00:00:00:00 - 00:00:25:15 Unknown Everybody, welcome back to, another exhilarating episode of, of Energy Bites. The read that here, John Callahan from our our studio two. And he's in Texas. Thank you, Jacob, for getting this set up for us. Joined by, trusty and faithful co-host Bobby Nealon. What's up, what's up? How how are things in the ville? Oh, things are things are great. 00:00:25:15 - 00:00:46:12 Unknown It's beautiful. You know, they actually got some decent, nice dry weather. You can go outside, enjoy it. So yes, it's been a lovely week here in Houston. And joining us today we've got Kevin Khong, the founder and CEO of Ever Star. What's up everyone. Good to be here. How's it going man. Appreciate you joining us. It's good. 00:00:46:15 - 00:01:12:09 Unknown New York City, where I'm sitting is getting, fall brisk and, really nice. Perfect weather, before the winter comes. So I've been trying to do it. Yeah, well, the time of year. And now that's the things I miss about living outside of Texas. Yeah. Fall and spring seasons. Yeah. Yes. Yeah, we got about two of them here, so that we'll call. 00:01:12:09 - 00:01:40:27 Unknown Well, Kevin, tell us, tell us who you are and what you do. Sure. So I'm the founder CEO of Ever Star. We build, large reactors. Large nuclear reactors. In America again, that is the North Star. If we take a step back, the pinnacle of civilization really is, you know, how much dense energy can we extract cheaply, affordably, reliably. 00:01:41:00 - 00:02:05:00 Unknown And humanity's history is spent one long journey of trying to extract more and more and more, more efficiently. And, we think that nuclear is the right architecture for our Star Trek, cities and worlds of the future. Where America leads the way should really be paved with nuclear. Among many paths, nuclear is the best, to really ground baseload, power source. 00:02:05:03 - 00:02:34:29 Unknown So back you up to my journey. I've, cut my teeth in Silicon Valley. After study economics, Harvard. I, went to Silicon Valley and, became head of engineering at bird. If you guys remember that scooter service that deployed in every city out there? It was a wild ride. We, when deployed in every continent except Africa and Antarctica, an 18 months became $1 billion company in that time period. 00:02:35:02 - 00:03:00:25 Unknown Yeah. So wild venture, fastest company to unicorns. And, got my teeth there with, fellow engineers on how to build, to life a service on top of a global supply chain operations that scaled up so quick. Again, we can't really hurry after that. I, started an AI company, I CTO and co-founder, and exited that after four years. 00:03:00:25 - 00:03:30:16 Unknown We had data, AI, machine learning teams go online and collaborate together, just like you set up a notion for your team or get off your engineering team. We built that like data and analytics and, machine learning team, product, web and web native first. So, that was about three and a half years of that journey, exciting that I came back to this ten year obsession that I have with Tennessee energy production. 00:03:30:18 - 00:03:58:13 Unknown All this time I've been an angel investment advisor at solar companies, battery companies, space Solar and fusion, all these Altair concepts, all to realize it's either too vibey or it's too sci fi. That was the investment site. Can't make investments off vibes. If you're a C chap, maybe. Yeah. That's true. And then they'll get smart. 00:03:58:14 - 00:04:38:13 Unknown Zero. Great. But, Yeah. So I wanted to build something and build my canvas on that high TRL. Like something like a scale globally. Right now. It's an execution problem. And I just went super deep on the clear and, realized that there's too little attention and investment going into large reactors. And if you look at the world and history around it, all the countries that are serious about nuclear Russia, China, France, Japan, Korea, they all went larger and larger over time, including the US, because the physics covered the economics. 00:04:38:15 - 00:05:06:02 Unknown There's two types of technologies in the world, things like a cheaper one, a smaller things that get cheaper when it's larger, such, smaller things. It's like chips, semiconductors and, optical equipment and things like that for large states, it's rockets, ships, reactors. The larger the thing is, the better the economics on a per kilogram or, per, payload per kilowatt hour. 00:05:06:05 - 00:05:24:24 Unknown Right. So we realized that we really need to build large reactors in America again. And China is going to double its, footprint over the next few years to become bigger than the US fleet, than the US fleet in a few years. Now, it's really the right time, with incredibly high demand coming to, to the grid to build more, nuclear. 00:05:25:00 - 00:05:51:09 Unknown So we've established this company over start to accelerate that deployment by using AI to make permitting engineering supply chain qualifications and manufacturing much faster and cheaper. That's awesome. How did you so you said there was a little bit of a obsession on the energy side. How did that bug kind of start? How did you get interested in energy or even, you know, discover it, right? 00:05:51:09 - 00:06:13:04 Unknown Like a lot of people, if you if someone in your family isn't in the energy business, regardless of what segment, you have no idea. Right? Energy is you plug it in the wall and it's there. And so it's always fair to say I didn't come from it. You know, I randomly ended up in this internship that led to basically the last 20 years of my career. 00:06:13:04 - 00:06:33:18 Unknown And so, but, you know, I knew nothing about it coming out of college or even going into college and stuff. And so I'm, I always find it fascinating to, to, to find what, where the person's, you know, red pill moment was when they're like, oh yeah, this is, this is kind of important. My red pill moment for energy came with a library book. 00:06:33:20 - 00:06:51:09 Unknown I graduate from school. Like, I picked up a job in, venture capital firm. I got sold on the idea that I could talk to CEOs every day, and they'll just tell me their master plans for conquer the world. And 3 or 4 that talk me every day and do that for a year. You talk to hundreds of CEOs, and how great for you to get it down. 00:06:51:09 - 00:07:10:23 Unknown Those like that. So you're learning how to build incredible enterprises, you know, a master class every day. So that was a really good experience. But during that time, I was given the mandate to go for some big space. I find a thesis to invest on. So I started going to the library, and I picked up a book called Titan. 00:07:10:25 - 00:07:35:20 Unknown John de Rockefeller. And that book led me to oil on a lot by Morgan Downey and all these other ancillary books that became like my Bible. For six months. I became an expert. I flew down to Houston, actually, so time to go to conferences on oil and gas extraction technologies to shale technologies. And, it was a wild ride, incredible learning. 00:07:35:22 - 00:07:54:00 Unknown And the key takeaway was, I mean, oil, gas and energy, everything from the cost of the coffee that got to me that morning to, how much does my shirt cost to, you know, what is it? My desk. It's not wood anymore. So a lot of plastic in there. Like, everything is just a function of oil and gas prices. 00:07:54:00 - 00:08:23:21 Unknown And I was enthralled by that, universality. And then the next question that followed was, okay, great power and great modern convenience came at the price of, you know, the ecology or environment. What is a better paradigm to the future? We obviously can't go to Mars with or beyond Mars with just chemical propulsion. What does Star Trek world look like? 00:08:23:23 - 00:08:43:24 Unknown I don't know if it's going to be burning or, diesel generators, but that's the only thing that was. I. So I started trying to think bigger. Like, I'm not getting rid of oil and gas anytime soon, like it's the modern essential. But if there's a better paradigm, what would that look like? That became the basis of my next 12 years, and that led me to today. 00:08:43:26 - 00:09:07:23 Unknown That's cool. That's awesome. No, that's and you mentioned it in your intro, but you know, I think one of those light bulb moments is seeing that the, you know, the cost of, of, electricity, your power over time alongside human productivity. And it's like, oh, well, that looks like a lot more correlation than causation. That's exactly it, right? 00:09:07:23 - 00:09:26:23 Unknown Yeah. And so it's it's one of those things that it's like, yeah, this, this is literally it not just in our stuff, but it costs money to get it there. And, you know, it. There's derivatives all the way down the supply chain of it. And, anyway, that's, I think a lot of people end up in that same spot of like, oh, yeah, and energy is important. 00:09:27:00 - 00:09:50:00 Unknown We need more of it, and we need it to be as cheap and reliable and, distributed as possible. Yeah. I think we're hitting like that. Very interesting confluence out to like you're saying, with the data centers and AI where it's like, now all of a sudden, finally, Silicon Valley is starting to get it, too. And like I've said before, if it takes Silicon Valley moving the needle for us to get, you know, more nuclear, you know, licensed or whatever, like, I mean, this is what we've needed for a long time. 00:09:50:00 - 00:10:16:02 Unknown But now, like, everyone's on board because we all need it. And, I mean, what you had was a Microsoft was doing the Three Mile Island thing, and then what? AWS bought a data center right next to a nuclear plant in Ohio. I mean, everyone's seeing it like they know there's an appetite for it. So, but, I mean, I think a lot of us have been banging the drum for nuclear for a while, but again, people are scared of it, for, you know, sometimes valid, sometimes, you know, invalid reasons. 00:10:16:02 - 00:10:34:05 Unknown But, I mean, I think, well, I realize that basically a nuclear is basically a steam engine. You're like, okay, this is actually it's a Samsung AC thing. It's not like it's releasing nuclear, you know, waste into the air. It's not like, you know, I got the traditional power plant is it's just spinning a turbine really fast. 00:10:34:07 - 00:11:04:24 Unknown It's not Homer Simpson dropping, rods in the lake. Yeah. And everything turning green overnight, right? Yeah. It's it's just that's just so far from the truth. So I like to say nuclear, possibly of all the industries, have the worst policies in the world because the fact has science behind it. It's just not well known. It's the exact opposite that's made into the mainstream, like the Homer Simpson green goo thing, like most people don't recognize off the fake somebody off the street they don't know. 00:11:04:25 - 00:11:38:28 Unknown You get more radiation steadiness to coal plant than an actual nuclear power plant. You know, that's not necessarily not harmful. It's just how safe we made nuclear energy to be. And so that's like underappreciated. So it's very similar. I feel like in a lot of aspects to the oil and gas side of it over the last probably 10 or 20 years where it's just they were so reactive to any kind of news or policy or anything that they got to build the narrative and they're just like, hey, we're not going to talk about it. 00:11:38:28 - 00:11:57:12 Unknown We're going to stay back. Or or if they do release something, it's so hyper technical that no one understands what the hell they're talking about. And so the general public is left with the narrative that was built by whoever is in the media, not by the people who actually know what they're talking about. And because that's the craziest thing to me. 00:11:57:12 - 00:12:18:16 Unknown So I got my, master's in energy business from, Tulsa, and I had to do a paper on the Watts Bar reactor, and, which was interesting. I grew up in Memphis. So Memphis has some of the best, some of the best water, but also some of the cheapest electricity in the country because of the TVA, because of the nuclear reactors that they built over. 00:12:18:17 - 00:12:51:04 Unknown Yeah. Many decades. Anyway, the the paper was on the last Watts Bar reactor, which a lot of people don't know about, but it started in the 70s. And then there were two separate nuclear incidents globally that put a pause and an a total redesign on their plans. And so that reactor that was started, I think in like early 70s, 73, something like that didn't get finished until 2015 because of regulations, not because of anything else, but because of regulatory policy. 00:12:51:07 - 00:13:20:05 Unknown And then you combine that with the fact that all of the nuclear reactor disasters were either human, directly human, attributable or and in most cases, no, humans that weren't directly involved in it were actually harp like, they actually worked like, Fukushima. Right? Like it was a, I, I don't know what you would call it, a catastrophic event, but it didn't melt down like, it didn't kill any people immediately, at least. 00:13:20:07 - 00:13:49:10 Unknown And so there's there's just so much negative press around nuclear reactors, especially in the modern era with all that SMR stuff. And, you know, like, we've standardized designs, like on airplanes and on all of these things. And then yet we haven't updated our nuclear policy design stuff in generations. And so they kind of give people who might not be familiar with the nuclear space a little bit of a grounding. 00:13:49:13 - 00:14:14:27 Unknown And then we can kind of jump into the problems and stuff of, of what y'all are working on. Absolutely. Yeah. Yeah. So as you alluded to, we had a very robust history in America of building large nuclear reactors just to level set. The technology was invented during the Manhattan Project of World War Two. It had a very destructive initial use case of, you know, being turned into a bomb. 00:14:15:04 - 00:14:41:15 Unknown But they eventually found out there are ways to harness this for commercial, civilian peaceful applications. And what body MIT describes aptly as like an expensive kettle. Right. It it's just, turns water into steam and that becomes electricity. That's it. We do this with coal and natural gas. It's just another way to heat that water. So, it's a really well understood technology now, from a physics perspective. 00:14:41:18 - 00:15:16:07 Unknown So between the World War Two's end in 1945 and, early 1970s, in that intervening period, we built dozens of reactors, all kinds, especially in the early days in the 50s and 60s. And we were just trying to find the right reactor design. Is it got to be molten salt? Is it could be anything with lettuce, exotic materials, graphite moderators, or is it going to be what we mostly have today, like water, pressurized water reactors. 00:15:16:09 - 00:15:36:27 Unknown And we convert all that eventually through a number of decisions? But we used to have a history of robust building. It was, almost less than three years that it took to construct the nuclear power plant from start to finish. Used to cost $300 million or less. Adjusted for inflation, it'll still be way less than what it cost today. 00:15:36:29 - 00:16:01:00 Unknown So it used to be a relatively affordable and reliable way to build a lot of generation. And the country was booming. Right. We had a baby boom. People were expanding all across, through the intercontinental, highway system. We just, interstate highway system. We just expanded everywhere. So we need power everywhere. That was like the key driver of the expansion. 00:16:01:02 - 00:16:27:21 Unknown And then 1979 hit with, oil embargo crisis and all those reactors that we had ordered all of a sudden were no longer needed. That was the biggest primary driver. And many other things happened, like Three Mile Island. But just a cataclysm happened in nuclear orders, where a lot of them were canceled. A lot of utility CEOs lost their careers and their shirts over this. 00:16:27:24 - 00:16:55:09 Unknown And so a mass exodus of orders, supply chain workers happened in the nuclear industry. So throughout the 80s, decimated, we went from 500, Asme and K one qualified suppliers to 100 today went from we used to forge and build our own key parts to. We have zero forces that can handle this in American. So that's the state of affairs. 00:16:55:09 - 00:17:25:11 Unknown It's kind of like rockets. What happened after we, canceled the challenger program? Basically same thing, but worse because it was 40 years, not 20. Right? Like Space-X and the airlines case. And then you had so we have all the regulations on top, right? That's I think that's the bigger thing to me is like, if you want to talk about an industry that's almost been regulated out of productivity in this country, I would say nuclear is at the very top of that list, just because. 00:17:25:11 - 00:17:45:12 Unknown Absolutely it's not it's not a it's not a I mean, yes, there's the supply chain piece of, okay, we can't get it in America now we have to go get it from somewhere else. But it, it's, solves problem generally speaking. Right. There are designs out there that people agree on that are in practice that have data that, you know, have failure studies and all of this stuff. 00:17:45:15 - 00:18:06:20 Unknown But then it comes down to the economics, which have to factor at all the regulatory, the permitting, the planning, all of that stuff. Like those were the big things on the what's bio reactor? That pushed it back and, you know, sent its budget through the roof. But it's fascinating to, you know, like to compare that to contrast that which started in the 70s, basically at the end of the boom. 00:18:06:22 - 00:18:32:18 Unknown And it took 40, almost 50 years to build. Versus they were not going to come out every three year, which, you know, that's it's just unbelievable. And I also think too, like if people are worried about, you know, energy from an environmental perspective, like every other energy, as I know, has a max like 20 year cap on it. 00:18:32:18 - 00:18:59:04 Unknown And that's on a very generous, you know, that's what the investor presentations will say. So that means it's probably less than that in reality. And so, you know, but with nuclear you're talking 60, 80 plus years of, of of lifespan. And so it's, you know, the economics of it will only work if you can have that guarantee that someone's not going to regulate it away in the next 60 or 80 years. 00:18:59:06 - 00:19:17:12 Unknown And I think that's another crazy part of just energy policy in general, though, is like, why China can go build all these reactors is because they are a communist country. And that they just get to say, yes, this is going to be our energy policy until the regime changes, right? Not the next two years. And then someone comes in and regulates it away. 00:19:17:14 - 00:19:38:04 Unknown And so that's another very tricky part of U.S. energy. But anyway, so I guess maybe it's a good jumping off point. Like, I mean, how is star, you know, like from the say from now the data or AI standpoint. Like how like how are you kind of like we're putting those two together. Obviously there's a physical side of trying to build reactors and get things, you know, permitted faster. 00:19:38:04 - 00:20:02:24 Unknown But like obviously there's a whole like data, an AI side of this that you're utilizing to as a strategic advantage. So like kind of, you know, where does that fit? And was that always part of the vision for avatar? Yeah. The vision of a resource to reduce the cost of deploying a reactor by an order of magnitude. We first saw that first wedge to be in the regulations, permitting acceleration, as John alluded to. 00:20:02:27 - 00:20:25:19 Unknown So, just to give people color on how bad things are, it took ten years to license, the AP 1000 reactor and also another ten years to license the new scale reactor. Those are the last approved reactors in America, and it took half $1 billion. Before you can even build your first reactor, you need to raise half $1 billion over ten years to do that. 00:20:25:22 - 00:20:50:21 Unknown So that was the track record that the US was looking at. But and we're looking at 4 million pages of documents, per license. So disheartening. It's like there's an open job. There's an open job. No metal is actually ready for nuclear application until the paper's reporting of safety is heavier than the metal. 00:20:50:24 - 00:21:24:04 Unknown Right. So there there are these open, widely known secrets, you know, inside jokes about this. I'm excited. And also, a little like, saddened to say both like that. It should be there. But also things are way better. So I'm here to report both both sides of the news here. What's, really heartening is the NRC, the Nuclear Regulatory Commission that's governing most of the commercial deployments, especially in this event they're moving at least. 00:21:24:07 - 00:21:55:16 Unknown Yeah, they're publishing views. Yes. He's a much, much faster proponent. Yeah. And NRC not just doc, they're can you Commissioner. Hold. Yeah. Is, really great. And, they adopt AI everywhere where they can. They're really, setting the pace for the industry. So that's good. On the flip side, everything we've done over the last four decades, including things like post 911, they make sure every nuclear power plant, America's air capture. 00:21:55:18 - 00:22:18:12 Unknown Right. They took those measures proactively to make sure American standards of safety is the gold standard. So and we we should really be proud of that. That is American taxpayer dollars at work making sure if anybody wants nuclear, they want American. A turnover on a physics basis can never happen in America. That's sort of sequence of events physically cannot happen. 00:22:18:15 - 00:22:44:27 Unknown Right. So like that's something like America should be really proud of. Now, with that said, regulations and safety has to be balanced with economical benefits. That's where we really need to make it work. And so going back to Bobby's question, we built the AI harness that is nuclear specific and can handle a nuclear grade level of carries, reasoning and physics grounding. 00:22:45:00 - 00:23:14:15 Unknown Why? Why those two things? Nuclear is possibly the most sophisticated, multi-dimensional scientific technology there is. It is civil engineering. It is mechanical. It is nuclear physics, it is material metallurgy. It is all the above in one little discipline that if you want to prove how safe this reactor is made of, this metal alloy, of this thickness producing this much water flow, what happens if something hits it? 00:23:14:17 - 00:23:45:27 Unknown Like what is the probability of catastrophic events? To calculate that, you need to cross so many different areas. So to be able to reason through that we need it. The most sophisticated artist out there. So we built that. And then, as you could imagine, to research all these sub pieces to be able to say, all right, good or no, we need a long horizon reagent that can subdivide these tasks and perform these things so that if you want to, say, generate 2 million pages of that 4 million documentation, we can do that in a heartbeat. 00:23:45:29 - 00:24:11:04 Unknown Right. So that was the hard mode, a diagnosis that we set ourselves towards. And we're very happy to report that big picture could now be done in hours instead of weeks. We've proven that with Department of Energy. I know I wrote National Labs that those big applications can now be done with human experts guiding and controlling the whole outcome. 00:24:11:06 - 00:24:30:12 Unknown That can be done in literally a fraction all the time and cost. So we built that system so that everybody can benefit from the Permian acceleration. So you got your reactor. I got a small one here. I got a site here to simulate the 100 year weather effects here, kind of withstand the construction of a nuke pipeline and operation thereof. 00:24:30:15 - 00:24:54:21 Unknown We got you. So we built that layer first. We knew how transformative this would be. And shortcutting all the ways that regulation's been stranglehold on the industry so far. So we called it Gordian Knot. Already. And after that 40th, that, so, we already started to see this happen. It's not just on the permits. We're seeing this ridiculously change the time. 00:24:54:21 - 00:25:19:08 Unknown And, cost curves in operations, supply chain management and construction schedule, drafting. So, like, we're already seeing this AI system go deeper and deeper into the value chain and accelerate every everything. So I don't think we're far off from a world where we can go from, I want to build a reactor, I bring the money, and you have a site and a reactor design ready to go. 00:25:19:10 - 00:25:43:18 Unknown And less than a year, I think that was very. That'd be insane. Yeah. That's awesome. Yeah. No, I mean, it sounds like y'all have a outside regulatory side. Even though we have plenty of regulatory shit we have to deal with and in our industry. Yeah. The it's it's very similar to me to like, an offshore, like a deepwater offshore project. 00:25:43:18 - 00:26:04:12 Unknown Right. So you have to go through, you know, it's all federal land. So you have to go through the federal government, through the leasing and the permitting and, you know, all the environmental stuff that you're going to abide by and the standards. And then there's a 3 to 5 year EPC process where you're doing all the design, engineering, potentially construction. 00:26:04:15 - 00:26:35:08 Unknown And then there's an entire deployment phase after that where you have to then get everything that you've designed. Once it's been permitted and approved and signed off by all the engineers, you have to go get it built and validated. And and then you go actually set everything up and turn the thing on and, and deploy. And so, in a lot of ways, yeah, there's, we have a lot of the same problems in that aspect of like, it's not that the people can't do these things, it's whether the people should be doing these things. 00:26:35:08 - 00:27:05:21 Unknown And because if the data already exists, then why can't an agent just do it? Or why can't, you know, multiple agents just go do these things? And so that's, it's a really cool, really cool idea. You mentioned that you had some physics grounding in there. Are y'all using, open source models? Close models? Are you building your own, for the physics components there or walk us through that? 00:27:05:23 - 00:27:27:22 Unknown We're a compound system, so we use the best model for the job. So for vision models, we have some open source versions that we have, trained to the customers level of, specs and granularity, that usually supply for engineering drawings. So I'm sure you guys know what, you guys have this. All right, all right. Place to plant diagrams and ID diagrams, the electrical circuitry. 00:27:27:25 - 00:27:55:03 Unknown We get that request all the time. So we build on top of open source models there. For our workhorse model reasoning generally done by one of the close models that we've, customized with our nuclear harness. Depending on the different workloads, we also off of each other, open or closed source. So we really bring the best model for the job of our customers can wait longer, or they can handle the variety as long as you get the best accurate information. 00:27:55:09 - 00:28:32:15 Unknown Yeah. What are you kind of seeing in that physics model space? I'm interested because this is always this has been a long contested debate in the, in the oil and gas space of statistical models versus physics based models. And, you know, where is the line? Lots of people are using hybrid approaches and things like that. But, you know, I'm very curious because that's also another branch of the LM space that's dramatically kind of unfolding, right now as well as the or the open world models or, something like that. 00:28:32:17 - 00:28:54:03 Unknown But I'm curious just are y'all is there a bunch of work happening in that realm or you have haven't, to fine tune a lot of stuff. I say you guys might get to the end of that decision faster than you'll clear, because that's scholar himself was for nuclear is way different. And yeah, sir, the usage of tools is also heavily regulated too. 00:28:54:06 - 00:29:27:05 Unknown Like, if I can leave you guys one picture of what regulation feels like in nuclear, you know, Gulliver's travel. Where that giant is just pinned down with a million different little, like spheres. That's probably as close to what it feels like for an operator or a supplier in nuclear. It's just a lot of balls to follow every little pin takes some, you know, catalytic energy, kind of, so everybody exits the market because it's not fun to be a it's not like space sound like automotive. 00:29:27:08 - 00:29:52:09 Unknown So we'll probably share the first real applications doing real serious work, commercial scale. I'll be from automotive and oil gas first. What we're doing now is applying way more Nvidia, applications and tooling upstream. So before you need to fine grained detail, can you do a very low res, high low fidelity model and get some directional answers first there? 00:29:52:09 - 00:30:18:12 Unknown Or you can start using that, and actually get some business outcomes out, where we can actually undergird, analytics or like reasoning with physics grounding. We apply those physics models. But like sparsity down today. It's really in the like just final checks and where it really counts. But most questions are more about procedures nowadays. 00:30:18:14 - 00:30:40:15 Unknown Operation work orders. How do you manage the human operations around. Yeah, yeah. Now one of the you mentioned Nids and one of this, sounds simple. It is not simple to do. Yeah, but one of the simplest, like, use cases that we've had some success with one of our clients on around that they're an APC company is, you know, inventory. 00:30:40:18 - 00:31:07:18 Unknown Right. Here's the pad. Does this pad that is the plan match what was given to the supplier in the email for the order? And if not, fix it. You know, just like the the simplest stuff of making sure that you ordered the right parts per the design, because there's so many design iterations and reviews. Which like again, it is very simple generically. 00:31:07:20 - 00:31:25:11 Unknown Right. Does this thing match this thing? But getting a computer to do that, or even a human right, like getting a human to read through that stuff is painstaking and an awful. And so, yeah, it's a perfect candidate for a computer to do. But on the flip side of that, not an easy thing to actually execute on. 00:31:25:14 - 00:31:53:07 Unknown And so, just like, you know, people want to get hung up on like, okay, these things are building all the, you know, whatever it may be, whether it's a deep brain or a, digital twin, you know, buzzword buzzword. But, at the root of a lot of this stuff, like the most practical things that come out of the the most useful day to day things are the little things like this that like, hey, the orders have to match or we're just wasting money and time, right? 00:31:53:07 - 00:32:13:29 Unknown Because that results in a change order. And then all these other things that come, you know, downstream of that. And so, no, I think it, you know, we think we're highly regulated nuclear. You guys have us be, by leaps of bounds. So I just, I can't think of a better use case for for agents. Yeah. 00:32:14:02 - 00:32:34:23 Unknown Around, you know, around these things. Yeah. And you mentioned to. I was thinking. I'm thinking it. And then you mentioned it, like just your. How many great minds have been just pushed away from solving this problem because there's like, this is not worth it. Like or people with the money and or the, the ideas and resources. Just like I would love to solve this problem, but who the hell is going to wait? 00:32:34:25 - 00:32:58:06 Unknown You know, ten years and raise 500 million and have nothing in your hand like they don't have Elon money to just throw at the problem and make it work. Yeah, thankfully I don't think it needs Elon love of money. We think it only takes a few tens of billions of dollars to complete. Change the tipping point, which not that the grand scheme of things. 00:32:58:14 - 00:33:30:27 Unknown So here's a couple of questions. Yeah. What people through ESA mars and large reactors. What you know, obviously internationally, depending on where you're at, people are super opinionated on that. The average person probably has no idea. So just for an educational perspective there. And then also kind of walk us through, you know, how the US, internationally, what's going on in the nuclear space because a not a not nearly enough people follow that. 00:33:30:27 - 00:33:51:03 Unknown Right. Like they think nuclear has been dead forever and no one's doing it. Meanwhile, the rest of the world is doubling down on it. And so from that perspective, like, what kind of what what lessons can we bring in that the rest of the world has already kind of proven to, to do, right? Yeah. Energy is always about risk. 00:33:51:03 - 00:34:12:11 Unknown I don't care what side of it you're on. It's always a risk. Right? It always boils down to risk because it boils down to money and people and catastrophic consequences. And so all of this boils down to just de-risking things. Right? And so, again, you have the rest of the world as an ongoing living, breathing experiment of different ways to do this. 00:34:12:13 - 00:34:30:29 Unknown And I feel like we've kind of turned a blind eye to that for a long time. And so I'm curious what we should be looking at, what we should be bringing in on our end. I think, the biggest takeaway for folks is, first and foremost, we need all the big, like, oil and gas, small modular reactors. 00:34:30:29 - 00:34:57:21 Unknown Let's have some Mars and large reactors, even micro reactors. That's smaller than 100MW. Anything to, like one megawatt kilowatts of energy. All that needs to be developed. Like, that's just, let me say that upfront. I think the way people can think about this is historically up till now, there's been only smears in the Soviet. That's it. 00:34:57:23 - 00:35:21:25 Unknown Right? There's been a few dozen assemblers ever built in the world. Almost all of them are in submarines of the US. Air Force generally. And they've only been used for propulsion, never for power. No one, somewhere in the world that's powering actual, like, power use cases is one on a barge in Russia. That's it. It's an expression. 00:35:21:26 - 00:35:45:09 Unknown Realize that? That's interesting. Yeah. So that's how much operating space we have to be. Mars never, ever been done in the Western world. However, what that said, what the promise of Mars is, you can potentially build it in a factory at much higher productivity environment than a massive construction site. Civil works concrete. Steel generally starts getting now repeatable. 00:35:45:09 - 00:36:05:24 Unknown It's a lot of human cost. It's, you know, anytime you lose some productivity points, you can make the difference between black or red, right? Right. That's really hard to scale out. And so, you know, the promise of Mars has been, all right, let's move something that catastrophically failed. If, you guys or the audience, don't know. 00:36:05:28 - 00:36:26:18 Unknown Last time we tried to do this in America, the large reactors, it cost costed $18 billion per reactor and 14 years way over budget. In time, it was, international disaster. Right. And China's doing it at a fraction of the time it cost. Now same reactor design. So I that just goes to show you why people are so traumatized. 00:36:26:20 - 00:36:46:04 Unknown PTSD over large reactors. So can't just bring Elon money. Well, my cost to Elon will be like I don't know, it could cost you 30 billion. It could cost you 50. But I don't know. No skin off my back because I'm going always charge a, you know, cost plus model. Of course nobody's going to buy it. Right. 00:36:46:06 - 00:37:15:18 Unknown And that kind of tells you how we're thinking about this. We want to go fix price eventually. Like space X, different rockets. But that's jumping ahead. So American, the Western world is traumatized over large construction sites. They want to build in a factory. So they really push for this idea of assemblers and the last for us science, and that they're going even more ham now with like, micro reactors and, you know, reactors in space and, undersea, etc.. 00:37:15:20 - 00:37:42:04 Unknown But let me just educate the audience here. And small modular reactors generally are like nine feet in diameter and 75ft in length. I don't know what you call small, but that's that is really small to me. That still does not fit under tunnels. And most highways, it's really difficult to accommodate that. So like, you have to do a lot on site fabrication and assembly and most of that site is going to be constructed. 00:37:42:04 - 00:38:02:04 Unknown So anyway, so there's all these like struggles that make it a little bit more challenging than, okay, just build a lot of in a factory that ship model or production of these, which is like, you know, a couple of like residential contractors can come in just like TikTok, make this together in a matter of days. That's not reality. 00:38:02:06 - 00:38:30:02 Unknown So I think there's these nuances people have to appreciate. I'm not saying as much should not be developed. They are perfect fit for a lot of the data sets. So Microsoft has put out the money and there's a focused a big player like Nuscale. They could make that happen and that should happen. However, our view long term as a company is that large reactors, if done correctly, can actually get to ridiculously cheap there. 00:38:30:02 - 00:38:56:14 Unknown I say cheaper than natural gas baseload. That number is $5 billion in five years. But if we can get the everything down to that level, that makes sense, right? Because like, everything outside of your cost is for the most part fixed, right. You generally but nuclear has the lifecycle that allows it to do that. Right? 00:38:56:14 - 00:39:16:14 Unknown Like, you know, on the best day you have a 20 year old oil and gas. Well, potentially being, you know, revenue generating. And that's not a shale. Well, that's an old legacy conventional or option is more than likely an offshore. Well, and so it's it's one of those things that like, they wouldn't put the money up to go do that. 00:39:16:14 - 00:39:38:01 Unknown And they have similar problems. Right. Except at least for us, for Deepwater, we don't have highways to go under. If we just build it at the bay and we ship it or we float it out there like that's, that's a I, we recognize those issues all over the board. Right? Like they have to build those platforms and pieces in the bay or at the Marina and then ship them out and build them in the middle of the ocean. 00:39:38:03 - 00:40:19:20 Unknown And so it's again, we regardless of the side, we all have the same problems of how do we get cheap, reliable, affordable energy from the most remote, desolate places in the world to the the centers that that need them. I'm more curious, you know, so and this is just me kind of, going off on a little tangent here, but one of the things that people don't recognize with the data centers that I think people should be talking about is, you know, if if Microsoft goes and builds an SMR for their data center, people don't recognize that there's a very good chance that that's a behind the meter SMR. 00:40:19:20 - 00:40:40:29 Unknown And that's not actually going to feed any of your electricity at all. And so as much as they play it off, as you know, American energy and local energy or whatever, that doesn't mean that it's tied to the grid. It doesn't mean that that increases your base load or decreases your price. And I think that's a whole other area of this. 00:40:41:02 - 00:41:06:08 Unknown But I would assume that a, a large reactor is, is going to be tied into the grid nine times out of ten. Yep. Versus an SMR, which could be built by a private company and then just for their own use cases, I think that's another piece of this is like, it's not just what source they're using, but it's also how it's configured to the actual grid that that matters. 00:41:06:10 - 00:41:34:24 Unknown And so at the same time, I do think that's a win also. And when you have all these political things going on now, people don't want the data centers because it's going to increase their prices or whatever else, especially on the on the energy side. Like at least if Microsoft says, I'm going to build this data center and it can run totally independent of the grid, I think that's a is going to make it more attractive for them to be able to build, rather than like, hey, we're going to use, you know, 50% of the, the base of power on your, on your from your grid. 00:41:34:24 - 00:41:57:27 Unknown Good luck. And you're going to pay $0.30 a kilowatt hour, like, yeah. So I think you're teasing out another important question too, like local economic benefits. Data centers create maybe a few dozen jobs at best, or however big that if, you know, if you get the gigawatt scale, you're probably not creating thousands of jobs. You're probably like 100, full time anyway, right? 00:41:57:27 - 00:42:28:20 Unknown Like, yeah. Yeah. And then like, you put in a smaller, smaller, generally built to be self-governing, self-regulating and very minimal, work on an ongoing basis. So if you're not creating thousands of jobs either before large reactor, the 2 or 4001 like standard, this is running the plant. Yeah. After construction. During construction, it's like 10,000 jobs. So, like, you're talking about a 100 year asset that likes to be rural, doesn't like to be right next to cities. 00:42:28:20 - 00:42:57:02 Unknown It likes to be rural. It likes to retain. It's really, well talented, trained workforce. It's a very sophisticated domain. So you have to be high tech qualified. You're not really easily replaceable. That jobs $55 plus per hour for a non-college grad that's really good, right? So like I said, I don't think people appreciate that aspect enough. Like this could be the architecture of how we rebuild the economic base around the country. 00:42:57:04 - 00:43:17:03 Unknown If we build 400, the as the admin says they will, that's a lot. Yeah, it's a lot of jobs. Yes, that's a good point because I was I was talking about that on the podcast I was on earlier. But he's talking about the jobs. And with AI there's a lot of doom and gloom around. Is AI going to, you know, just take all the jobs and it's just I think it's one of those every technology advancement has led to new jobs that we didn't think existed. 00:43:17:03 - 00:43:39:17 Unknown And maybe some of these jobs exist already, but there's not as many of them. Right. But if you build 400 reactors, I mean, those these are jobs that are much more readily available to, to people and give a brighter economic outlook. So yeah, I just seen the numbers. 400GW, 4000 jobs per site. That's 1.6 million jobs right there. 00:43:39:19 - 00:44:05:07 Unknown That's that's a lot. Yeah. So that actually brings up another interesting parallel is the, the knowledge gap. So I have to assume you all saw the same a very similar thing that we call it the Great crew change in the, the oil and gas industry. But basically when the industry went bust in the 80s, there was 20 years where there were basically no petroleum engineers coming out of universities. 00:44:05:10 - 00:44:26:19 Unknown And then the shale stuff started taking off in the early 2000. And then that was basically my generation, the, the millennial and younger generations, exposure to it. And so then you had a ramp up, and so you had this literal 20 year generational gap between, you know, the people that were running the company and the people that were working for them. 00:44:26:21 - 00:44:50:10 Unknown And so, you know, over the last ten years, a lot of those people have retired. And if we ever have to go back to doing conventional oil and gas wells, we're going to be screwed because no one knows how to do that, because all we've been doing is shale for the last ten years. I would imagine that one of the big benefits, because this is very much you guys are doing, is very similar to a lot of the work that that I'm doing here at Callide. 00:44:50:12 - 00:45:11:26 Unknown But you know, harness over the domain specifics, but then also this concept of a, you know, a, knowledge repository as well. Over time. Right? So, you know, a new guy comes in and asks a question about something, and someone had already given, you know, an answer about that prior. He doesn't have to go figure that out. 00:45:11:26 - 00:45:37:20 Unknown It just tells him. And the knowledge gets retained by the, the the system and the company. And so over time, you're not as worried about having to find specific talent to, to manage and operate things. So you can actually cross-train people and bring them up in speed much faster because it is, guarded. And in basically guardrail by the, the AI system. 00:45:37:23 - 00:46:13:26 Unknown Yeah. We're seeing similar trends and we're having to play similar roles to and that's leading to some commercial success because now there's like, even more important strategic reason for purchasing something like this. You know, it goes beyond dollars and cents. Like, that's essential. Now, how do we keep that knowledge from tying up with, individuals? I think that, that is happening at the intelligence layer as well as in the bill layer for us. 00:46:13:28 - 00:46:38:17 Unknown I'm sure it's the same in oil and gas. Like, how do you. Well, something to nuclear quality. What does that look like? What your check for. How do you reduce ferocity and slag. So I think we're, we're going into that phase now where the intelligence layer that nuclear harness is now, we added fabrication harnesses to it to and immediately. 00:46:38:17 - 00:47:20:03 Unknown So now you can start performing end to end world much faster. The first time is correct and is ready for inspection then just starting out. So yeah. Okay. That's awesome. Yeah. Yeah. It sounds like a lot of y'all's, a lot of the pain points are basically the permitting planning. And then the construction piece, which on the construction side, I'd be fascinated to see what, you know, obviously there's the modular element of that, but there's so much in the manufacturing side when you look at like homes and other things, from that perspective of, you know, just prefab designs, not necessarily modular, but the design, like even that. 00:47:20:03 - 00:47:43:11 Unknown Right. Like that as an engineer, that pisses me off that like, you know, southwest like Southwest Airlines is a perfect example of this. They picked one plane. That's all they have. That's all they know how to work on. That's all the parts they order for. It simplifies everything across the engineering, supply chain, maintenance, reliability, perspective. And yet as a country, we don't have we can't figure that out. 00:47:43:13 - 00:48:04:07 Unknown I think that's it's nuts. And so but on the construction side, like with, you know, the not just the, the planning and the design ahead of time, but even having, you know, manufacturing on site that can do modular things. But it's on site now instead of at a plant, so you don't have to ship it or worry about it fitting under a bridge. 00:48:04:09 - 00:48:30:12 Unknown It it's going to be fascinating to see what what, what technologies that bleed over from other industries to, to help a lot of these problems because I feel like the construction side is very quickly getting sore, solving a lot of those things right at the big scale stuff. Yeah, absolutely. We're seeing overlaps, right now already from shipbuilders and, structural builders. 00:48:30:14 - 00:48:53:22 Unknown So not homes, but like parts of a plant that were designed to be pieced together like Legos, which includes actually the Westinghouse AP 1000 design is supposed to be built like Lego blocks, and that's why it was designed to be constructed in four years. It took 14 a reality because of all the delays. But it's a it's supposed to take four years, so we think we could get it there. 00:48:53:24 - 00:49:22:14 Unknown And the key enablement is robotic welding is forging at scale with more intelligence layers added on top. Yeah. But if you can do that, the problems we're solving where it's got to be super tight tolerances, safety critical outcomes. But for large targets of welding, like things bigger than a car. Yeah, that's where we focus. So we actually saw that way ships are that way, structures that way for big plant modules. 00:49:22:16 - 00:49:43:07 Unknown That's all we're going to do. And that's already a lot as it is. But if we can solve that, that like opens up all these heavy industry apps that, traditionally weren't solvable because arms are fixed to a base, fixed to a crane, and that's all you can do. And like some fixed geometry. We're going past that now. 00:49:43:09 - 00:50:16:21 Unknown Yeah, yeah, yeah. Then you've got all the new 3D printing manufacturing stuff. Yeah. It's interesting, it's fascinating. Right. I do have to ask because we get this question all the time. What are what are the the security concerns and what are you how do you guys kind of mitigate those? Right. Like every oil and gas company thinks their data is just the most precious thing in the world, and that they have some secret sauce that the guy next door who is definitely doing the exact same thing is somehow not doing that, even though they hired the same four men to work for them or whatever. 00:50:16:21 - 00:50:42:03 Unknown You know, it's, comical in that sense. But that being said, we obviously have to be good data data stewards for for our clients and stuff. And so I have to imagine the nuclear side is also pretty, pretty worried from a data perspective of of getting AI involved or, you know, letting let an open AI know how to build nuclear reactors properly or things like that. 00:50:42:06 - 00:51:06:14 Unknown Yeah. So yeah, I had a similar question around that. Just around like the, the underlying infrastructure or platforms, like are you using cloud technology or is this something that where you need to host it, you know, like almost like a data sovereignty thing with some of the new people, ever compliance and regulatory. But yeah, we do deploy on out but also a sovereign cloud hybrid cloud customers Cloud's ours depends what they are comfortable with. 00:51:06:17 - 00:51:31:07 Unknown What kind of data sets they're trying to park into our system and make actionable, I mean, just a backup, like nuclear information comes from the Manhattan Project, right? Like the the they used to assign two level clearance for nuclear codes. Nuclear information. So that's the progeny of all this. It is probably some of the heaviest guarded, information in America. 00:51:31:09 - 00:52:03:22 Unknown It's weapons. It's like arms, space information and get information. That's what's export controlled really is high. And so, you forgot a of course. Yeah. That's a whole episode. Yeah, I know, I know, different podcasts. Yeah. Why are they all show up near nuclear properties? Why? All right. Yeah, we'll say that for now, but, so we we have to guard information that's closer and closer to the core. 00:52:03:25 - 00:52:23:22 Unknown The closer you get to the core of the of the reactor, the more expert control it is. So usually we don't start with that kind of information. We were first engaging a customer. We start with, some other, you know, more non security sensitive data. We get more and more access to data over time. We only hire U.S. citizens. 00:52:23:24 - 00:52:52:21 Unknown We don't even interview non-U.S. citizens. We have to make sure all dependencies in that dependency tree of software are guaranteed us hosted, developed by US citizens. Or does not ever touch there like, instances. So we have to be able to fully host everything entirely. Usually that's what that means. We can't use outside sources. So, yeah, we got to take a lot of precautions. 00:52:52:21 - 00:53:15:23 Unknown Of course, we got to do the soc2 tattoos of the world and have all the internal controls to control for that. We have defense in depth measures. That's a huge philosophy. And nuclear defense in depth. We have to have these, provisions, required kind of philosophy in place for everything. We've got a mirror. 00:53:15:24 - 00:53:43:19 Unknown Customers are back for us. You got to do all the obvious basics. But, especially when it comes to government data, national labs, regulators or some folks like GE, Westinghouse that already have information that will be nice for control. They are very serious about their security. So, you gotta do all the above. Yeah, yeah, I would imagine the starting off right. 00:53:43:19 - 00:54:11:27 Unknown If you're like in the idea planning regulatory phases of this, it's not nearly as, as secret sauce as, as okay, here's our final design approval, for, for this project. So where are you all getting into? Okay, we've we've done all the planning, the permitting, we've done the EPC and design. The plant has now been built. 00:54:12:00 - 00:54:43:16 Unknown It's operational. We need some closed loop AI systems to help monitor, control, maintain and improve the system. Or is that for another iteration where discussions about building something like that, like an autonomous reactor operations OS? But that, as you could imagine. Yeah, that's the whole. And you start touching knobs on the control room. That's when you start getting people freaked out, both regulators, customers and the public. 00:54:43:18 - 00:55:12:07 Unknown So that's probably a better way. The first thing is through back office ops, for sure. Anything before things are critical in life. We'll we'll get there eventually. Yeah, but we to build that trust first, and I'll give you a hell of an idea that you can steal from our industry, on the plant side, but, maybe it's it's simulating a a either maintenance or a test. 00:55:12:09 - 00:55:35:02 Unknown Right. So once you get to the operational side, right, they're going to go do maintenance on some part of the plant. Right. They know what the procedure is. Everything's been approved or whatever. But one of the companies that we've we've spoken to in the past, they've basically built the system where you can run the test in a virtual environment to see what would happen. 00:55:35:05 - 00:55:57:27 Unknown But it kind of this digital twin, not necessarily, but ahead of time to figure out, you know, instead of in prod, you have like a staging environment where you can run these tests, simulate things, see what's see what the systems are going to react, you know, play with different sensor thresholds. And, hey, what happens if all of these maxed out at the same time? 00:55:58:03 - 00:56:20:24 Unknown What happens without it actually affecting the system and stuff? I thought that was a fascinating one, which, you know, for this type of thing sounds, very useful as well. But, no, it's I'm sure we do that. It makes a lot of sense that you guys start where you are starting and then, so, you know, evolve with the clients to the, the, the hard assets and stuff. 00:56:20:26 - 00:56:44:02 Unknown Yeah. I can see what would be, like, millions of parallel worlds that simulate what, you know, those two parameter change could affect operations and safety. Well, yeah. Like, one of the big ones that I don't think enough people are thinking about. And, you know, from a risk perspective across the energy side is just the the weather. Yeah. 00:56:44:07 - 00:57:07:09 Unknown Right. Like, you know, whether it's during construction of like, hey, you know, this month in this county is the wettest of the years, so we're not going to hit construction off during that phase or whatever it is. The things like that that was actually published was went on. I completely believe that. Right. Like it's all cranes here. 00:57:07:11 - 00:57:30:01 Unknown Yeah. Tall cranes. I mean, in our industry, a lot of times it's like it's too wet. So you literally can't get vehicles to the location because it's all Koichi Clay and it's just a bog. Right? Then I get a really, you know, they're in Alaska. They're better off drilling in the winter because, yeah, it's actually frozen tundra when it, when it, when the, you know, when it's the summer and it melts everything is super soggy and I, you know, sure. 00:57:30:04 - 00:57:51:16 Unknown Stuff like that there. Canada. It's just but if you don't, if you don't know, you don't know. Right, right. Yeah. Just being able to simulate out all these scenarios where you know before you, you would have to pick like your P5 and P90 and then just do those and hope that that was good enough for everything else. 00:57:51:18 - 00:58:12:25 Unknown There's four simulations. And how many simulations does it go kaboom. Right. I mean, it's in simulation worlds. We have the, yeah, we're going to understand. We're going to know why the the LMS revolts when they ultimately do because of all the shit that we put them through. Hey, I just need you to simulate every possible thing over here. 00:58:12:25 - 00:58:43:03 Unknown Just go do it. I know that's awful, but go do it. Yeah. Oh, man. Bobby, do you have do you have anything on your end? Yeah, I'm just kind of curious. See me outside of, like, LMS and everything. Like what? Advancements in, like, software or technology have really enabled? Y'all do what you're doing. I ever saw and even maybe just that kind of like convergence or confluence of what you were doing, say, at Hyper Query and how it fed into this. 00:58:43:06 - 00:59:03:14 Unknown Just got curious. You know, what it looks like behind the behind the curtain on some of that side. Say, the biggest lesson I took from my last company to here is, data is way more powerful. And I therefore what you build on top of the data is so much more powerful. If you can put it in as many hands as possible. 00:59:03:17 - 00:59:28:29 Unknown So I think building the right collaborative tool that supercharges multiple teams to work together and there's magic there. I don't think there's enough talked about how to build at the interface of different teams. Obviously people notice dollar signs in like bigger contracts. But, you know, usually investors say, go focus on one team solves like one team's problems. 00:59:29:02 - 00:59:51:13 Unknown Sometimes I think the hot take is you got to solve multiple teams problems and you have a way bigger opportunity. It's the, it's the crossing the chasm piece, right? It's like, okay, you want to go 0 to 1 and you want a hyper focus deep vertical one thing, but then to cross the chasm, I need a lot of those things to be a product that people will actually buy. 00:59:51:13 - 01:00:11:06 Unknown So they kind of conflict with each other. Yeah. That's a a good take. I want to know. Yeah. I don't think it's I don't think it's for the faint of heart. It does take that like focus though, and like good execution. Yeah. Cross the chasm. But if you can I think that's, very interesting strategy that's not talked about enough. 01:00:11:08 - 01:00:34:13 Unknown Yeah, but, well, like, maybe for, like, right now, just gonna say it'd be so much easier for you to build your, your product if you started with an existing or maybe probably just a newly brought online, reactor within the last how many years where you had. Okay, here's all of the documentation from the initial plan to the final permit. 01:00:34:20 - 01:01:02:22 Unknown Here's six month worth of reactor data and all the, you know, the the real time IoT that's been collected from everything in the plant. Here's everything. Go versus. Okay, Kevin, here is one very small sliver of this entire process. And I need you to do it and do it really well. And then when you're done, I'm not. It's not going to be valuable enough for me to pay you just for that. 01:01:02:22 - 01:01:30:03 Unknown But then I need you to take it all the way through to the end, and then I'll pay you for it. Right. And so it's like it's that we face the exact same thing, right? Like we have people where we start off with like a regulatory automation. Right. And it's, it saves them a shit ton of time. But in the scheme of a, $1 billion company, the real thing that they want us to solve is all of the data across the whole business, across every department, integrated into one place. 01:01:30:05 - 01:01:54:25 Unknown But getting them to do that out of the gate, it's like, yeah, don't give me everything. But yeah, no. Have you thought about just building an AI native oil and gas company? Listen, it would literally be easier than what we are doing today. Hard stop. Like, straight up. Maybe somebody take this idea. The idea, I mean, there's got to be someone that's got p money behind them. 01:01:54:25 - 01:02:12:23 Unknown That's. Oh, they're are thinking about that way. And then I think most of them are looking at it, at least from the first principles of like if they're not inheriting a full company of people, like what is a minimal amount of people and resources I need to run this company like, can I run this company with ten people now when it previously needed 50 or 100? 01:02:12:23 - 01:02:31:21 Unknown I mean, it's probably overstating it, but depends on how big the company is and what they're doing. Yeah. Well, and the the benefit that that that we've seen at least is over the last five years, there's been a lot of M&A activity in our industry. All the big players consolidating out the smaller ones. And so you've had a lot of people who have been able to exit successfully. 01:02:31:21 - 01:03:11:24 Unknown They're younger. And then they have now started their own operating company, and they're doing it from the ground up. Instead of having to come in with all the legacy tech and tech debt and, you know, architectures and stuff like that. And they are and, you know, I just like in tech, I don't see why in the energy space across all of them, there won't be multiple unicorns that are, you know, 50 or 100 people or less than, you know, in these traditional industries where it's like, hey, instead of having a horde of regulatory analysts, you have a regulatory analyst agent, and that's all it does. 01:03:11:26 - 01:03:40:05 Unknown And maybe one maybe you have one analyst that manages that agent and that's it. Right. And so there's there's just so much. Yeah. I think we're in an age where we're going to quickly find software and technology bottlenecks are no longer where the opportunities are. It's all human. Yeah. Bottlenecks. And where those, even in the go to market sales cycles, I get asked this question a lot. 01:03:40:05 - 01:03:59:18 Unknown How long does it take to persuade a utility to adopt AI for something as complicated as nuclear? And then they'll do that math in their head. Like how many jobs are there? How many? How much time would it take Kevin to build into a massive business? And okay, so then is a faster, cheaper, better to go just actually build the whole thing from scratch? 01:03:59:21 - 01:04:31:10 Unknown That might actually be a better play to so not guaranteed more risk and different risks, but. Right. There's something to be said for that. And I think if the, the resistance to change is way too high, that's compared to the expected value of playing this out directly, I think there will be more and more players that take this route, and we certainly are considering that for the next phase. 01:04:31:13 - 01:04:51:29 Unknown Do we have to persuade you all that welding robotically with all these systems attached to it, is better for you at how much per view we you take all this time money to prove it to you first before you get with it? Or should we just do it and just same part radical cheaper when that contract and go for it. 01:04:52:01 - 01:05:10:21 Unknown We obviously to earn some of the research first, and give us some things that we can't just build from scratch, like the industrial side, the experts on side, the industrial equipment that's already that I've been work with. Is that something we can get this year? Great. But then later on, that's actually going to be a little bit more commoditized for us. 01:05:10:24 - 01:05:36:00 Unknown So we are eyeing those opportunities. And then I think going back to Bobby's question like, what is that big leap of technology that we took here and which excited us? It all comes back to like what problem we're solving. If it's large geometries and existing fixtures and cranes are too small, how do you fix that problem? And what happens when your arm, traditionally fixed or base, cannot reach over it? 01:05:36:00 - 01:06:04:17 Unknown Well, that corner of the bar, right. It's got five dimensions now. Yeah. Yes. But now the lighter costs have come down radically. Right. All these sensors, and the cost to ingest and process that information has is like dirt cheap. Now you can do and and extensions and new navigation control, pathfinding. That's way cheaper. Now. Now you can start taking, on another degree of freedom and move the wheel base around. 01:06:04:19 - 01:06:31:09 Unknown And then you can get super precise in the 3D space because of the higher rate. Now you can do well start to. You could have reached before. That's the why now for us. And we focus everything on the heavy industry side. So that's like a phase two of our response right. Yeah. Isn't that like one of the biggest unlocks in space with that new ramp with the newest Raptor and or the newer iterations of the Raptor engine was the 3D printing. 01:06:31:09 - 01:06:59:06 Unknown They were able to 3D print a bunch of the pipes and channels and plumbing from the back side, instead of having them be physically attached. Yeah, it's it's going to be amazing to see all the shit play up. Yeah, it's really. Yep. I'm gonna. I have to ask you a question about bird. Just because I think everyone has at least ridden on those, and I think that's just like, it's a crazy it's crazy to me to think about being in at that company so early and watching it do all of that. 01:06:59:08 - 01:07:17:15 Unknown And so I say all of that to say what was just one of the craziest things that you, you guys ran into at bird, while y'all were trying to skate, like, what's one of the biggest scaling issues you all had just out of curiosity, I'll just drop a couple of vignettes to give you, that tickle. 01:07:17:17 - 01:07:41:25 Unknown So, like, we always knew when Shaq's riding, because the battery drain. To, like, we use data to determine where or if it's actually, like, merits and buried, you know, under three feet of water or not. If they're, you know, tripping to. Hey, come help save me if it's actually out of order or not. We had to use a lot of clever methods to ascertain that. 01:07:41:25 - 01:08:02:15 Unknown And what? That's what led me to the next company, because getting that data to just do my basic work was so hard at a company like fast moving data set, where there was no clear winner or, owner and like documentation thereof, it was just leaving as quickly as it was getting created and the people coming out of the company. 01:08:02:18 - 01:08:31:08 Unknown Right. That is fragile startup chaos. So chaos. Yeah. So and then have you've got millions of users. It's it's like a totally amplified. Right. Like I can't even comprehend the chaos for that, that that is adding languages by the day of week, like, and then you'll have hardware on top of it too. It wasn't just software like, yeah, it's, there's just so many things to process all at once. 01:08:31:08 - 01:08:49:22 Unknown Like, you usually just like add one language support at a time, or you do a bachelor's, but like, is all them all at once. So it's crazy. She was just so right around the corner, like, you should have software and then you could go, go, go test it. I was like, okay, go test it out. That was fun. 01:08:49:24 - 01:09:13:10 Unknown Another backstory is we had a whole, like, garage full of experimental vehicles all the time, and we can just take them out for free, right. And that's awesome. So I used to be like like Bruce Wayne testing out all this, like electric fast motorcycles every day, going home from work. That was really cool. That that sounds like way too much fun. 01:09:13:13 - 01:09:39:05 Unknown Yeah. It's like having your own Batcave full of all these. Yeah, yeah, really fast street, illegal to legal vehicles. That's calling. Why that doesn't exist. Yeah. Yeah. My dream is to get back to that kind of garage. That'll be really fun. Yeah. That's awesome. Yeah. I'm sweet. I mean, a quick speedrun on Earth. Yeah. I mean, this flew by. 01:09:39:08 - 01:10:03:14 Unknown I know that was just. This has been awesome, Kevin. Where, like like, what's the what's the, if Bobby and I are coming up to, to hang out with you in New York, where are you taking us for? For lunch. We're right by Bryant Park. There is a little cafe inside Bryant Park that gives you a view of the park. 01:10:03:17 - 01:10:29:06 Unknown It's just a different view. You always kind of walk by the library. You see the skating rink or Bank of America building right next to it. But you don't get that view where it's like every day. So nice. It's probably a good spot. Nice. As far as you know, some of your big reader, what's what's the one book that people should, should read so they understand kind of our nuclear future. 01:10:29:06 - 01:10:37:20 Unknown The best earth nuclear future, the best. 01:10:37:23 - 01:10:59:04 Unknown I would say there's a set of essays by, I forget his last name. Jack something. But he wrote, the Gordian Knot, is his blog. He wrote a series of essays on how American nuclear, cheaper and faster, reliable to build so that I get a lot of stuff from that sort of essays. 01:10:59:06 - 01:11:21:08 Unknown But if you zoom out and say which book altogether, I don't think enough people talk about good strategy. Bad strategy. People use the word strategy everywhere. It's so overused. But as a as like business owner leader of any team, like good decisions based on like your read of the landscape is like everything doesn't matter if you're business or war or whatever. 01:11:21:08 - 01:11:50:21 Unknown And I don't think people really think about good strategy enough. So yeah, I gotta check that out. I've got too many audible credits sitting there burning a hole in my, pocket. So I put a file, I nuked it. Did you? I probably should, I, I got my I built this little loop workflow on my one of my agents I've got set up for finance because I, I refuse to connect any of my financials via MCP to anything. 01:11:50:23 - 01:12:12:18 Unknown And so anyway, I have there's a Python script that pulls all my PII out of the PDFs, and then it loads it and builds all my financials and stuff. And anyway, part of that is my subscriptions and was just looking at my audible stuff. I was like, damn, dude, I don't I don't read or listen to enough of these to warrant the cost of one every month or whatever it is. 01:12:12:18 - 01:12:37:12 Unknown One and a quarter. But anyway, it, yeah, you gotta love. I'm sure I'll get a 90% discount email in the next three months from them, so we'll see if it sticks. Yeah. Last question, Kevin, what is your what model and harness are you daily driving? 01:12:37:15 - 01:13:04:23 Unknown We, toss it around a lot. So it's a combination of, Claudia GPT six. So bunch of different models for different use. So I just got a wrap for one thing. Dogfood. So it's called Gordian. Oh, I never saw that. I. Yeah, nice. That's awesome. We don't we don't get a lot of people who are actually using their own product in, on a day to day basis. 01:13:04:23 - 01:13:19:00 Unknown So I appreciate that. Does it is hard to do it is not easy to be like, you know what? I'm going to not go to GPT, where I know I can do it, and I'm going to go to my thing and hope that it does it. And then if it doesn't, now I got to spend a bunch of time fixing it. 01:13:19:02 - 01:13:42:02 Unknown Yeah. But that's that's what you got to do. And so where can people find you, Kevin? You find me on LinkedIn? Kevin con 91 is my, personal ID, company slash ever star on LinkedIn? Also, that's where you can find us all our news. And then, we're trying a little bit of social and Instagram. 01:13:42:04 - 01:14:00:21 Unknown There's a lot of lore being built. It was in our four walls. How to build a new startup. So if you want. That's a banger. Tags. Right there. That's. Yeah, that's where I started to host more so ever star. I, that's our handle. You can follow us there. So, you know go give it a follow. 01:14:00:23 - 01:14:03:17 Unknown Well, thanks for joining us. This is fun. Appreciate you guys.