EPISODE TRANSCRIPT

Ladies and gentlemen, we are back. This is another episode of Things Reasons, the podcast for IT leaders. Jono, I’m pretty excited about this. I know he is sitting right there, but let’s just pretend that he’s not. We have a special guest today. I, of course have to introduce him. 

Let’s see if I can do this justice. Our guest today is Dan Wright. Dan is the co-founder and CEO of Armada He’s an entrepreneur with deep experience driving digital growth, operational excellence with disruptive technology. Dan has previously worked as CEO of DataRobot, a leader of augmented intelligence, and prior to that was the COO of AppDynamics. 

Pretty cool. A leader in uh, uh, Gartner Magic Quadrant for application, performance and monitoring for seven years in a row. Excellent. Beyond operating roles, Dan is an active board member as well. Uh, an advisor to high growth startups, a key contributor to the tech and academic community. 

Welcome to things, reasons. Dan Wright. Thanks for having me on the pod. Great to be with you guys. Welcome Dan Wright. Fresh off the plane from San Francisco. It’s awesome to have you with us, Dan. 

 We’ll get right into it. Tell us one thing that we should know about you and, uh, one thing that leaders often misunderstand about AI right now. I mean, I think one thing you should know about me is that I’m on a mission. The mission of the company, uh, that I, I built, uh, with my, my team and my co-founders is Armada. 

We’re bridging the digital divide all over the world. So man on a mission. That’s, uh, that’s fantastic. And there’s, uh, that, that mission is largely centered around sovereign AI and what’s happening at the edge and all of the craziness that’s exploding all around the world right now. Tell us more about that. 

Um, what a time to be alive. I mean, it, it is amazing. If you’re into technology, this is the time we are very fortunate to be living at this particular point in time. What’s happening is you’ve got a few specific things that are coming together to, to change the world with ai and one is, uh, UBI ubiquitous connectivity, the fact that you have internet everywhere. 

Uh, most people don’t realize Starlink was only launched in public beta in November, 2020. Mm-hmm. Now it’s in, I think, 155 countries, maybe 156 now. Yeah. And it’s, you know, everywhere. And it will be in the next five years. Um, ubiquitous, you’ll have connectivity everywhere. Second thing that’s happening is you have just huge advances in sensors, drones, robotics, and those technologies are getting more widespread, in part because they’re getting cheaper, more cost effective. 

So now when you look at oil rigs, you know, mining conveyors, ships in the middle of the ocean. They have sensors all over them, which are spitting off terabytes of data every day. And then the question is, well, what do I do with all of that data? The third thing that’s happening is that these AI models are getting more and more powerful and they’re enabling things like autonomy. 

You know where I live in San Francisco, you’ve got autonomous cars everywhere. We’re talking about this. Yeah, way more. There’s more. I think there’s more autonomous cars than there are cars with drivers. Now I haven’t met anyone who’s taken Waymo. That’s not a raving fan. Correct. I mean, it’s, once you take one, you’re like, well, why would I do anything else? 

Why don’t we have them in Sydney? Yeah. I mean, if they can navigate San Francisco with all the steep hills and craziness and going on, Sydney’s no worse than that. I feel like we should have, oh yeah. It should be a piece of cake, right? Yeah, yeah, yeah, yeah. We’re ready. Yeah. Your point on, I’m just gonna say to your point on connectivity is near, near and dear to Australians. 

Yeah. I mean, we suffered as a nation with connectivity for a long time before we had our. National Broadband Network. This is well before Starlink. Um, I mean you, you suffered to have just, you know, three or four, uh, megabits of connectivity and now we’re we’re cranking along, right? That’s right. And so Australia has transformed, transformed, sorry, markedly with connectivity Now, I reckon prior to that it was really hampering our progression as well. 

I feel like Australia as a nation with tech, and I know we’ll get into this, has progressed rapidly since Australian households and businesses are now much better connected. Yeah, it’s amazing. I mean, once you get the internet, then it brings the question, well, what else can I do? Right. You know, and, and the the beautiful thing about technology now, and this is again, part going back to the mission of bridging the digital divide, is you can do it literally anywhere. 

Um, and so you could be in the middle of, um, you know, the Outback, you could be in the middle of, you know, we do deals where it’s in the middle of the Amazon, middle of the ocean, and you can build, um, you know, cutting edge AI factories literally in the middle of nowhere. 

For people who, um, aren’t familiar with, um, your technology, Dan, I mean obviously, you know, Armada are leaders in your, your space and your capacity. 

Just talk us, talk to us a little bit about the innovation, um, that’s occurring within Armada Yeah, well it’s, I think it’s helpful first to talk about how has the cloud evolved, right? So the cloud launched, um, you know, whatever, 20, 25, 26 years ago, and. At the time, the whole idea was to build these massive centralized data centers and then provide compute capacity that could be readily accessible and rapidly available to, to everybody. 

 The problem was it wasn’t actually available to everybody. It was available to. The specific parts of the world that had these big hyperscale data centers. But if you look at a map of the world, that’s only about 30% of the world. Yep. That means 70% of the world is completely underserved, has been left behind. 

 And I think actually when you look at some of the problems that exist in the world, and certainly in the US where I live, that divide between the haves and the have nots. Causes a lot of downstream issues, and it turns out that that 70% of the world is where there’s huge disasters. Right. You know, whether it’s, um, you know, a split second life and death scenario in a disaster response context when you’re responding to wildfires and floods, or if you, um, are in mining and there’s a, you know, a, a mine disaster or an oil spill in the middle of the ocean, right? 

 Those areas have been left behind. And so what Armada does is we provide full stack modular, modular, rapidly deployable AI factories that can be deployed in weeks, literally anywhere in the world, and we’re providing distributed compute. Mm-hmm. Right. It’s flipping the whole paradigm where before if you wanted to do any data processing, you had to have the data sale to some far away data center. 

 The reason that we named the company Armada is that we actually have the data center sale. To the data. And so, so the, the, the name of these modular AI factories, they’re Galleons. Gallian is a type of ship. Got it. And they sail to wherever they’re needed. I didn’t know that. There you go. That’s actually pretty cool. 

 And uh, when you found a company, you get to come up with cool names for stuff. I mean, that’s probably the best part. Yeah, it’s definitely one of them. I mean, what’s more fun than that? Exactly. I told you my 9-year-old came up with the, the latest name hasn’t launched yet, so, yeah. That, that’s, uh, top secret. 

 We’re not that here today. Secret. Secret. No, maybe not. Yeah. Keep it off Twitter. Um, but mate, yeah, it, it is, it is a very cool story and it’s so important, uh, so important in the age that we’re in. We spoke about this on a, on a previous pod, Narin and I talk about this all the time, that the rise of AI could actually lead to the next sort of revolution in, uh, in human development development of our society. 

 Uh, something akin to the internet or the launch of the smartphone, or even the development of the printing press. So infrastructure that enables that to bridge that divide so that we don’t have haves and have nots that everyone can participate. Uh, it’s a really worthwhile mission. Yeah. That you’re on. 

 Thank you. Yeah. And I mean, I think that’s the highest and best use of technology. I mean, technology is just a tool. Yes. Right. And it should be a tool to bring us together. Yes. Not, you know, create a. And it should be a tool to solve the most important problems in the world. Yeah. And that’s how we look at it. 

 Love that. Love that. I mean, we are very lucky. Um, previous podcast to this, was with, um, uh, Dr. Kate Gwynn, um, talked about human-centered design Yeah. Around AI adoption. She does a lot of work with, um, higher education as well. And it’s just the importance of augmentation, right? 

Yeah. And that’s, that’s I think, the way we need to think about it. That’s certainly augmented my life. Right? I enjoy the tech. I like working with it. Um, I wouldn’t think for a second. Certainly not within my business. And I know you wouldn’t in yours, we’re not looking to replace people. No. We just wanna enhance them and enhance output. 

 And it’s consistent with, you know, all the messaging that we’ve delivered to universities of la I do a bit of, uh, guest lecturing, um, and the graduates just wanna know that they’ve got a future and they do have a future, but maybe there’s a different expectation on output and so what, that’s great. Yeah, exactly right. 

 You can produce more and do more and maybe do it better and smarter, et cetera. And I think that’s the way we should think about it, for sure. Yeah, absolutely. And Dan. Uh, it’s a privilege to have you here, uh, from the US and we know that the US market is often likened to Australia, but perhaps a little bit ahead that 12 to 24 months ahead. 

 A lot of our listeners would be super curious to learn a little bit more about, uh, the United States National AI Plan. Yeah. And, and how that’s occurring on the ground over there. Obviously, Australia’s just released. Uh, the Australian version Yeah. For Australia. And that’s probably, I’d say been launched to mixed reviews mm-hmm. 

 From both industry and, and, uh, and the enterprise. But, uh, what’s happening in the us can you give us a bit of a feel for the dynamics on the ground, the, the pace of adoption, uh, where you’re at in the adoption curve? Yeah. Yeah. So I was actually at the event in Washington, DC. When the, um, AI action plan was announced by the White House and the, the plan has three pillars. 

 It’s, you know, refreshingly simple. The first one is all about rapidly deploying infrastructure, uh, all over the country, all over the us and that’s related to the second pillar, which is. Cutting red tape, you know, there was a lot of red tape to build these big hyperscale data centers. Mm-hmm. That has flipped. 

 Where now actually there’s a lot of incentives to build out this infrastructure. Uh, for example, in, you know, the one big beautiful bill, there was a bonus depreciation rule. And what that allows organizations that are profitable to do is buy infrastructure and then write off the cost of the infrastructure. 

 Wow. But they get to benefit from the ai. Internally. And then also if they’re like, say a big energy company, then they can actually create new revenue streams, monetize any excess that they don’t use with the, use themselves with third parties. Right, right. Incentive. So created some incentives to do this rapid build out, which is working. 

Mm-hmm. The, the build out is, is really in warp speed. And then the third thing, and I think that this was a lesson that the US took from 5G, is that speed really matters. And not just speed for domestic employment, uh, deployment, but speed for exporting this technology to allies around the world. So the third pillar was all about exporting the US AI stack, uh, really faster than China can, but faster than anybody else was the goal, right? 

 And at the same time, there was this realization that that can’t happen in a vacuum. You have to work with allies to do that. And so there’s lots of incentives to do that, for example. Um, you know, director Michael K CIOs, who’s great, who’s the kind of White House tech advisor. Um, Ethan Klein, who’s the, the CTO of the White House who I met with a couple weeks ago in dc They’ve been working with, uh, export Import Bank with commerce, uh, with the State Department. 

To basically go around to Allied Nations around the world and say, Hey, can we partner on some of these big AI infrastructure projects? And it’ll be a win-win. Well, you’ll get access, you know, in some cases privileged access to the latest and greatest AI technologies delivered rapidly wherever you need ’em. 

 Um, and then for, for the US they’re thinking we’d much rather live in a world five years from now where. We have allies that are working with us on AI initiatives versus working with say, China. So it’s interesting, isn’t it? ’cause there seems to be two things going on at the same time. We’ve got this export of American AI technology, which makes sense. 

 Mm-hmm. Right? Getting those frontier models out there combined with governance, government, sorry, effectively looking to nationalize their compute capacity. That’s right. And it feels like these things are happening at the same time. A hundred percent. Our government’s talking sovereign, that word sovereign, how often do you see that now on LinkedIn? 

 Right? All the time. All the time. It’s everywhere, right? Yeah, absolutely. I mean, I was at Davos. Um, you know, earlier this year and all anybody wanna talk about is sovereign ai, right? And it’s continued. And this is where we feel like we have a unique advantage because you can take these latest models, whether you’re talking about models from open AI or xai, or you name it, you can run them disconnected from the public cloud right at the edge. 

 Hmm. And we think that that is the future. Every country and every company. Is gonna have its own sovereign AI factory. And even like this week there was gtc Yes. In San Jose near, uh, where I live. And we had a bunch of our team there. We had a gallion there. And you know, Jensen is talking about this, like every single company, every country is going to have its own sovereign AI factory. 

And when people think sovereignty, historically, they’ve always thought about sovereignty at like the country level. But that’s not the way it’s being thought of anymore. It’s being thought of, of. Sovereignty down to the site level. Hmm. If I have an oil rig or I have a mine, or I have a university, I want all of those models to run at the edge. 

 Air gapped disconnected from the public cloud. Why do I want that? Well, I want that because ultimately the number one time when data is hacked, when it’s most vulnerable is when it is sent back to the cloud. So the more you can do the data processing locally at the edge and then rather than sending like the entire model back or all the data back, you’re just sending the metadata back. 

 It, it really increases your security posture. Yeah, right. Speaking of security posture, I think, uh, in the US defense considerations have, have a big role to play. In alignment with this strategy. It’s quite an aggressive strategy, I would say, compared to, uh, Australia’s current iteration mm-hmm. Of our strategy. 

 I think that may change as, as demand changes, but, uh, can you tell us a little bit more about, about defense technology and how the US government is thinking about this rapid scale out of distributed ai? To compete in this race with China or the global race for ai. Um, and what that distri, what that distribution does for your security posture. 

 Also your defense posture. Can you elaborate on that? Yeah. So I’ll give you an economic answer and then I’ll give you like more of the traditional defense answer, right? If you think about, um, you know, what affects the standing of different nations in the world? A lot of it is your economy. Right? And if you win in ai, it’s gonna have a huge impact on your GDP Mm, right? 

 Your, your companies that you have domestically are gonna be more productive. Especially as we start talking and, we’ll, I think we’ll talk more about this physical ai we will, the use of robotics. Yep. Right. The use of automation. You know, we work with some of the largest energy companies in the world, and they’re talking about fully autonomous rigs, not, you know, 10 years from now, like within the next couple of years. 

 Right. So you think about the downstream impacts, if you’re able to produce more oil, if you’re able to produce more raw minerals, if you’re able to, uh, have factories that are fully autonomous, right? You’re just gonna be able to, number one, have a, a better, uh, more efficient internal. Uh, economy. But then secondly, you can export that technology to others, right? 

And so that is kind of a core pillar of, of the US strategy. But the second part of it is you need to be able to, um, utilize these technologies that are increasingly how, uh, deterrence is done. It’s how conflicts are resolved. You know, for example, drones, right? People think, oh, I can just use a drone anywhere. 

 But the reality is that drones dramatically lose their utility the further you are away from these hyperscale data centers. Yes. Right. And, and a good example of this, we put out a video public, uh, with the state of Alaska, you know, very remote US state, and they use drones for responding to floods and avalanches, you know, life, life and death type stuff. 

 And the latency to process all the data from the drones was more than 24 hours. Huh. You think about that. It doesn’t work. Doesn’t work. No. You know, avalanches and floods, they’re not gonna wait a day. Yeah. It’s just outside of tolerance. Massive. You gotta know like that a hundred percent. And so that, that is where, um, I think when you think about it from a defense standpoint, infrastructure is the key enabler for all of these other types of technologies. 

 Whether you’re talking about drones. Uh, counter UAS Yep. Uh, you know, autonomous technologies and increasingly that’s going to be critical in a contested scenario. Yeah. I mean, look, we’ve got the two conflicts, you know, and I know they’re, they’re terrible things, right? Obviously the Ukraine and more recently in Iran, and, and we’re just seeing, um, specific assets being targeted, right? 

 Yeah. Data centers in particular being targeted. And I’m just wondering whether that in itself is gonna influence distribution and the way data center are considered. Yeah, absolutely. I mean, there’s always been this concept, and I mean, you guys know this, but. Uh, rack, rack resiliency with data centers. 

 Mm-hmm. But now it’s like you need to think about node resiliency, right? You don’t want this like mass of concentrated data centers in a small geographic area that becomes, you know, a big concern. Mm. Instead what you want is distributed compute where, number one, these things are geographically dispersed, so they’re much harder to target. 

 And number two, if one is ever threatened, you can detect that well before something happens. You can trigger some defense mechanism. Right to try to knock, let’s say a drone out of the sky before there’s a disaster. Um, and then three, if that doesn’t work, you can remote wipe it and send, you know, back up to another node. 

 Yep. Right. So it’s very resilient, hard to kill infrastructure. And I think what you’re gonna see is that other types of critical national infrastructure too, whether you’re talking about airports or oil rigs or mines, you’re gonna wanna have a similar type of setup where you have. You know, distributed infrastructure to do 24 7 monitoring of those assets, not just for defense, although that will be one thing. 

 But then also, um, you know, for things like. You know, conveyor fires or safety issues. Yeah, just making sure that these things always work as they’re intended and that there’s no problem. You know what I mean? I think we, we talked a little bit about connectivity before, and obviously that’s near and dear to all Australians, but also the advancements in, in latency and more importantly, um, being less, uh, impacted by latency tolerance. 

 And we’ve spoken about this before. I mean, these advancements have been happening in parallel, particularly as connectivity’s improvement. COVID helped with that. Everybody working from home. It wasn’t that long ago where we used to panic over five or six milliseconds. Right. For things like synchronous replication of databases, et cetera. 

 And there’s been solid advancements in all of that. Right. And I think there’s a lot of myth and misconception out there around just how important latency is. I’m not sure many people in Australia know exactly. Where their data resides, particularly maybe as it relates to inferencing and the models they’re using. 

 And half the time those, those models aren’t even in Australia anyway. Right. And email particularly is nowhere near latency sensitive. I mean, where am I getting with that? We’re going with that. Sorry. I think the tech, as it relates to latency, has been improving markedly, um, over the last few years combined with, I think, um, a greater propensity for distribution, uh, of AI data centers particularly. 

 Yeah. I think it’s those advancements that are gonna help. Uh, unlock the ability to dilute that concentration. So if you start to move to that more distributed model, especially as society becomes more reliant, uh, act frankly dependent on AI systems, uh, to run their lives or their businesses or important services and utilities, uh, you can see. 

 Uh, that playing forward in Australia as well. We’re somewhere, somewhere else in the, in the adoption curve, maybe in sort of the foothills. Yeah. Uh, for this technology, the US is a bit further along. Yeah. And uh, and you can see there if distribution of this tech is becoming important, that’s probably the way that we’re gonna head as well, because it seems to be the natural thing. 

 I mean, the next few years when it comes to AI is going to change everything. You know, like what we were talking about with the Waymo’s, that’s the tip of the iceberg. You’re gonna have fully autonomous, uh, recycling trucks, you know, fully autonomous, uh, robots in everybody’s house. Like this stuff is not way out. 

It’s like. It’s coming. And the, the demand for compute, not just for training these AI models, but critically as we were just talking about for inference and fine tuning of those models, is going to increase dramatically. And I think Australia has kind of a once in gen, a generation, once in a lifetime type of opportunity why you are, uh, strategically, you know, positioned. 

You have your, your land rich and your energy rich. And that matters a lot. That’s everything when it comes to ai. The question, and this is the same question we ask ourselves in the US, is how do you fully utilize all the available energy and the land, wherever it is to support this global AI boom so that you can fully take advantage, um, of it. 

And I think, you know, some things like, uh, what we’re doing with. You know, WinDC. That’s, that’s the way to do it, is like you use whatever energy in, into, including renewable energy as you were just saying, you know, the, the latency is really not much of an issue. It’s not a problem. And, and you use the, the land and the energy wherever it is to fuel the ai boom. 

Yeah, indeed. Well, look, the technology’s evolving rapidly. Um, and I think we’ve, we’ve had new words enter our lexicon, the, these words of neo clouds. I mean, was anybody talking about neo clouds five, six years ago? I don’t think so. We weren’t right. It’s definitely a thing now. It’s a new thing. And so I guess a question for you really on, um, it’s that balance between hyperscalers and neo clouds. 

I feel like there’s a role to play for both of them. Um, talk to me about that, like your, your view on sort of positioning in the, in the broader ecosystem of capability, how they go together. Yeah. No, I mean, it’s, it’s actually kind of amazing. I mean, these neo clouds have grown so fast. I mean, it wasn’t very long ago, everybody thought Neo, they thought about the matrix. 

Now everybody’s thinking about these, these huge neo clouds, and I think there is absolutely a place for them. Yeah. And, and the reason is the, the cloud providers. You know, they built for a different world. They built for a world before ai, certainly before the physical manifestations of AI is more focused on CPUs than than GPUs. 

Sure. Um, and certainly not thinking about a world where you have these complex workloads, um, where you need to orchestrate across multiple types of GPUs. You know, it’s not just the Nvidia GPUs now it’s multiple types of Nvidia GPUs plus a MD plus, you’ve got other players like Cereus and uh, obviously, you know, Nvidia just. 

Bot rock, you know? Sure. And so you, you have different types of, uh, GPUs and you need to handle the orchestration across all of that to get to an optimized infrastructure that can support those types of workloads. And that is one area where the neo clouds have a big opportunity. The other area that the neo clouds have a big opportunity is just the geography, and especially with this whole focus on sovereign ai. 

 Sure. You know, a lot of people are not content to send all their data across. Some border to some far away data center anymore. They want their own local Yeah. Uh, you know, AI infrastructure. And that’s where I think the neo clouds have a, a very large opportunity. Um, you know, but then again, they also have to be willing to think about things, you know, differently. 

And that’s part of what we say too, is you don’t need to just build the hyperscale data centers. The hyperscale data centers will continue to be important. They will, but you need to have, uh. Distributed compute that supports not just training of these models, but efficient inference, uh, you know, the running of these models located at the edge against Sovereign. 

Yeah. And that, that to us is, is where, uh, the neo clouds have a big opportunity and, and so we’re partnering with them, for example. Sure. Um, you know, at the Super Bowl, the big, uh, the big game in the US we were there with N Scale. Uh, we co-sponsored the Super Bowl breakfast with them and we announced our, our letter of intent to partner with them. 

 Fantastic. Um, and again, thinking about these things are better together, the big hyperscale data centers with these smaller edge data centers together makes sense. And look, let’s just look at Australia, right? I mean, I. I’ve recently had solar hooked up to my house. Right? Yeah. I’m fortunate to have a battery and I’m doing the amber arbitrage thing and I love it. 

Right. It’s a good game if nothing else. Right. But I’m lucky to be able to do that. I mean, Australia’s had 20 years to get ready for solar panels. If you think our grid was ready for that. I’m gonna say it out there on the podcast. I, but I don’t get in trouble. But we haven’t kept up with it. You know, the grid hasn’t, uh, kept pace with even that. 

Right. And so transmission is a problem. Uh, we should have community batteries everywhere. And shout out to my mom and dad, they’ve got a community battery project Yeah. On the go, right? Why should it be homegrown retired people like my mom and dad trying to get batteries up and running in their, in their area, right? 

Because they don’t exist. So that’s a little bit of a rant from me on the state of play for our grid and transmission. Um, I feel like demand in Australia and everywhere with ai. He’s gonna set the pace of this stuff. Yeah. Right. And if it’s, it’s no due to no fault of any of the metro data center providers, they’re doing all the right thing. 

They will always have a place in our market. Of course. Um, but we need something else too. That’s right. Right. It’s not a case of one or the other. It’s a case of, and, and because we need to deliver the tech to the people, to the businesses such that we can keep pace with everybody else. Right. And I think if our grid was different, I mean, God bless it. 

Right. I wish. With the MBN project, they should have been doing transmission upgrades and all the rest of it as well. Right. At the same time. That wouldn’t that have been wonderful? Well, I think the, the thing is nobody really saw AI coming. Well, no, I didn’t. I, there’s probably people who were working on AI that saw it coming from an academic perspective, but certainly, uh, being monetized and becoming mainstream tech, it sort of felt like it just exploded. 

And it is now outside of, I’d say in Australia, uh, if you look at. Uh, projections for consumption of power outside of the mining industry and sort of smelting awe and things that consume a lot of power. Mm-hmm. Uh, AI is where the all of the new demand is coming from, and it’s exploding. And I think, yeah, grid infrastructure is gonna have a challenge to keep up, and that’s where our distributor models start to make sense in helping us keep up with that demand. 

I was at the data centers, uh, conference, the leaders conference yesterday. There was a great panel discussion all about this. Mm. Around, uh, the demands for grid hookups and the, the debate that’s playing out live in Australia around how do we get access to this power? We need to do ai. It’s important. Uh, it takes five years to get approval to cut a new transmission corridor. 

And, uh, not to mention the money. So there, there’s a mismatch with how fast we need to move and how Yeah. Quickly we can practically move as it relates to some of that fixed infrastructure. See, that’s right. I mean, uh, to, to the point you just made, and this is something that’s top of mind for us in the US as well. 

I mean, your ability to utilize all of the available energy, it’s an and is critical to winning this in this global AI race that’s going on. I mean, as another example. We just, uh, did an announcement with the state of North Dakota and they’ve got two gigawatts of stranded energies. Huh? Stranded natural gas. 

And in many cases, they literally have to, to bury the gra the gas in the ground so that they can produce more oil by regulation. And so why wouldn’t you use that to power AI factories? A hundred percent. And you do that in such a way where you don’t impact individual. Uh, you know, citizens, electricity bills ’cause you’re using behind the meter power. 

You’re creating new jobs and you’re benefiting the GDP of the state. You’re helping the economy of the state and you’re helping the country. It becomes a no brainer. And I think that’s, you know, the opportunity that Australia has as well. 

 Tell us more about this concept of physical ai. 

And I know when we spoke the other day, it was kind of the, the first time I’d probably heard it that succinctly. Um, I think everyone’s aware of Tesla robots and other things. Yeah. So when you think of physical ai, is it that, I think it’s more than that, and I think it’s advancing at pace. Uh, in your neck of the woods? 

Yeah. Compared to Australia. We don’t have Waymo here yet, for example, although, fingers crossed. Fingers crossed. I think Teslas are allowed to drive themselves now in New South Wales. That that’s pretty new. Well, you gotta have the dead man hand on there, you know what I mean? Oh dear. You gotta buy that thing on Amazon that looks like so you can have a sleep on the back seat’s. 

 Didn’t say that. Yeah. Disclaimer, don’t take our advice, but you could do that. Hypothetically. Tell us more about, uh, physical ai. What are some of the advancements you are seeing that we might expect or our listeners might expect to see in the next two years here? I mean, one of the benefits of living, uh, you know, in San Francisco where I live, is we get a little glimpse into the future. 

 It’s like the Jetsons. Yeah, a little bit. I mean, so, so you’ve got. The Waymo’s. And it’s funny, you know, even the US like I had one of the members of our team come in from Chicago and I was like, you gotta take a Waymo. And he was like a kid in a candy store. He is like, this is, this is like the Jetsons. 

 This is amazing. You know, it’s like living in the future. But you know, we’ve got new robotics companies springing up all over the place. You go into a laundromat in San Francisco and there might be a robot there that’s gonna fold your clothes and do the laundry for you. Mm-hmm. But that stuff’s not, again, it’s not decades out. 

 No. It’s gonna happen in the next few years. And it’s gonna blow people’s minds and it’s okay. What are all of the things that are highly repetitive that I could train an AI on and have it do for me? Yes. And probably do better, faster, cheaper, uh, than me and make my life better. Give me time back to go do some higher level thing. 

I think when you start talking about that in the business context, people get a little bit sensitive. Um, but if think, think about it first, like in the personal context, like who wouldn’t want. A personal robot that can clean your home, uh, can cook for you, you know, can maybe walk the dog in the backyard. 

 Yep. So you feel like a good owner. Um, there’s so many different things that these, you know, robots, which again, are just physical manifestations of, of AI are going to do for us, that are going to make our lives better. And then I think the, the ultimate example is when you think about healthcare, we were talking about this Yeah. 

 A little bit. You know, you think about. The previously unsolvable problems, the diseases that will be cured as a result of this technology, again, not decades out in the next few years, that is when people are gonna be like, oh, this is a good thing. This is a great thing and then that’s only gonna help speed up this whole cycle that’s going on. 

 Yeah, look, I mean, I think, um, world models particularly, I’m glad you brought that up, right? I mean, we talk about, you know, are we on a path to a GI and do the frontier AI firms have a solid path? I’ve got a view on that. I’ll share that in a minute. Um, but world models particularly, I mean, it’s about causation, right? 

 And obviously within what we call system one tech, it’s all correlation and relationships. And what we need is for the tech to be able to make decisions, not anchored in data points or learning. It needs to be intuitive the way we are, right? When we crossed the street, we don’t need to have crossed a street just like it with same cars, et cetera, to make the decision. 

 We infer and we, we look around and we can, we can do the causation ourselves with that. And I think world models are gonna seek to try and solve that. Um. How do you feel about a GI, I mean, we’ve talked before about a two speed economy. There’s the tech we know and love today that has perfect application to augment and help every one of us in our lives and in business, et cetera. 

 It feels like the markets are anticipating a GI, I know Altman was talking a good game about it, our dear friend, Elon Musk. Right? I mean, he’s talked about maybe needing that as well to make his camera based optical, uh, autopilot if he’s allowed to use those words. So I dunno how the court case is going on that right. 

Um, so my point is, uh, how important is a GI in the current market, and do you see world models contributing to getting there? Yeah, I mean, a GI is, it’s like many different definitions from many different people, right? But I, I come back to it and I say, well, ai, what is ai? AI is a tool. Mm-hmm. And I think we’re starting with more basic tasks that it is automating, you know, driving a car, driving around San Francisco is a good result of that. 

You know, something that it speeds up the time that it takes me to do research on a question that I might have. That’s another good example. Something that, um, augments my ability to code. Mm-hmm. Right? Like claw and all these things. Um, and then you’re gonna continue to get more and more sophisticated. So I don’t think there’s gonna be like a aha moment where, you know, okay, we’ve reached a GI, I think it’s going to be just a series of more and more, um. 

 Uh, powerful models that are able to automate more and more complex tasks. Yes. And that is gonna play out very fast. Mm-hmm. You know, um, and I think that what we’ll see is that that is going to grow exponentially in the coming years. Like, it’s, it’s amazing to me how. Especially the physical manifestations of ai, but, but Chachi PT and all these other things, they’re, they’re a good example too. 

 Like once you get a taste of it, you just wanna do more. And then it, it causes you to think like, what’s the next thing I could do? What’s the next thing I could do? It does, it is the imagination. Yeah. It’s good at prompting you too. Hey, I, I’ve got this idea now that you’re on, you’re kind of thinking about this. 

Would you like me to explore that for you? Sure would. Yes. Yes. Well, and then, then the other thing is, you know, the, the capital available to create companies that solve these types of problems is at an all time high. Yep. You know, it’s only going to continue to accelerate. Um, and so again, that will just speed things up. 

 Alright, so Dan, we’ve uh, we’ve made it to the section of the podcast. We’re gonna give you our quick fire three questions we ask every guest. Alright, let’s do it. Alright. Now you can say whatever you want. Okay. Alright. So you can be as controversial or otherwise. Here we go. Question number one. What is one piece of advice that you would give your past self before taking on a major tech initiative or challenge? 

 I think that the number one thing I would advise my past self. Is, write it down, huh? And, and what I mean by that is if you’re building a company or you’re really starting any initiative, you need complete clarity so that you can, in one, understand what you’re doing. Uh, number two, you need to tell you mission, uh, your vision for the company or the initiative in a really compelling way. 

 They can get others to say, Hey, I wanna leave my very comfortable job and go do this. And also for investors, it matters a lot. And so it’s a very simple thing. But before we ever started Armada, before we incorporated the company, the first thing we did is we sat, uh, you know, in our case, in Founders Fund, one of our investors, and me and my two co-founders, Jono and Pradeep, we wrote down the mission, vision, values of the company. 

And it’s amazing how, you know, that took us a day. But the amount of mileage that we got from that. Um, and also just peace of mind knowing what you’re doing and why you’re doing it. Mm-hmm. Lot if you write a lot that today. Lot just, I’d love that by the way. Thanks for, for sharing that. If you, if you were to write that today, would you write the same thing? 

 I would exactly the same. I’d write, yeah. I’d write the same exact thing. And in fact, some of the things that we saw have actually happened faster than we even thought. Hmm. You know, like, um, you know, just as, as, as an example. Sovereignty. This whole trend around sovereign ai, we saw that as part of the need for distributed compute. 

 Um, and I actually did a blog, like the day we launched the company outta stealth. And it was all around, you know, distributed compute for, you know, kind of distributed world. And that has gone faster than any of us thought. Hmm. And I think it’s because of just everything going on in the world. You know, we didn’t know that we would have these conflicts around the world. 

 We didn’t know that there would be these massive cyber attacks and breaches. Um, but we had a feeling that that was gonna be important and that it’s played out. Can I tell you, that’s bold, right? Because you’re almost saying in the same breath that I’m gonna choose my words carefully, but you’re almost betting against the hyperscale. 

 It’s there. Right. And I know you work closely with them as well. Right. But it’s like, you’re almost saying this is a problem that’s gonna get solved in conjunction with, as opposed to just with hyperscalers. Right. Yeah. I mean, I, again, I think it’s an, and the hyperscalers are not going anywhere, of course. 

 So No, nobody panicked. If you work at any of the hyperscalers, your, your jobs are very secure. Yes. Um, but what’s going to happen is it’s going to be an and yeah. Where you’re going to have the hyperscalers, you’re going to have, uh, you know, neo clouds. You’re going to have companies that have access to land and power that work with distributed infrastructure companies like Armada to put that to work for as part of this big AI boom. 

 Um. And that is only going to continue to accelerate, but I think it’s an and. Yeah. Great answer and some good advice for our budding entrepreneurs in the audience. What’s something you used to believe in leadership that you no longer believe? One more thing on the the hyperscalers, then I’ll come, come back after that. 

 The other thing is actually, not only is it an and but, and we were talking about this a little bit before. We actually help extend the hyperscalers. Yep. And what do I mean by that is that you take the services that people really like. For example, we have a great partnership with Microsoft. Of course there’s things that people really like about Azure. 

We take those and we run those local. Like actually I met with Satya Nadela. Hmm. Um, earlier on at Armada. And he said, I, I love this. ’cause basically every time the connectivity rolls out in a new geo. You are like extending the capabilities of Azure and you’re taking it local. Yes. And then he actually went up, um, and I didn’t expect this, but like his next, uh, speech was at Ignite the following Tuesday. 

 This is like on a Friday. I was there up in Bellevue. Yeah. And he talks about how Armada helps take Azure local. So that’s the way we see it as it’s an and, and we actually, uh, make. What people love about the, the hyperscalers even better by extending them to these other locations. I love that. I mean, look, that’s near to my heart. 

 I mean, I’ve gotta mention it, right? Yeah. I mean, I, I’m an Azure guy. Yeah. I, so, so I do work very closely with Microsoft and it’s their adaptive cloud strategy, right? Exactly. And you’re, you’re able to essentially extend the native perimeter of services anywhere you want. And I think that’s been probably the greatest innovation in addition to Azure, natively, I think in the last couple of years, is being able to do that. 

 So I love that. Yeah. And I mean honestly, a lot of the customers that we work with, you know, the largest energy companies, mining companies, um, you know, public sector companies Sure. You know, agencies, they love it too because again, it’s taking something that they know and they have a very well-defined understanding and value prop. 

 And you’re allowing them to use it the way that they want to use it. That’s right. Which is local. Yeah. Um, and then to your question, I think your question was around like, what advice would I give to entrepreneurs? Uh, something you used to believe in tech leadership. Ah. It’s a little more spicy. Okay. That, that you don’t believe anymore. 

 I, I think, um, the number one thing that I have learned is that you need to trust your gut. I used to think, like, before I make a big decision about the company, I need to, um, you know, ask, you know, 10 people and have a long conversation about it. But the truth is, if you’re building. A company, especially a company that’s doing something new, you know more about it than anybody else. 

 And so your gut is usually telling you the right thing to do, especially if your gut is informed by data. Hmm. And then, you know, the ability to actually make very quick decisions has a lot of compounding benefits. So I think that’s the number one thing that, uh, has changed. I used to be a little more contemplative.   

I used to always want to, you know, get everybody’s opinion before making a big decision. Uh, now it’s like when I know that something is the right thing to do, I do it very quickly and decisively, and I think it’s, um, you know, maybe a better founder, a better CEO. Do you think that comes from subject matter expertise, belief in the mission? 

 Uh, what is it that drives Yeah. Your, your gut ability to connect the dots maybe more efficiently or faster than. Say others who are, who have only recently been read in, for example. Yeah. I mean, I think it’s, it’s an interesting combination. It’s this combination of having been doing this for a while. So this is my third technology company, and so you, you learn something. 

 Yeah. You know, and, and everyone has had something to do with, you know, how do I better use my data Hmm. To solve and pro problems, you know, in the world for my business. Um, they’ve just been taking it from different angles. But then the other thing is. We’ve gotten very deeply immersed in some of these new use cases that nobody has any experience with. 

 Yeah, right. And so if I go, it’s novel. If I go try to talk to, you know, 10 of my entrepreneur friends and I say, how would you, um, you know, use these cutting edge frontier models on an oil rig to create value? Nobody’s gonna know better than we know. ‘ cause we’re doing it every single day with the largest companies. 

 You know, in the world and we live that. Mm. Um, and so I, I, I think that’s part of it, is this, this tension between having done this for a while and taking the lessons from that with doing something that is brand new that nobody else has the answers. 

We’ve spoke before about entrepreneurship right? 

 And, um, and the embracing of generative ai, and I’m just wondering whether, um. People should really take heat of that. Right. And be brave. Right. ’cause they’re gonna need to be, right. I mean the market’s wide open now. There’s so many different little startups that are firing up left, right, and center embracing the technology. 

 And if you’re not running on instinct, what are you running on? You’re not moving fast enough. Yeah. It has to be, you gotta move fast. Gotta move fast. You have to Right. People talking now about an individual can launch the next AI unicorn probably all on their own. That’s right. I believe that, by the way. 

 With agents. Yeah. Yeah. No, it’s gonna happen. It’s, it’s a hundred percent gonna happen and it’s gonna be somebody who. Deeply understands the data and the workflow. Mm-hmm. Um, I actually think the subject matter experts, you know, have this huge opportunity. If you’re somebody who, let’s say that you work, you know, in industrials for your whole career, you don’t have to be an expert, you know, coder. 

You don’t have to be an expert software engineer. You just need to deeply understand the data that’s available and you know how you can actually automate some of these very manual workflows. And then you know, you can find somebody to build or you can use something like Claw to build it yourself. 

 Astonishing. Look. Great, Jono. Here’s the thing, right? We are to believe that if we are in a simulation, it’s the most interesting path that we will find ourselves on. I feel like this time is the most interesting. You said it yourself. Yeah, a hundred percent. What a great time to be alive. Amazing. We’re definitely in a simulation. 

 Working with you tells me that more than anything else. His life is funny the way that works. Well, Dan, it’s been an absolute pleasure to have you on the podcast. Thank you very much. Um, amazing work with Armada. It’s exciting to watch. It’s like every second day there’s a new announcement for your business, by the way, on LinkedIn. 

 And the fact that you can be on LinkedIn more Than US says something as well, by the way. Yeah, that’s right. That’s right. We’re everywhere because we’re everywhere. Uh, Dan, it’s been an absolute pleasure. Likewise. Thank you for your valuable time and insights. I know, uh, for me, the biggest takeaway for me is I’m pretty excited about physical ai. 

 Mm-hmm. Yeah. Uh, I can’t wait. I can’t wait for that to, to happen. And, and we can start benefiting, benefiting from that, uh, incredible stuff. And it’s always good to learn more about what’s happening with our, uh, American. Cousins, uh, in tech because it is a couple of years ahead. Mm-hmm. Um, our partners, it’s a That’s right. 

 And it’s a little glimpse into, into what’s to come. So I know our listeners will have got a lot of value from that. Uh, where can listeners listen to this Narin? Well, it’s always where you find your podcasts. Ladies and gentlemen, I’m talking about you, you get your podcast from Apple Podcasts. It could be Spotify. 

 It could be YouTube. Thank you very much for subscribing. We are of course, all over LinkedIn. We are produced by Pru Loon and Karina Aguilera, and you can find us at pru@thingsreasons.show Thank you very much. That’s a wrap. That’s a wrap everyone. Thanks. Thanks again. Next time. It was a blast.