Ep. 016: Horizontal Scaling of Synthetic Environments - Jason Kennedy

0:00:04 Tom Constable: Hello. Welcome to episode 16 of the War Fighter Training and Simulation Podcast. I'm Tom Constable and this is Colin Hillio.

0:00:10 Colin Hillier: Hello. Hello.

0:00:11 Tom Constable: So I've got apology up front to make to you, Colin. I'm forcing you to record this way in advance of the actual release date. And I think before you mention it, I thought I'd I'd mention it up front. So, yeah. Apologies, Colin.

0:00:23 Colin Hillier: Well, apology accepted, but you know, that's the price you pay for being too organized. You need to own that.

0:00:30 Tom Constable: You say organized. This is pure desperation. Moving house and then followed by Third Child in close succession. I don't know what state I'm going to be in, so I didn't want to risk it.

0:00:39 Colin Hillier: Yeah. So in the podcast land, the time is different, isn't it? And it doesn't quite sync with the real world. But you have a good excuse, so we'll let you off.

0:00:46 Tom Constable: I try. So what are we looking forward to today?

0:00:48 Colin Hillier: This one's been planned for a while and it's really a deep dive into some of the projects that Improbable have been working on. So we have Jason Kennedy, who's their VP of engineering for improbable defense. Jason's got a great background. He's worked for some really credible organizations both on this side of the pond and in the US. I knew him before, I've sort of had a few discussions with him. He knows his stuff and that's what's nice. It's a bit of time to go into some of those details and some of the vision they've got in how they see bringing together sort of disparate pieces of simulation, which, let's be honest, it's not easy.

0:01:21 Colin Hillier: I mean, if it was easy, it had been done by now. And I think Jason has a really interesting perspective on how they're going to solve that problem. Are they looking at things differently? It's not just, oh, well, we'll just magically connect stuff. So a really useful chat, I thought.

0:01:35 Tom Constable: Yeah, absolutely. And what I really liked about this chat was that how open and kind of honest Jason was with his views, the good things and the bad things. And I think that was really came through and I appreciated that. So. Thank you, Jason. Thank you for your time. And should we go into the interview?

0:01:48 Colin Hillier: Yes. Okay. I'm very pleased to introduce Jason Kennedy from Improbable. Jason, how are you?

0:01:55 Jason Kennedy: I'm doing good.

0:01:56 Colin Hillier: This is kind of the counter to the Joe interview. So with Joe we're talking very much about high level concepts and we thought it'd be quite fun to dig into some of the details. So as an engineer, this is quite interesting to me, but I think it has wider appeal for those that might just want to understand some of the concepts and the different approach the team are taking to solving the distributed simulation problem. The title of this is all the subject is going to be about something we've or you've titled horizontal Scaling of synthetic training environments. And I think before we get into it, probably first of all, please give us a bit of your background, where you come from, and then give us a bit of an idea of what this is all about.

0:02:34 Jason Kennedy: Sure. So, just a background of myself. I'm the VP of Engineering here at Improbable today. I started my career actually in Avionics, so engine Control systems. First half of my career, did programs as old as the C 130, then worked on the F 35 Ryan Space Shuttle and a whole slew of others. Then moved into simulations that environments, call them what you will, about ten years ago, really focused on collective training out of the US. And then when I joined Improbable, I continued that journey from synthetic environments, simulations into the UK Mod, as well as NATO. And it's not just training, actually.

0:03:07 Jason Kennedy: We are now looking at synthetic environments or simulations across training, but across decision support, war, gaming, and a whole bunch of other kind of things. Policy analysis. So when we talk about horizontally scaling synthetics yes, absolutely. It still has a foundation in training, but now it's also moving into all.

0:03:24 Colin Hillier: These other spaces in terms of the type of problems we're going to solve. I mean, obviously we're pretty familiar with the defense piece, but from your point of view, you're looking much broader, aren't you?

0:03:34 Jason Kennedy: Yeah. So when you think about the kind of problems we're trying to solve and good, better and different, I think COVID helped really crystallize the idea where simulation can help is when the COVID crisis came along, we had to figure out how this new virus was going to spread. And a lot of people reached into their back pocket for simulations, synthetic environments, to answer that question. However, we didn't really build a community of disciplines and tools and technologies to accelerate that.

0:03:59 Jason Kennedy: So then you had lots of people doing the best they could through predominantly academic units, trying to figure out how we build out these right simulations of the spread of COVID and different policies that we could pull in. And that's what I think we're really looking at, is if you look at governments, defense, national security resilience, you're seeing a more increased demand of I understand what I have through big data and big data analysis, and what has happened over the last ten years with that great. I have tons of data. I understand what is what I need to know is what if I changed a policy, if I changed the defense structure, if I know I have election meddling, what could I do to stop that? And what would those impacts look like? You're seeing that. So yeah, it's not just defense anymore. And honestly, as we all know, defense has changed. It's not just kinetic activity. And because of the move into Gray Zone, into these sub threshold activities, market money manipulation, resource control, all these worlds are blending together. And what you're seeing is government, defense, national security, resilience, all asking for, I need more. What if I need more simulation?

0:05:03 Colin Hillier: And that was something echoing back to Joe's discussion was definitely about how simulation is really good for trying lots of different things very rapidly. You press reset, you restart. Let's try it with a different plan. Yeah, it's very interesting what can be achieved and how that complexity is growing. But in terms of that complexity, bear in mind what tools we have today. What are the sort of things you're developing and what do you think is needed to get that complexity?

0:05:26 Jason Kennedy: Great question. I think this gets us back to horizontal scaling of synthetic environments and synthetic solutions. Right now, to make simulations is expensive. It's slow. They are used often, but they can take months, two years to create, and then they are very hard to sustain. So what you end up with is, I've spent a pretty significant investment to make a simulation. I use it for a few questions, then I put it on a shelf, and that's it sits there.

0:05:51 Jason Kennedy: As a matter of fact, the COVID model is a great example. That's not the first time we've ever modeled the spread of a virus. But every time we spend all this time to model it, then we put it on a shelf, can ever get to it, it's gone. And this is what horizontal scale is about. It's providing tools and technologies that allow you to discover what content is available, to deliver it quickly, to integrate it together with other things.

0:06:12 Jason Kennedy: It's really helping a shift from a pipeline business model, which are very vertical delivery. I make the solution, I make the content, I make the technologies. I deliver it all together to a horizontal delivery or a platform business model where you're looking at, okay, somebody specializes in building out the tools and technologies. Other people specialize in building out content, models, solutions. Another group specializes in building those solutions that consume those things and provide those to users. And then all of a sudden, you have an ecosystem that you can deliver rapidly. You understand what you have. You can discover it, you can find it, you can reuse it. And I think that's the fundamental issue we find ourselves in is we've been using simulation for 40 years. I mean, you go back to the original flight simulator, which is this big blue box that even have a visual system. It's not that we don't know simulation is important and useful. It's that it's expensive.

0:07:01 Jason Kennedy: And it's traditionally sat in things like training. Because a vehicle can take so long to build. It's going to sit out there in the environment for so many years. So it was worth the big investment of months or years to build the simulation because you knew the life of that vehicle or weapon system was going to be around for ten or 20 years. The C 130 is still flying right now, we're trying to bring those same kind of useful what ifs, those simulations, those techniques into rapidly expanding worlds. We're still trying to deliver it in a pipeline delivery methodology, and that's the real friction point. And that's what we're trying to deliver. And to do this right, what we.

0:07:36 Colin Hillier: Have to shift and Defense is really good at buying single thing to do a single job, because you can write requirements for that. It's very clear, funding is clear. But what you're asking there in terms of procuring platforms is almost like, okay, we're going to buy a thing, but we don't know all the things we're going to need it for yet. And I guess the question is, how are those procurement or processes changing that allow this thing to be bought in the first place?

0:08:04 Jason Kennedy: Yeah, it's really interesting because I think this is at the root of the problem is if we want to buy an aircraft carrier, we have the processes and mechanisms to do that. We know how to do that. And Defense has been doing this honestly very successful for a very long time. I think sometimes the procurement arm of Defense gets a lot of negative press. But actually, if you want to build incredibly complex manufactured good, they're probably the best of the world at doing that. It's when you move into the software side that they I think we really struggle. And we've tried doing things like the Trex program in the US.

0:08:34 Jason Kennedy: With OTAs other transaction authorities. Now you see we're replicating that in the UK with the Defense and Security accelerator dasa but I think it's kind of missing the point. What they're trying to do is, oh, the problem is we have to procure faster. But actually what you need to procure is different. What you need to procure is the enablers, and then you need to contract on top of the enablers. So the way Defense has currently looked at doing it is we'll just shorten the procurement cycle or allow for more people to enter the procurement cycle, shorten the proposal process.

0:09:04 Jason Kennedy: And although that's an improvement over what could be a one to two year process, it's still procuring the wrong thing. You're still procuring siloed solutions. And what you really need to be doing is procuring enabling technologies, platforms, and then opening up an ecosystem, then changing the future procurements around that ecosystem. You see some examples of that. So the Tempest program, Selborne, Nelson are kind of we're going to build consortiums and we're going to think about not just buying the thing, but how we're going to buy the thing, and maybe some of the underlying elements.

0:09:34 Jason Kennedy: You see some programs in the US platform One, for example, where they're like, we're going to build the enabling technology for deployment. And then in the UK through Digital Foundry and the DCEP program, they're talking about, okay, now they are thinking about buying or procuring the services and the tools and technologies, but then you still run into the but now you have to procure differently on top of that, and you have to start mandating those things. So you see some shifts, which is really exciting. This is the first time in my 20 year career predominantly focused on defense. I've seen these actual shifts in the way we're thinking about procurement, the way we're thinking about delivering software. But it's still the minority. It's not the majority for sure. And then we're struggling to get adoption and lift because it is the first time defense is really going into that.

0:10:16 Colin Hillier: Freddy, within that, I guess there's an area where standards are important and we have existing bits of software, so we don't want to throw away what we already have. How do we reuse those?

0:10:27 Jason Kennedy: Yeah, I fully agree with you. Right. And I think this is some of the challenge is we have lots of standards, SHLA siggy. We have groups that are in charge of standards. CISO, Dmask and others. Dmask is now DMSO we have actually really good technologies, once SVR forces VBS four, or some really good technologies that have shown good value. The trick is, how do you shift the delivery methodology and how do you move toward a different business model without throwing out the baby with the bathwater?

0:10:56 Jason Kennedy: And I think this is part of the challenge is even sometimes we've had frictions. I've done some presentations or discussions with CISO, and it almost feels like no. To move to a platform business model or horizontal delivery means you have to not have Federated models or standards anymore. It's like, no, we're not saying either or. We're saying that standards, open architectures, the existing softwares that we have, have helped quite a bit, and it has moved the industry. But they've also existed for well over 20 years in most cases, open architectures. And the concepts of that in software go back to the 60s. Dis, the original version of Dis. It predates what we would consider Internet 1.0, like the modern Internet. These have been around for a long time and they haven't gotten us to the promised land, so it's not get rid of them, and it's not that they're not part of the solution. I believe as an industry we have to accept the fact that they are insufficient by themselves.

0:11:44 Jason Kennedy: And what we need is more.

0:11:46 Colin Hillier: I think it'd be useful for some of our listeners if you can go into some of the examples of what isn't working, because the standards do improve, they do try and keep up, but it would be useful just to cover off some of the tangibles of where we think the problem is, and then we can talk about your view of how that gets fixed.

0:12:04 Jason Kennedy: Yeah, I think that's a great kind of transition point to get into, maybe a bit more specific. So if you think about standards and open architectures as part of the solution, which they certainly are. They're still supporting a siloed delivery methodology. So the standards, you might mandate a standard as a requirement on a procurement, but you're still procuring a solution top to bottom. So if I'm going to go out and buy a new, what we kind of generically call within a probable operation Decision support tool, but you might call it a C four C, five ISR Tool, Jadsey, Tool, pick whatever your great name of the moment is for these kind of operational decision support tools, when you buy it, you're still buying the solution. You're buying all of the underpinning technologies, you're buying the service, you're buying all the modeling and content as one big procurement. So although you might put in a requirement of using a standard to support, quote unquote, interoperability, you still procured everything wholesale. So now if somebody else from a different agency was to procure something, they're rebuying all of that. And at the base of your simulations, at the base of your support tools, there's so much technology and so much tooling that all of these solutions are using, but you're buying them over and over and over again. Now, the standard might help those things communicate better, but the standard hasn't gotten rid of the fact that in every one of those procurements, you're rebuying technologies, you're re paying for integration of engineering of those technologies, and then you're paying for some common services.

0:13:28 Jason Kennedy: Your security, your deployment, your DevOps all of that stuff that could be common, you're repaying for. And I think that's the ultimate problem is standards do help interoperability, full stop. We know that, we've been doing it for a long time. However, what they haven't gotten rid of is all the duplicative spend across the piece. What they haven't done is allow an estate across an entire defense organization to recognize what content is available. The discoverability of content.

0:13:54 Jason Kennedy: Now, defense has tried to come after that from different angles. And this is where you have different frameworks or different marketplaces. They haven't really latched on. And I think part of that comes into tying an incentive with motion. So how does the commercial industry do this? So commercial industry has been doing this very well for a long time. If you were to look at, let's say, the Unity marketplace or the Epic marketplace, that's from the games world. So I can go to the Unity marketplace and I can find a whole bunch of content, everything from AI to models, to 3D assets, to terrains, to weather, a whole bunch of stuff. And people make that and they're incentivized to make it really good, because if they make it really good, I'll pay them for it in order to lower the cost or increase the value of the game I'm making. And Unity wants me to find all that content because if I find that content, it's pre integrated with their development technologies.

0:14:43 Jason Kennedy: They get me to pay for the development technology, the content provider gets me to pay for their content. And both of those folks have lowered my cost to entry so I can increase my profits. And I think when you look at the way defense is set up because they are still buying through silos, they're not really incentivizing the right behaviors. If you want more commonality of use, if you want more collaborative behavior, you have to incentivize that. If I do collaborate, I make more profit. That's my job, my job for a business. And I think sometimes we hide away from this is I am a business, I believe in supporting defense. I've made a career of it. I'm going to retire out of defense and supporting defense.

0:15:19 Jason Kennedy: So I do care deeply about the industry, but I do work for a business. At the end of the day, I have to make EBITDA profit and deliver that to my organization. And I have to chase what incentivizes that solution. And in the commercial world, they've incentivized that through having platform delivery where everybody gets to share in profit, everybody gets to share in revenue. And if we work together, we all make more profit together than we do apart. And standards don't do that. This is really about a business model. And that's why when I talk about platform business model, I'm not talking about platforms to technology, I'm talking about the business model shift that is empowered by technology.

0:15:56 Jason Kennedy: But that is the essential shift that has to happen in defense. And that's where the procurement methodologies open standards and these requirements just aren't tackling the real problem.

0:16:04 Colin Hillier: I think you make a good argument for the platform approach and it's not hard to understand because there are good examples out there in the real world, as we say, of platforms that work and provide value and marketplace and ecosystems. But I think the thing that would probably shock defense when they start to think about it is the cost and the effort to make these platforms successful. Because for every successful platform there are several that haven't worked. And we're going to work well, takes money and time and skill.

0:16:33 Jason Kennedy: Yes. So that's really interesting because this goes back into the idea of what, let's say fair procurement or fair bid practices or the procurement agency as a wholesale is and improbable has experienced some of this ourselves. So we've been on a multi year journey now with the UK in supporting a synthetic environment platform. And we've spent a considerable amount of funds to get ourselves into a position where our synthetic environment platform, skyro can deliver tools and technologies across a very wide range of use cases and deliver these kind of complex synthetics at a significant reduced cost.

0:17:05 Jason Kennedy: However, that cost a lot of investment to put in place. Now, if we want to go then pivot and go try to use that, eventually we'll come up against a contracting organization that then wants to control our profit and they will come in and say, well, if we single source this to you, you can only make eight to 12% profit. It's like, yes, I appreciate that. However, we have spent tens of millions in R and D cost over here, and some of that's allowable, some of it's not allowable. Then we start arguing about IP, and again, it kind of goes back into that incentive model. If me as a company have taken a risk of pre investing a lot of money and you hear this a lot, you hear this a lot in the US. As well, where they keep saying they want industry to invest.

0:17:42 Jason Kennedy: They don't want to pay for innovation. They want industry to invest in innovation. You can't have it both ways. If you want industry to invest in innovation, you're going to have winners and losers. And as you said, to make a platform, you're going to have four or five, six people make a platform. And if you look at other platform ecosystems, really only one or two platforms ever survive because there's a whole network effect that has to occur where the ecosystem adapts to a certain business model, and then that's the survivor. Everybody else is lost. But the ROI on that, from a business perspective, if I do win, I now get to reap rewards of additional profit or high gross margin on what I delivered. Now I have to keep my ecosystem happy, so I have to constantly reinvest and do these other things, but I do get to reap a higher return for the business. Right now, we're saying, hey, companies, go out and preinvest take the risk, you all won't survive. But then I want to control your ability to make profit on the other side of it, and it makes the wrong kind of tension. And I think this is probably one of the sticking points on this kind of shift over to if you want to incentivize pre investment, then you have to allow for some level of exploitation of additional profit on the other side. Not exploitation in a bad way, but in a way of I have to make the books balance if I'm going to have five big bets out there, the one or two that make it through.

0:18:57 Jason Kennedy: I have to be able to then recover the R and D of the failing programs, as well as put positive contributive margin back into the business.

0:19:05 Colin Hillier: I'm recalling that Joe made a pretty convincing case that that investment needed to be from private money because the private industry is going to have the good ideas and be able to make something work. And as you say, then the challenge is, how do you get your return on investment? Which I agree. I could think of a few examples where you then provide a conundrum for the procurement organization, well, this isn't fair anymore because you have an advantage. Like, yeah, no, we invested in that to make give ourselves an advantage. That was the point.

0:19:32 Jason Kennedy: Yeah, but it's so interesting because that's at the root of it is the procurement organization is trying to do what they believe is best. They are trying to make sure the taxpayer and the defense organization is getting the best deal and quote unquote, we're being fair and we're not exploiting relationships. And you see this all the time around this idea of monopolies. They don't want to be held to a particular vendor. They don't want to provide a monopoly of a technology.

0:19:54 Jason Kennedy: And again, I think this is where this becomes a challenge for defense, is platforms, by the nature of how platforms deliver value, do create some level of vendor lock in. If you think about these other industries that have platform delivery methodology. So let's go to smartphones. You're either Android or your iOS. That's pretty much it. Microsoft tried for a little bit and they failed. So if you look across and everyone in the world has a cell phone or practically at this point, you really only have two ways of delivering solutions in those environments. So we have an inherent vendor lock in across the smartphone industry. Now, as an end user, do you actually care?

0:20:31 Jason Kennedy: You really don't. Matter of fact, the fact that we have that vendor lock in across those two entities means whole companies could optimize the way they deliver end user value around those motions and they can deliver better value faster to end users. However, it does mean that those two entities take a skim, take 20 30% of all the revenue coming through that. Now they use a big portion of that to reinvest back in the technology to make it even better so more people use it and such and so forth. And that's called the network effect of what a platform business model is. But in the defense space, that's seen as a very bad thing.

0:21:07 Jason Kennedy: And I hear this all the time, but if I use Skyro, I'm going to be vendor locked into your solution. Yes, to a degree. Now, if we stop performing completely, you can take an instance of it and then you can replace this in the future or you can deprecate it out. But why would you want to? I'm incentivized to ensure you're always getting a good product and you're not vendor locked into me for the solution because I'm the back end technology.

0:21:30 Jason Kennedy: I'm not actually delivering your solution. And I think this is where we get stuck. Is this looking at your technology or your solution? Stack in sections and recognizing that platform delivery models push the lock in into the common technologies, but not the stuff that users care about. So again, going back to your phone analogy, when you use an application, you don't actually care about the developer tools or any runtimes or other things that are involved with the application. You care about what the application is providing to you, whether it be enjoyment, entertainment, fitness, that's the value and you're not locked in there. Matter of fact, platforms give you more choice of end solution than you've ever had. And this is where we have to start helping defense recognize that you don't want vendor lock in on solutions, understood. And you want that to be competitive.

0:22:13 Jason Kennedy: But some level of vendor lock in at lower technologies isn't a bad thing and you already have it with cloud. So when we move over to Cloud, there's only three major cloud providers in the world at the moment and the chances are you're already and most defense, if not all defenses already using one of those or all three of those technologies. And they've accepted that level of vendor lock in. And it's interesting because it seems like defense is more comfortable accepting vendor lock in on tangible hardware kind of things, or data centers, or even see this in particular radio or other kind of very niche areas. They accept vendor lock in at those areas. But in the software space, we seem to get very nervous and I do think it's based off of bad industry behaviors.

0:22:54 Jason Kennedy: You talked about simulation and training. If you go back to the original image generators, the original IGS for Flight sims yeah, the industry took some advantage there. They would lock you in for an exorbitant amount of license for that IG, and if you ever left, they would pull the IG so you can never leave them. And then they would use that to bootstrap the rest of the solution and services, essentially making them an immovable incumbent. So I'm not saying we have to forget the past, but I am saying we have to kind of redefine what we mean by vendor lock in, redefine what we think a monopoly or an unfair bid position would be. If we're really going to shift the dynamic into this rapid delivery of solution that we see in other markets, I.

0:23:33 Colin Hillier: Think we can think of lots of examples that involve infrastructure in our daily lives, like the electrical grid, railways, where you have an infrastructure monopoly effectively. Do you choose to effectively nationalize that bit or is it a pseudo monopoly? I think with smartphone platforms, we're probably lucky we have a duopoly, but there can't be three. Doesn't work. Just economically, it doesn't work. We're probably lucky we have two because we could just have one. There's people that will only develop for one of the platforms for whatever their reason is, could be either. Or we could let the excellent get in the way of the good, couldn't we, by saying, well, unless we have a perfect ecosystem where no one has an advantage, we're not doing it. Well, you might have to live with the fact that, yeah, you have one or two providers that have that advantage because they've got good technology, but it evens out the rest of the market and we're just going to have to manage our way through that. Slightly imperfect world.

0:24:25 Jason Kennedy: I think you're spot on because you are right, it is imperfect. Right. Obviously the perfect position for defense would be able to rapidly interchange any technology they want at any moment without any vendor lock in. And I believe this is what we've been trying to get at with open architectures and with standards is this idea of the whole solution is federated and this way I can just hot swap everything in and out. We just know that's not how it works.

0:24:46 Jason Kennedy: Even if you have consistent APIs or consistent interface points, that's not where the hard work is great. You wire up all these different technologies. Now you spend weeks, months, up to years doing all the hard integration work of making all those things actually work the way they're designed now. So again, I think the perfect world is what we keep aiming for, but we're giving up the good world we can have today. And I just don't think the perfect world is achievable. And part of it goes back to and it's funny that you mentioned infrastructure. I really believe some of this goes back to the industry behaviors that I've already mentioned and there were some very predatory industry behaviors over the years.

0:25:20 Jason Kennedy: Cost plus six fee is another one of those contracting mechanisms. That really bad industry behaviors made. That a very unlikely contracting structure of today. Some of this is we have taken advantage as industry of defense and of governments in the past. But I think some of this goes into what people think monopoly is or vendor lock in is. I think people go back to the railroad barons and they're picturing almost a comic strip of the fat guy with like a monocle just raking in money and shoveling in money because nobody has any choice. Yeah, but I think the big change here and what we have to realize is we're talking about everyone's using the same rail. We're putting the vendor lock in to the rail, not the train service itself.

0:26:00 Jason Kennedy: Now you have choice. You can pick any train service you want and hopefully they provide you a really good train service. We can't have every single person running their own rail across the country. It's just not going to work. It sounds a little ridiculous, but that's what we're doing today. Every single time we buy a new software solution in defense, we're building all of it top to bottom or procuring for all of it top to bottom. And sometimes it's offset through reuse of some software technologies such as a CGF, computer generated forces like a VBS four, which is really helpful, that saves you a lot of money. But then we're paying for all the integration services again and the deployment services.

0:26:35 Jason Kennedy: So it's just inevitable that you get back to the place that we're procuring over and over and over and over again for everybody to run rail between two cities. And then we're asking, well, we just need to make this cheaper and we have to make this one single rail. All right? If you want it to be a singular rail, you have to accept one person is going to run that and maintain that. And that's okay, because that's not what you really want. What you want is the train service and then that you can compete.

0:26:58 Colin Hillier: Coming from the UK and the United States, we've probably picked a bad analogy with trains.

0:27:03 Jason Kennedy: Yeah, we probably got a lot of people fired up.

0:27:07 Colin Hillier: But no, your point is valid. You might actually pay a bit more for that infrastructure, that standardization that you run the thing on in order to get the value out. And I think that's the devil's choice, isn't it? You're going to have to invest a bit more. And you may not keep this analogy going, you may not know how many trains are going to run on that track, but the track has to be built whether it's one or ten. It's pretty much the same thing.

0:27:29 Jason Kennedy: Yeah, but then the question comes in is how do you incentivize industry to want to pre invest, to run across that track? And I think this is what the UK is trying to do with their digital foundry and their DCEP programs is they're pretty much saying, hey, we are going to set up some commonality of deployment technologies through D two S, some commonality of synthetic environment solutions through the Diesel program, synthetic environment platforms. Now we are going to want people to come in and use those as a foundation to deliver solutions. If it's done right, it should work if the incentivization is there. If I use D two S, I get accreditation, continuous accreditation or secure by design solutions. So my accreditation cost come down and I'm using a bunch of technologies in a proven process, so that saves me money. So when I deliver a solution, I should see an increased profit and that incentivizes me to use that solution because I have to change as a business.

0:28:21 Jason Kennedy: If I use a synthetic environment platform out of the DCEP program, that should save me some money and I should be able to then see that in some kind of increased profit or decreased risk. And I think this is the challenge we run into is that program works when you incentivize people to come carrot and stick analogy. If I give you a carrot and say, hey, if you use these technologies, it's going to reduce your costs, it's going to reduce your risk and you should recognize that increased profit margins or gross margin, that's a big incentive. I imagine you would see a lot more companies pick it up. And if you look into the commercial world, they do pick it up. In commercial worlds, people have adapted to platform delivery methodologies very quickly. And it's because if they do and if they make a good product, they will see a return in defense. It's kind of interesting because we monitor or measure how much you can really make because we're trying to make a fair environment. It's hard to incentivize some of those behaviors. On the flip side, if we mandate it through a requirement, you must use Dis, you must use HLA, you must use VBS four. Then what you have is this alternative behavior of, okay, how do I minimally tick the box to say I've met the requirement, but not really buy into what the end solution is going to be because all the extra investment I'm not going to get to see as a return anyway.

0:29:36 Jason Kennedy: And it's a real challenge because I'm not saying we should be looking to procurement organizations to not monitor or measure those items because that's obviously going to lead us back into the same predatory behaviors that we saw in the past that caused a lot of those policy changes. I know I'm not giving an answer here, but the alignment of incentives to use common platform technologies to deliver different business model, a platform business model so we can horizontally scale delirious simulations is the solution to reduce cost and deliver faster. We've seen it in too many markets to believe it's not the case, but that doesn't leave us with a pure answer. That's technology only. We have to figure out how this other stuff works together.

0:30:15 Tom Constable: So you were talking about the potential increased profits for these companies that are providing those end solutions. Obviously, the flip side to that is if there's more people who are on the platform which will drive increased competition, bars to entry will be lower due to this increased interoperability between standards. Is there a risk it becomes more of a price war for the end solutions? Because they've got to differentiate themselves somehow. And currently the way that we procure is very much based on do you have a model of a tank? A challenger two?

0:30:44 Jason Kennedy: Yes.

0:30:44 Tom Constable: No. Well, actually, not all challenger two tanks are created equal. But that's unfortunately the way that the way that we currently procure in the UK. It's a tick box.

0:30:52 Jason Kennedy: Yes.

0:30:52 Tom Constable: No. So how do you differentiate yourself? And one of the ways is of course, on price. So is there a risk of that and is there a way we can mitigate that?

0:30:59 Jason Kennedy: So there's definitely a risk of it. Let's not avoid the fact that if you have a platform solution in place, you're expanding the people that compete in the market. And that's by design. So by design is some of the hardest technologies to build and some of the hardest tooling to build. Hopefully, you're pushing into your platform and making them common commodity across the ecosystem. That's how you maximize savings to your ecosystem and your developers. It's also how you maximize savings to your end user. But yes, that does mean a nontraditional can enter the space and all of a sudden compete with a traditional who has spent ten or 15 years making some of their own common solutions. They often reuse, but now they charge for which they've earned the right to do. By the way. Now that you have to look at a couple of different ways.

0:31:45 Jason Kennedy: One, it's incentivizing a good behavior across NATO, across five eyes, across pretty much all defense organizations. They're talking about wanting to bring in innovation, bring in Silicon Valley, bring in nontraditionals. This is the way to do it. Now if you are a traditional that's bringing lots of people into my space and eating my rice bowl, I'm not a big fan of that. However, you can shift your monetization.

0:32:04 Jason Kennedy: So now what you can do is you can actually make your technologies and make your content, your models, your solutions, stuff that you're using to make your solutions available to that same community. So although, yes, I know it raises competition on the end solution that does force you to deliver more value and it also lowers the boundary for entry for everybody. It does give traditional companies, companies have been in defense for a very long time a different way to monetize. And the nice part is that monetization comes with a higher gross margin. So rather than go after a technically acceptable lowest cost, a TELK kind of win, you can now make your content available through the ecosystem or your tooling or technologies available through the platform.

0:32:43 Jason Kennedy: You can charge a recurring fee on that or a one time fee on that, but it's something you've already built. Now you have a new way to monetize that and you don't necessarily have to make all your money in the end solution competition, you can now make your money through other avenues. So yes, it definitely makes more competition there and it reduces the total cost there, which is great for the government, great for defense. It will cause industries to change the way they monetize and that's going to be a bit of a friction point. And this is where we have to make sure that they are incentivized to do it. If it's just through the stick, you're going to have people fight against it. If they are incentivized to do that and if they do reprioritize some of their technologies, your content, to be able to be delivered through a platform, they have to be able to then reap the rewards of increased profits on that pass through. So it is the shift in market, but you saw that shift happen in a lot of different markets to date. So every market that went from a traditional pipeline delivery methodology into a platform delivery methodology had to go through that exact same growing pain. And you had survivors and you had people who didn't survive. Nokia is a great example. They've owned most of the phone industry. They didn't adapt, they disappeared.

0:33:48 Jason Kennedy: Other people did adapt and actually reaped lots of rewards from it. But if we do make this kind of market shift of software delivery and we're successful with it. You will see companies that adapt and thrive and you'll probably see companies that don't adapt and would struggle with that. Now, the benefit we have in defense is this is a great business model and using platforms to deliver software solutions is a great concept. It works. We know it works. It reduced cost.

0:34:13 Jason Kennedy: We still have manufactured good and manufactured good is still a pipeline delivery methodology, although we're getting smarter with just in time delivery and all this other kind of stuff, it's still best done in a pipeline business model. And those traditional companies that work in that space, traditional defense companies that work in that space, that's still going to be delivered that way, there's still plenty of ways for them to monetize in that area. So we just have to separate the.

0:34:37 Colin Hillier: Two areas just to pick on something. You said just then a few minutes back about maybe governments and militaries looking at Silicon Valley and looking at all the great things they can do and the things like open air AI. Two things spring to mind. One is that their eyes might be bigger than their belly in terms of the amount of money it takes to do these things because there's some amazing technologies, but defense doesn't necessarily have that level of investment. I just read this morning that the Facebook so far spent 24 billion on their metadata.

0:35:06 Colin Hillier: I've just tried to sort of understand, I think that's the GDP of a small Latin American country, isn't it? It's incredible sums of money. So there's that piece. But also Silicon Valley and the like have a very difficult relationship with defense. One day there could be it's all about the good that it can bring and then they can turn on the sixpence and say, well, we're not doing that. And I think, Jason, you'll know, of examples where internally there's people that are just not willing to work on military projects or it's not seen as important to their bottom line one day. So they turn around and go do something totally different. So there seems like there needs to be a place for the Babelfish, the organizations that can interpolate and interpret taking the right technologies, understanding them and going these things.

0:35:50 Colin Hillier: We can use these other things. Not necessarily for you, but how about this? Was one of your thoughts on that?

0:35:55 Jason Kennedy: Yeah, so it's interesting, the Silicon Valley, quote unquote, silicon Valley relationship with defense and defense's relationship with Silicon Valley in the opposite direction. There's constant talk about bringing innovation into defense. And you are right. It almost seems like every two or three years or maybe three to four years, defense has this big surge of I'm going to take a lot of money. I'm going to put it into innovation companies.

0:36:16 Jason Kennedy: And then you have a whole bunch of people from these nontraditional Silicon Valley go after that kind of money. And then the relationship never seems to solidify. It's really challenging. And part of it goes into what you were talking about. If you have a company that is designed to do AI for cat pictures on the Internet and it finds the best cat memes that's ever existed and now you're going to pivot that technology into finding tanks behind enemy lines, you're probably going to have employees that did not sign up for that. So some of this is if you're making something in one market and going to another market and a lot of people do see defense as a high risk or perhaps a confrontational market. Now, obviously I don't I think it's a necessity, but some people do and that's okay. Everybody has to be able to live the way they want to live. So you have that part, then you have the raw technology part. If I'm from Silicon Valley, if I'm from a commercial market, if you ask me to invest up front, I'm expecting to exploit that investment to reap some kind of reward or profit. And the second I have to go into a competitive procurement process, I could just get turned off. Wait a second. You worked with me for the first three years investing in something, and now you're telling me I have to go to a competitive process and I don't even know how to compete for a defense contract. I don't understand scoring or anything like that. So one, I might be turned off and just walk away, or two, I might lose because I don't know how to write a defense proposal. And there's people who are in from the defense industry that make careers out of this, right? So I might just trade lose. So you have these real challenges. But I think you're right. There's space for companies that can straddle that line because you want to bring that $24 billion of investment in the Metaverse into defense.

0:37:50 Jason Kennedy: Defense doesn't want to spend that much in that investment, but you have to have almost a translator. And this is where Probable sits today. Probable has made its business because we sit on the Metaverse and the game side and this game technology kind of side of the world, but we sit on the defense side of the world. And we made a division that just does that because our broader division across Metaverse probably wouldn't want to work in the defense space. So we made a division that all they do is sit in the defense world and they help bring the best technologies and services from the commercial world, Auburn to defense. They help mingle in that money. So the money that's invested in these commercial technologies are now a starting position for defense as opposed to them having to bring it themselves.

0:38:31 Jason Kennedy: And we're not the only company that's doing this. You're seeing more and more of it if you go back in time. That's how Bohemian Interactives or Simulation started with EBS that started off as an entertainment product. They moved into defense and they stood up a whole company around it. So it's a proven model. But if you're really going to exploit the best innovative technologies, you either need to have companies that are willing to straddle that fence or you're going to need to really invest into a platform solution to the point where I can be a part time player. I can be in the commercial world, but then I can have something that's really great a piece of technology, a tool or a piece of content that I can put into a platform with pretty minimal investment and then be able to see that get a return.

0:39:13 Jason Kennedy: And we're seeing that with some of our academic units. So we work with a lot of academic universities throughout the UK. And the ability for them to be very scientific, go after their doctorates or their doctorate thesis or their own internal research to build out these complex models of civilian behaviors or interactions, but then be able to make that readily available through a platform is pretty great. They don't have to get deep into defense, they don't have to be there, but that allows them a way to monetize that, to bring more money back into the university that they can then reinvest in innovation and the stuff that they really like to do. So I think there's a couple of different ways to do it and you see all of those in the market, but we have to be able to support those and continue to do that. If you want to keep pulling that outside innovation, that outside investment into defense, which I do think we want as a common notion, I'll leave it there.

0:39:57 Colin Hillier: I think that's a really important point and I don't want to water it down. So jason, thank you. That really puts a lot of context to the previous interview do with Joe, so thank you and depth to what it means for the technology. So I really enjoyed that discussion.

0:40:10 Jason Kennedy: Yeah, thank you very much. I know it's interesting to talk about technologies that are really back of the stack. These enabling technologies are not really the showstoppers, but it's through those enabling technologies that we can give showstopping solutions. So it was really great to talk about it and I'm excited where the industry is, I'm excited where the industry is going and I see the shift happening. I really believe if cross industry, government, defense, academia, if we can all lean in together, we can make the shift happen and get to a better spot for tomorrow.

0:40:41 Tom Constable: I just really enjoy talking to guys from Improbable because they just bring a completely different angle to this and a different conversation topic and a way of seeing the challenges within defense. And it is just always nice to hear and be open minded to new ideas to add that to your kind of repertoire of understanding and knowledge.

0:40:57 Colin Hillier: As I say I've known a few people, very credible people that work in probable and yeah, it's never dull having a deep conversation in terms of how you solve these problems.

0:41:07 Tom Constable: Brilliant. I think now is the time for our intrepid journalist Andy Fawkes from Military Simulation Training. Here he is.

0:41:16 Andy Fawkes: Hello, great to be back. There's been no shortage of news you'll be pleased to hear. So we'll crack on with that, shall we?

0:41:23 Tom Constable: Yeah, let's go.

0:41:23 Andy Fawkes: So in the Canada coming up, is itech? Which is in Rotterdam this year, so looking forward to that. And there's an article in Ms and T saying there should be around 2000 visitors and more than 75 exhibitors. I've certainly looked at the exhibitor list myself and I'll be there myself. I'm looking forward to doing a little talk. Should be good.

0:41:40 Tom Constable: What's your talk?

0:41:41 Andy Fawkes: I'm calling it from SIMNET to the Metaverse. Why is it taking so long? And it's about kind of multi domain networked training with simulation. They were doing it in 1990 and we're still talking about it all these years later. So come and come and have a listen. Awesome part of a panel and just.

0:41:59 Tom Constable: Tell me the date again.

0:42:00 Colin Hillier: 24Th to the 26 April.

0:42:02 Tom Constable: And I think if anyone else is, I think Colin, you're going to be there as well.

0:42:05 Andy Fawkes: I will, looking forward to that. And it'll be great to get a European dimension. It looks busy.

0:42:10 Tom Constable: I've had a flavor of these news articles and I'm really looking forward to discussing it. So what's the first one, Andy?

0:42:14 Andy Fawkes: So I think it's essential reading if you're in military training is please read Dim Jones's article in Ms and T and it's called Ref Pilot Pipeline not Keeping Pace. Dim was an RAF pilot and instructor and he's written for Ms and T for quite a while. So I would say he's certainly an authoritative figure in this discussion. So I've heard talks myself from various people that there are problems in the pipeline that basically is just taking longer to get people through the pipeline. And of course all these things are complicated, but there are issues like instructor shortages, poor availability of some aircraft, training aircraft. At the moment it's the Hawk and also changing mod requirements.

0:42:55 Andy Fawkes: So please have a read of that. But I think the story is obviously this is about the Ref World Air Force, although it is about training army and Navy pilots as well, but it's about how far you can sensibly contract out training. For those who don't know, this started late 90s in the efforts to sow private financial initiatives to essentially contract the buying of aircraft and then that moved to much higher levels into the actual people training the people who were training, not just the aircraft and buying simulators and the like. So in a sense, it was a bold effort to try and contract virtually all training out. But they claimed time.

0:43:33 Andy Fawkes: It was a partnership, military flight training system. Now, it'd be good, obviously, to get in the future, another point, someone to discuss that, who's actually involved day to day in that. But it does seem from the outside that it's taken an awful long time to get right. And even now there are huge problems.

0:43:51 Colin Hillier: Yeah, 20 years ago there were a lot of holdovers. It seems to have got worse from what I hear and I'm conscious I'm not near enough at all to it to comment but it's one of these issues that doesn't go away and I think Dim's article was a really good coverage of it where it is now. I mean, these things are complicated but I think we slightly do a disservice to the military service men and women who are delivering the training to say oh, we just contract, drive it and solve our problems because actually there's a lot of extra hours that goes into making it work. You won't necessarily get that when you contract or if you do, you pay for it.

0:44:24 Tom Constable: Here's a stupid question surely we've got A, and I probably know the answer to this, but we've got a contract that's written that says they are required to have X number of aircraft available at all times. They are required to output X number of qualified and trained pilots at number of times. So what's happening? Are we a did we not contracted well enough? Or B are we not holding these bigger companies accountable?

0:44:44 Andy Fawkes: I think in this case it's probably six or one and half dozen, the other the mod and it should be able to change its mind with training. Over the period there's been a number of SDR, SRS, Strategic Defense Reviews and that's changed the demand. The world has changed also the aircraft have changed significantly. So where maybe when it started a lot of fast jets had two people and now predominantly it's just one person or one pilot with no navigators. So the world has changed and how do you contract when the world is changing? And particularly if you have operations, I'm thinking the enemy decides what the requirement is and you've got to change it. So I'm not entirely convinced how far you can contract things out when it's so intrinsic to defense the military is that training is such an intrinsic part of what you do.

0:45:31 Andy Fawkes: How far should you contract out such a vital function is my question really.

0:45:35 Colin Hillier: I mean, I think the original idea is MFTS is lead in training so it's not frontline squadrons. So that's pretty common. You go on like a Uracopter if you rotary or you do on a Hawk and or a Ticano before you go to your frontline. So these aren't frontline aircraft, they're leading trainers and that was why they originally said well, at least we can control that because it's common and it doesn't really change. I mean, it's interesting and you say that things change. Actually some of the aircraft I flew in were 50 years old.

0:46:00 Colin Hillier: They've been around for a while.

0:46:02 Andy Fawkes: Well, I think my emphasis was the front line aircraft have changed.

0:46:06 Colin Hillier: MFTS isn't frontline, that's the point.

0:46:09 Andy Fawkes: No, but it's providing the front line, even the numbers of pilots and what training they need even before they go to the operation of the Ocus, the operation conversion units, as I say. I mean, Tim explains this in his article about how the demand side has actually changed. That obviously affects the numbers of instructors you need as well. The instructors are essentially preparing people for different types of aircraft. As soon as you get faster and.

0:46:34 Colin Hillier: Faster sorry, we're going off on one. But actually the interesting thing is the really basic, the EFTs trainer, the little single prop aircraft, the very first aircraft you go into even that they had problems with. So if you take it to the most basic level and say, right, we could do what Oxford Flying School does, which is a basic flying training, I mean, civilians do it, it's no different. This isn't even any military content and they still had problems with having the aircraft availability.

0:47:05 Tom Constable: Interesting.

0:47:05 Colin Hillier: I think we should some really close examination in terms of it's easy to wave a hand and say, oh well, the military stuff is complicated, but actually even at the most basic level, we can't do it. And why can't we? I don't have an answer to that, by the way. I don't know why, but it was always a problem. Jets was a problem.

0:47:23 Tom Constable: Just going to interject there and although I know, Andy, you would like to continue this conversation, I just think I'm keen to stop it there. I'd like listeners to have an opinion and obviously please comment on the post on LinkedIn for this episode if we want to kind of continue this conversation in that. And you'll have Andy and Colin both exploring different angles, certainly. Right. Andy, if you don't mind, can we move on to the next one?

0:47:44 Andy Fawkes: Okay, so the next one I'd like to move to is for those who've been following the news, the wider news outside MSNC, there's been the story of a leak, a number of highly classified papers from the Pentagon. And I think the interesting story here is about the connection with the sort of gaming world. There's an article in the BBC and many others talking about this leak, but apparently one of the leaks came from a discord server. For those who don't know discord, that's a way that principally gamers can communicate very easily either or verbally or text or whatever, and it's used for other purposes. But there's a so called Minecraft Earth Map, which I thought was really interesting. Apparently there's lots of people creating maps of the world, essentially creating the world in Minecraft.

0:48:29 Andy Fawkes: I don't think this particular one, but some are going down to 1 meter resolution. So sounds very metaversian to me. But anyway, I digress some. Of these papers have been leaked through a number of these servers. And I know we were talking earlier, that's not the only one. There's a game called War Thunder, which is a combat simulation which apparently has 70 million players worldwide, and they for a number of years have been leaking information. So in in October 21, according to to the the BBC anyway, they leaked the design details about a French Leclerc tank that was about winning an argument about turret rotation speed.

0:49:08 Andy Fawkes: Yes, indeed. And in July 21, a user claiming to be a tank commander in the British Army posted documents about the armor structure of the vehicle to win an argument. It is kind of interesting that how gaming is now about how powerful I am in terms of my leak, rather than how good I am playing a game.

0:49:25 Tom Constable: Obviously those leaks are that's I think quite an interesting one for us, isn't it, about the relevance of like, simulation and then gaming and how that interlinks. Now the question is, is it a misinformation activity from the US DoD in order to cede incorrect information and cast dispersions and doubt over operational plans? Or is it just someone being an idiot and sending something over platforms they shouldn't be doing?

0:49:47 Andy Fawkes: Interesting in the heat of the moment. Yeah, possibly regret it afterwards. No, I think it's just interesting how gaming is just more, much more than just the playing games. It's a whole ecosystem that people are living in, spending their time and competing with each other. How highly classified the documents are they're going to share? Well, I'd like to think of all the intelligence services are on these forums all the time.

0:50:13 Tom Constable: Cool.

0:50:13 Andy Fawkes: The final one I'd like to just point you to, there's an article in Ms and T again about US Space Command. They were set up in December 2019, so not long ago. So it's a brand new force which is the same kind of level as US Air Force, and it's the only dedicated high level space command apparently in the world. But what I think is interesting, and please have a read of the article, is that in a way they could set up a brand new force and decide how they were going to structure it. As far as I could tell, there are three so called field commands at two Star, and one of them is Space Training and Readiness Commands. So they see that as obviously really important in terms of the overall aspect. And this command covers these so called Deltas. They're not like squadrons or air forces, they have these Deltas. And this Space Training and Readiness Command has five Deltas which unhelpfully are not numerically in order.

0:51:06 Andy Fawkes: But the Space Delta one for the whole command is Training and Education is 13. But it's interesting that they've put them together with doctrine and war gaming range and aggressor. So that's looking at threat and test and evaluation. So very interesting. Way they've joined up those areas. Which reminds me a bit of my old time in Mod when I was in the directorate analysis, experimentation, simulation, so trying to cover all those areas. So I thought that was very interesting.

0:51:31 Andy Fawkes: Obviously they're quite new, but they are increasingly, if you read the article, they'll see they're moving much more into war gaming and bringing in other areas of the US forces, like US Marines, but also other countries. So they're running a big exercise this month in April.

0:51:47 Tom Constable: And so these are the guys that unashamedly have the essentially a Star Trek symbol emblem that go with it, which I love the level of geekiness. I mean, I had a conversation with an officer responsible for setting this up for a country at their own kind of space force, and he was the one responsible for looking at training, and he said exactly the same thing. They were so privileged to be able to say, how do we want to approach this? Can we do this differently? And again, that level of open mindedness and obviously in geekiness willingness to engage with technology works with those kind of organizations. And he was looking at creating whole virtual space stations where you could put a VR headset on and go into your training, but not to do it in a classroom or not to do it in a in a vanilla environment. It was actually no. Well, we're a space force. Well, actually, let's do your phase one, two, and three.

0:52:29 Tom Constable: Let's integrate with a virtual synthetic sorry, Andy environment which has been created that just makes it different and makes them stand out. I quite like that approach and it's quite refreshing.

0:52:39 Jason Kennedy: Yeah.

0:52:39 Andy Fawkes: And I think anyone who's looking at this transformation of the military, maybe, dare I say an army, navy or Air Force could say, hold on a minute. Why don't we reorganize with a fresh piece of paper like Space Command. So, yeah, please have a look at that article. I think I found that really interesting. And obviously they're putting a heavy emphasis on training and readiness.

0:53:00 Jason Kennedy: Cool.

0:53:00 Tom Constable: I'm happy. Colin, are you happy?

0:53:02 Colin Hillier: Yeah, all good. Sorry, it's noisy and someone tries to deliver something just at the one half hour. I was not available all day. Well, no, I mean, you had to know, isn't it?

0:53:14 Tom Constable: Andy, thank you so much for your time. As ever, I really enjoy our time together and I look forward to having you, I think. Is it the last season? One episode is the next news. Andy? I believe it is.

0:53:25 Andy Fawkes: Well, that's exciting. I'm looking forward to that. Have a roundup of what's been going on.

0:53:31 Tom Constable: Enjoy your well and rest afterwards. So, thank you sue time and speak to you in the next episode.

0:53:36 Andy Fawkes: Great stuff. Thank you.

0:53:38 Tom Constable: Thank you, Andy. So I'm sure I would have mentioned this in the news, but if I didn't, then of course, I'm actually really sad that I'm not going to get to itech. But on the flip side of that, the shining light is that colin, you're going, aren't you?

0:53:50 Colin Hillier: Yeah, well, some of us have to work, so if you're there in Rotterdam, do come and say hi. Apologies in advance if I'll be in some deep conversation, but always nice to hear from people if you're there.

0:54:01 Tom Constable: Yeah, I mean, mate, you're welcome to look after the newborn four year old and two year old, and I'll go to itech, if that's what you prefer to do.

0:54:07 Colin Hillier: Yeah, we could do that, yeah. What you like on HLA, after last episode?

0:54:13 Tom Constable: All over it. You hope itch goes well and I'll.

0:54:16 Colin Hillier: Speak to you soon. Yeah, see you next time.