process that's clearly defined in your roles known versus CEO of a startup or particularly with eight or ten people what you say goes. What did you take from being a middle manager? What were the useful lessons that you think were brought to this? Understanding how to motivate people, understanding how people are staffed effectively, how to select talent. I had a center director come to me in my first week on the job as a section manager at Raytheon and say, hey, we have this big proposal we're going after.

13:26How are you going to staff it? And just off the top of my head, I go, 36 people from the existing program. We're going to hire six new ones, and we'll pull three out of the rotation program. He says, OK, do it. So I had to learn how to go do hiring and how to really identify who are good engineers and who are bad engineers and who are great engineers. And then finding those great engineers and really putting them in opportunities where they can grow and contribute to the organization and feel valued so they stay around. Got it. So you got up Briscoe and you guys built. You learned a lot. I know you ended up replacing yourself in some different ways.

13:57Moved on to Spartan. Tell us about Spartan Radar. What inspired you to start this? It's an automotive radar company. Yes, yes. So, Spartan, our mission is really to make the world's roads safer, you know, for both cars, trucks, specialty vehicles, all of it, through better radar software, right? I had a, about a few months after I had kind of left, you know, stepped away from EPRS, I got a call from an old mentor of mine, Dr. Theogenes Sabatza-Glue, who taught me almost everything I know about radar signal processing. And he had retired a couple of years before, was working as a consultant for a couple of automotive radar companies.

14:33And he said, Nathan, I think I've come up with a way to build an automotive radar that has five times the resolution with a quarter of the parts. And I go, well, that sounds intriguing. Maybe that's something that we can turn into a product. So I looked at it and said, well, you know, maybe there's more to this. So I started diving into it, and I realized that they were still building automotive radars in a lot of respects the same ways that we had built airborne radars 30, 40 years before. Very single-function devices that do one thing. What do they do? So automotive radar, if my car is going to hit something, it goes beep, beep, beep.

15:05Is that the automotive radar? Yeah, it's either the radar, the cameras, or the ultrasonic, depending on the range and the context. So what else is it doing? What's automotive radar useful for us? It's for lane keeping assist, for automated emergency braking, for really tracking where you are relative to other moving objects in the around. So, it helps with either self-driving in some cases or in other cases, but just braking or warning you about things. What's called ADAS systems, automated driving assistance systems, Joe. So, that's like... And those are pretty common now in a lot of cars. Ubiquitous.

15:34Yeah, it's pretty much in every car. There's probably something like 60 or 70 million radars that go out the door every year. I would say probably 60 % of the cars on the road have one at least in the Ford basically. I'll admit something to you. I'm a distracted driver is one of the reasons I actually hire a driver almost everywhere I go. And actually, it wasn't my choice. My brother and my wife basically forced me to hire a driver. And the first time I was driving, maybe eight or nine years ago, when this ADA thing kicked in, I was grateful for it. I think it's probably saved me twice because it's just like, you know, you're maybe doing something you shouldn't be doing or distracted.

16:03And boom, it breaks for you and saves you. So, it is pretty cool. These things seem to work pretty well. Yeah, so what our software does is it allows the radar that's this big, and I'm pointing to something about half the size of an iPhone, to behave like something that's the size of a large dinner plate. So it's able to basically, if you think about it just intuitively, like telescopes or lenses, radars operate on the same function. So you can have higher resolution. So the proportion would usually be to the square of the size. And so what does that mean in practice? And why does this matter?

16:34Like, why do I care? Are you doing something that's like a millimeter better? Is it like inches better? Why does this matter? So what radar does better than other sensors is it can estimate distance and velocity very accurately, but it has poor angular resolution. Right. It can't tell elevation and azimuth very well. Where this becomes a problem is like, for example, if you're trying to see a motorcyclist that's going by another car and lane splitting or a person that's standing outside of a car door that just closed. Right. These are scenarios where, you know, a car, whether it's warning you or it's actively driving the car, needs to put ample space between you and that other object for situational awareness.

17:11And what's the bigger radar do? It just sees these people much better. It sees them with higher clarity and with it's able to distinguish those objects and separate them better. Further away. Yeah. Further away. In practice, give me some some numbers here. Is this like you're going to see the motorcycle 20 yards away instead of five yards away? Or like what kind of things are we talking about? It depends on the context, but about a four to five times increase in resolution. So you can scale that in terms of distance. So if something normally would have been detected, right, was five yards just passing quickly, you could see it for a while back.

17:41Exactly, yeah. And also the ability to distinguish things that would otherwise kind of like fall into the background, right? That's pretty cool. So at an intersection and stuff, you probably see what's happening much further away and who's turning and who's coming to turn and who's coming through. The other piece of the software stack that we have is the ability to kind of do electronic foveation and to dynamic research. For our listeners who did not do as well in certain areas of college science, what is foveation? So what electronic foveation means is that we can change the radar's focus so that if you're turning left, it can look more to the left and focus and have a higher response rate in the direction of travel.

18:18I knew that, by the way. I'm just asking for our listeners. Of course. So there's that. There's the resource management piece. There's what we call radar modes, which if you think of it, are kind of like apps in your phone, right? So the ability to look at a situation with multiple contexts and pull out new information. We have one mode that we're developing called instant acceleration detection, where in a single look we can tell if somebody's hitting the brakes, which means that automated emergency braking you talked about can kick in sooner. And if you've got a 40-ton semi-truck rolling down the road, a couple hundred milliseconds is the difference between him ending up in your back seat or him stopping just before he hits you.

18:59Interesting. So, you really get to check what's going on in much better ways. It seems like every big car company should want the very best stuff here. So, you're just, I mean, if you become the standard, that's the best, like, is everyone starting to use you? Like, where are you in terms of people adopting this? We have a number of customers that we're working with on both the imaging radar side and on the conventional radar side, as well as OEMs that we're in evaluation with. So, some of our customers include Paccar, Under, Valio, Philips Connect, etc. We're actually working with a number of the trucking customers to try and figure out how we can put more radars around commercial vehicles so that they can have fewer blind spots and more reliably stop before they hit something they're not supposed to.

19:39That's great. And I guess it takes a while to evaluate these things. How's a big car company go from meeting you, evaluating, to using it? This is still a startup. It's like a few-year process, is how you're thinking about it? Yeah. So, the design cycle, one of the reasons when we came up with our business strategy for this company that we leaned into commercial vehicles so heavily at first is that the passenger car design cycle, depending on if you're talking about a Chinese OEM, an American OEM, European, or Japanese, can go from two to ten years even. So, we said, OK, well, we need to make revenue fast.

20:12So, we took away to package our algorithms in our own aftermarket radar and then take that into the trucking market as a product that we can sell, which we call Hoplo. Got it. So, you're selling things in aftermarket, but you're also building things in. So, you can make a lot of money in five years with these groups. We have at least one tier one at this point that has integrated our software into their mule vehicles, and OEMs are currently evaluating it for selection. So, one thing I want to ask is that you mentioned this earlier, one of the very best at finding and hiring world-class hardware engineers.

20:41You've done this now, again, acknowledged by some of the top talent people I know. And we could tell you are the best people as well. How do you find them? How do you distinguish good from great engineers? And how do you get them to work together? So I kind of take an unconventional approach. One thing that I notice is really popular in Silicon Valley is doing lots of brain teasers and stuff. I don't do anything like that. Is that because you're not good at brain teasers? Oh, I'm terrible at brain teasers. You know this. You graded my test when I applied for the job at Querium. You may remember this.

21:09You know, actually, that was the most intimidating interview in my life. That was super obnoxious. Yeah. Well, no, I mean, you sat me down in front of Peter Thiel, and it was my first coat and tie interview. I was scared. But anyway, I would say that, like, what I look for is people who have a multidisciplinary approach, and I look at their body of work. And I say, where was, you know, show me an example where you were faced with a problem that didn't require an in-the-box solution. It was an out-of-the-box solution that was required and how you solve that, right? So like, and that's sometimes hard to find in aerospace.