0:00 We're back with another exciting episode of War. It's a born, it's a born supporter of John. We're trying to fight against the superlatives. I'm an engineer, man. I'm not good with the English
0:12 language. I mean, I was a basic kickass. So, I mean, that's kind of better supportive than what I was able to pull out. But yeah, no, I think, seek for both of us. We're really excited to
0:22 have Rudy Lachabera here, the founder and CEO and janitor and developer of a petrobrix A lot of hats, man, a lot of hats, thank you. Thank you. I was glad when I heard you're coming into town,
0:35 I've been wanting to get you on here for a while, and it's always better to do it in person than it is over the Zoom camera, so. Well, thank you both, I'm glad we can make it work. Yeah, and
0:44 Rudy's down for Houston Energy and Climate Week. Yeah, I think Jacob ran into you on Tuesday or Wednesday, whenever, and then I ran into you yesterday, and so it's been good. It's nice to sit
0:57 down and have some time with you. Yeah, the whole week, it's funny compared to Colorado, where I'm from, base in Denver. And there's so much more going on down here, especially around oil and
1:08 gas, but even just the startup scene and the tech scene, you got a lot going on. No, it's pretty cool. I mean, I got, what is it, a month or two, I was walking around the cannon. I was like,
1:16 man, I need to do this every week. Like just the people I was bumping into and just the stuff that was going on. And it was, I didn't engage that much, but it was just like, no, it's very true
1:25 If you're just going to the cannon or the ion and just like walking around and seeing all the, it was just very energizing. I had a lunch meeting at the ion with Foresburg and Nancy earlier this
1:36 week. And I was just like, oh, this is cool. Yeah, it's like, I hadn't been to the ion since they built it. And so, I needed to get over there too. Empty the last time I was there, basically.
1:44 But it's - Yeah, the ion's fantastic. And
1:48 I, I might be wrong about this, but I thought I heard some people saying that they're gonna expand that area and they're building a whole innovation district. Have you heard about something like
1:58 that? Houston's trying to do that. One, it started with that stuff, and then now you've got all the tech companies throwing money at data centers and different things that they're gonna build in
2:07 Houston, and so it's nice that you're getting not just energy tech. But, well, I think even with medical tech, I know what, 'cause AM and Methodist partnered up on a medical technology,
2:21 you get your MD, but it's like a medical engineering, or something like that, too, recently. So, I think a lot of cool stuff's gonna happen in here, for sure. I'm glad, 'cause Houston has
2:30 needed it for a very long time, I feel like. So, yeah, I mean, I think I'll differentiate us with Austin, even in all this stuff, too, where I think people are starting to come around a little
2:38 bit. But, yeah, now if someone figures out to remove humidity, they'll have it solved Yeah, humidity and traffic would be golden. Maybe the carbon capture guys can figure out the humidity
2:50 problem too. Yeah, they're sucking all the air out of the. But we got to stop producing all the carbon dioxide in here too. But,
2:58 well, awesome. So Rudy, you're here with Patrick Bricks. What the heck is Patrick Bricks? So, Patrick Bricks, it's data connectivity software for oil and gas. And it's basically a solution for
3:10 that problem you run into a lot of times where you got 20 different systems that everything needs to work together for let's say a production engineer to do their daily operations. And so, Patrick
3:22 Bricks is kind of a connector that has pre-built components that talks to all those all those different systems and then you can create data pipelines by just kind of snapping those things together.
3:24 So, it's kind of like a five-tran or zapier make for energy. Is that
3:30 kind of
3:38 the idea? Yeah, it's funny. So, that's right. You're one of the only people that I talk to who's sort of a zapier. Is it Zapier or Zapier? I have no idea, it's one of those. But crazy enough,
3:52 that was like the original vision was, I kept having, you know, I was thinking about like the next kind of sensor thing, sensor data play that I wanted to build something around. I'm like, yeah,
4:01 there's all this important stuff going on with spark plug B and MQTT. And then, you know, I figured out data structures and architectures that I thought would be kind of a generalized data pipeline
4:10 solution. And I kept running into people, trying to talk to people, I kept trying to explain to, you know, operator service companies, folks like that, what I had in mind and the response I'd
4:23 always get is something along the lines of, I don't know what you just said and I really don't care.
4:29 But I tell you what I do care about. I've got these two systems out here and I need to pull data from them twice a day and I need to put it in my backend system. And I also need to, you know,
4:38 convert the data to my common data format that I use in my backend system Can you write me custom software to do that? And I would always tell them, no, no. You don't want me to do that. That's
4:49 the worst way to solve that problem. And after about the third time that happened, I finally got a clue. And probably more importantly, I looked at it from their perspective. I'm like, well,
5:00 what's a good solution? So it's all well and good to tell them, no, I'm not gonna do that. That's a bad way to solve that problem. But yeah, what would be the good solution? And I realized that
5:09 there really wasn't a very good solution. And we probably needed a Zapier for oil and gas where instead of talking to social networks and Twitter and Slack, you talk to Enbase and Py Historian and
5:22 all the systems that oil and gas companies care about. Yeah, so for those, I guess, who don't know Zapier or Make, are these kind of low-code platforms that have kind of blown up over the last
5:34 five years or so, where they have a bunch of, it's basically managed, it's APIs as a service in the simplest form.
5:45 APIs from all these different softwares and platforms. And then you come in, you connect to your platform, however you do it. And then you can start doing things with the data from that. So you
5:54 can pull it and then transform it and then push it somewhere else. You can write it to somewhere else. You can do all kinds of stuff with it. But we use it all the time for a lot of our marketing
6:03 and backend stuff. And so it's incredibly valuable. For the longest time, I was like, I tried to avoid using it just out of principle because I'm like, I know I can write this API, but then as I
6:15 got more mature, I was like, I know I can write it, but I'm not gonna maintain and manage it. So it's worth paying the money to just let them do it, especially when you get enough of those lined
6:25 up, you're like, I could have a whole person just managing APIs all day long. And so I think it's an incredibly needed thing in our industry. So I'm really glad you are building it. Yeah. So but
6:37 let's talk about, 'cause I mean, you brought up the five-tran part too. I think it's kind of a, it's a very extensible, platform, right? Or like the concept. So I mean, there's a lot of
6:46 different use cases here. I can think of, all right, we can just extract and load data, say from all your different sources into your centralized snowflake, SQL server, Databricks, whatever,
6:55 right? You know, you can do that, which people are doing before with scripts or other services, right? But then then, but then you can do your stuff there, but you could also push it from those
7:04 sources back out to other sources, right? After it goes through the sausage machine, or I think petribricks can be the sausage machine as well, right? Like it can pull from one API, do some
7:13 things to it, and then put it into, you know, it can get from one and post to another, those kind of things, right? So, I mean,
7:21 yeah, absolutely. I think of it as a data automation platform. And, you know, the idea that is that it's doing all the Zapier stuff, but it's also connecting to things that we care about in the
7:35 oil field, and maybe you don't run into if you're you're doing marketing? Yeah, there's, it's the biggest downside of, of make and. It's like, if you want to interact with any actual industry
7:46 tools, you just better hope they've got a database connection on the back end, otherwise you're not getting it.
7:52 Yeah. So, like, Petro bricks, we'll actually have automations where they can start off by doing the API connection thing, where you're connecting to some system, to their public JSON API, or it
8:04 could be
8:07 MQTT, or it can be a SQL query against a SQL database It can be an FTP server, like we run into SFTP and FTP quite a bit still. No way, not in this industry. Sorry.
8:20 Yeah, I'm, I'm red-pilled now, I mean, you know, unless a tool like this works really well against APIs, like start posting your shit on the CSV to it FTP all day over some of these awful APIs
8:31 that companies have. Yeah, I've learned, I've learned after being, you know, when I was younger, I started off pretty critical of stuff that I thought was done. And over the years, I've
8:42 learned that, oh, there's actually a reason you're doing that. So yeah, I probably should have a little more humility when I'm talking about, hey, that makes no sense while you're still doing
8:52 that. But yeah, there's a lot of stuff like that that you run into. And more to your point, Bob, you started talking about more automation stuff. You know, it's just as easy to push into an API
9:02 as it is to pull from an API. And, you know, there's, it depends on it You know, the API on the receiver side has to be there. But if you've got, you know, like an API to change a
9:15 configuration on an ESP or something. Yeah, Patrick Brooks can actually, it's got a whole eventing system. You can inspect the data going through the pipeline. You can raise events based on, you
9:29 know, conditions that might happen in the data stream. You can then like send out commands to an API through MQTT. If you want to do spark plug B, that's another way you can send commands to
9:42 external systems. So yeah, you have the ability, you know, the first use case is always, let's consume data, let's get all data exists in like 20 different places. We need it in one place where
9:53 we can look at it and kind of manage operations. So that's the first use case. But absolutely, the next step is going to be let's start doing real automations Yeah. And what's about where does
10:04 petro bricks live? I mean, I think we're talking about Zapier. That's a sass. I log in and I keys my Google email with SSO and all these things. But like, as I understand it, you can, petro
10:13 bricks can live in the cloud. You guys can host it. It can live within their firewall or it can live on the edge, right? Yeah, we're actually not hosting anything yet. I think that's going to
10:21 happen as we go to smaller operators. Sure. We're going to run into folks who don't want to host their own petro bricks node. But really part of the key design is petro bricks is meant to solve
10:31 this problem for oil and gas. And I think that requires that you have to run on-prem, especially with security concern and high firewalls. Yeah, you can't be sending all your sensor data off to
10:45 the cloud or opening up a hole in your VPN hole to your private skater network. People are not going to do that. No, they're not doing that. The Purdue model going, they're only going to push out.
10:56 You can't - I don't even think with petro bricks - so we got a lot of folks who are basically dropping a petro bricks node on their private skater network I think even that violates the Purdue model.
11:07 Fair enough, yeah. But yeah, it's - I guess that's within it. And it's acquiring data and then pushing out, as long as there's not an external one that's in, then I think. Right. But no one
11:19 asks me. Yeah, so yeah, to
11:23 your point, yeah. I think it's really important for oil and gas that our customers are upstream operators So I think it's really important. for those customers that they be able to run this type of
11:37 data automation where they need it. And in some cases, you even have access to unprotected systems, and you know, because they're on a private skated network that's never going to see the light,
11:47 you know, no one's ever going to see it. So, yeah, being able to be able to drop something with those those data automation capabilities on a network like that that just has access to everything
11:58 it needs access to makes everything work much, much better. Big time. Yeah Now, that's, yeah, that makes a ton of sense.
12:06 Yeah, because I mean, even then you're, but you're talking about, I mean, I guess all the protocols, I mean, can you even talk almost directly to the PLC level or mod bus type things as well?
12:14 Or have you looked at that or? There's, so, you know, the answer, could we do that at some point? Yeah, it's, that's the nice part of it, you know, being custom softwares, you can build
12:22 whatever you need. But we intentionally try not to,
12:27 well, we don't want to reinvent the wheel either. There's good solutions for that already Okay. Where we find, the closest that we probably get to that is, there's a lot of stuff going on with
12:39 spark plug B these days. And so yeah, we can definitely pull data from and push data to spark plug B. Okay, can you speak about spark plug B for people who may not be, you know, or try to dumb
12:49 it down as much as you can? So, or why it's better than what existed? So by the way, I have to put the qualifier in software guy, Maysos, not a real oiling desk, let me say something dumb. And
13:02 I'm also not the, you know, my expertise is not an operational technology, but yeah, spark plug B, with that said, spark plug B is basically a
13:15 standard for both messages that you're gonna send, the format of those messages, and the topic space, the namespace of the topics that you're gonna publish to on an MQTT broker. And by doing that,
13:30 you know, MQTT is it's just a really basic pub sub. So there's, you know, that's great. But if you want to do like a SCADA system, you have to maintain state, you have to track devices and edge
13:42 nodes as they come online and you know, that kind of thing. Yeah, that's fun. And so yeah, Spark plug B basically is the standard that just through being really smart about how they pick topics
13:52 and really smart about the format of messages, it kind of allows you to provide that SCADA type functionality where you're doing all that extra stuff and you're maintaining state and you're tracking
14:02 devices with just a simple pub sub. That's pretty slick, nice. What are the kind of most common integrations that you're seeing right now as far as like between what softwares or different tools
14:14 and platforms? So, you know, we kind of have a boiling the ocean type problem where the platform can do anything, but to make it actually useful, you have to kind of, you have to start somewhere.
14:24 And so we've started with production engineers
14:30 Uh, a lot of the initial use cases, uh, have been methane emissions, methane emissions data. So you got all these new companies like, you know, they have a, a new IOT, uh, methane sensor
14:41 platform, like Kuva or cube or, uh, Earth view and, uh, there's no, there's no wired skated connection. So how does an operator get that data into their historian? Uh, so that's, that's a
14:54 good entry point for petro bricks where we kind of have a turnkey solution for that Yeah. That doesn't exist really. Doesn't exist. Or it requires, you know, really it's requires, yeah, that
15:06 makes sense. And so once, uh, the interesting thing is once folks see how that works for something like math animation, say, realize, Oh, wow, you know what, you know, it's really important
15:16 that this would do the same thing for is, can you get us, uh, compression data for gas lift? Yep. You know, ESP is a big one too. I feel like everywhere I go, we were even doing an RDS like,
15:26 can you, can you hook into my ESP and like. stream that data like because some of them have there like you've got 10 different vendors with all the different APIs and like and they just want it
15:37 consolidated into one standard format. It's not what it's time stamping. Yeah. It's
15:45 the vein of our industry's existence and all honesty is having 20 different softwares with all the data that you need in them, but none of them talk to each other or none of it's unified in a way
15:55 that you can actually do something with it. And that's, there's such a need for it. And one thing I just thought of though too is like something like this where say you can normalize multiple
16:05 vendors that provide the same thing, it gives you a lot more, I would say leverage on a supply chain side, because I mean, I know some people they've built, you know, in a smart, say on Liberty
16:15 Sider from folks, you know, where they've got, you know, they're standard. They're standard and like people get used to consuming it that way. And so then it's like, man, is it really worth us
16:23 negotiating with, you know, Liberty whoever else like on the frack side, because like, I know I want to get the data this way. And if I switch, now I got to rewire everything, but then you give
16:33 up a ton of leverage on the pricing side, right? Yeah, Liberty's great about that. They have good standards. Yeah. Oh, that's no. And we had what Brian Weaver on bottom, probably about a year
16:43 ago or yeah. Yeah. Unfortunately, not many people are very good with their data standards on the field services side. And that is a big problem, at least historically speaking So it's, but it's
16:56 one of those things that it's like, it's solvable. These things are solvable. And I'm glad there are people like you that are working to help solve these problems. So an interesting thing occurs
17:05 to me, you're asking about use cases. Tell me if you've ever heard about these kind of problems with gas lift in particular. There's kind of this new, there's kind of this, I don't think this has
17:17 been a problem in the past where suddenly operators want data from surface compressors. There's a SCADA connection to the Surface Compressor, but that's owned by the service company. And the
17:30 service company wants to make the operator happy. So they'll make that data available, but they do it in a way where you need a programmer to get at it. You know, Event Hub or JSON Web API,
17:39 something like that. And so I don't know if this is because we're suddenly realizing we want that data or
17:48 what the situation is, but I've seen this coming up a lot where operators, they're either trying to develop custom software to get at that and pull it from the service companies. Or a lot of them
18:00 will actually, they'll have an initiative to get a wired SCADA connection going to all their surface compressors. Is that something that we've seen that you've ever heard of in the past where it's
18:10 like, it's not even a new system. Right, it's just somebody else's. Yeah, you suddenly realize, oh man, the service company actually owns that data, or they have the wired SCADA connection to
18:19 that data, not us. And yeah, in order to get it, we're going to have to do something. Yeah, I mean, I feel like that happens all the time. I mean, there's a company detection technologies out
18:28 there that, you know, that's their whole jam, pretty much, is like the compressor side, but it's becoming a big deal. I mean, I mean, whether the GME or different people I'm doing work for now,
18:37 I mean, the compressor stuff is coming up a lot. Well, yeah, if you can produce it, but you can't sell it and take it even optimizing the compressor, there's all that stuff too. Like, are they
18:46 over-capitalizing all the electricity or, you know, there's a lot that goes into it. So man, it is fascinating to see how granular we've gotten with the efficiencies that we're trying to look at
18:57 in the field, never did I think they'd be like, oh yeah, the compression is where we need to start optimizing things, but these are getting in like the LOE. analysis. Yeah, no, I mean,
19:05 spending a ton on that makes sense. It's capital expense, right? Yeah, and it's a lever on production, so that's nice. Mm-hmm. How did you get the idea for petrol brakes, or like, let's back
19:18 up a little bit, how did you get into a kind of tech andor energy, Can I go in back?
19:25 coding or taking stuff apart, or we've asked this question before, but what was your gateway drug into coding or tech? So I was a software guy, and I
19:38 actually went to school for aerospace and physics, and those are two terrible degrees if you want to get a job.
19:46 You didn't want to go work for Lockheed somewhere? I did want to go work for Lockheed, but
19:52 it's interesting The aerospace industry is kind of similar to the frack service company side of the business. Whenever
20:04 there's a cut in the commodity price, they may fire a whole bunch of people, and there's no jobs to be had for a few years, and that's what aerospace is like. With every new administration, if
20:13 they cut NASA and they cut defense, forget it, that's no job. That's very true So I caught the wrong side of that cycle. And was working in grocery stores and that's kind of what got me into tech.
20:26 I'm like, you know being a programmer That's probably better than stock and Coca-Cola on the shelves. Yeah
20:33 So I did that and it was kind of a self a self-taught Programmer, I did do some programming stuff in college, but I really I didn't I really didn't come out of school with those skills and What did
20:46 you start when you started kind of getting into it? What what were you programming in
20:52 I? Think I was doing Pascal visual basic and for Tran Nice Yeah, I was gonna guess be visual basic ever sure but
21:03 I feel yeah So many people have touched visual basic at some point in their lives It's kind of amazing actually yeah, and then for Tran is Hilariously common as well like people don't think that that
21:13 would be a common thing But one of my professors she worked at Lockheed. She was in the missile Ballistics department and I asked her when they was like, what do y'all you know code missiles with
21:23 and stuff. And she said, Fortran. And this was in 2007,
21:30 I was like, Why? And she said, Because man, that's what management knows. Like everyone reviewing the code and all the upper VPs in management, that's what they coded in. And so that's what
21:41 they do. I was talking about this so before, but like even just the exposure or the risk of - Right, who knows for refacking that code to another language is like - Yeah, it's whatever, it's
21:53 probably a security risk at that point. I mean, and safety, right? You think about, I don't remember which NASA mission it was where they used a metric instead of Imperial and it blew it down.
22:03 That was when they sent like one of those things to Mars, wasn't it? I don't remember which one, but it can't go. And you're not gonna give it much lower level than Fortran, I don't think. Well,
22:12 that's what she said too, is like, if you can't do it in Fortran, it's not efficient enough. And I was like, that's actually smart. And also from just like the compute you don't want to spend a
22:22 bunch of money on the electronics if they're in a missile that is intended to blow up, right? Like you want the most basic lightweight thing you can have on there, but also even from a hacking
22:32 perspective, it's like, not many people even know for sure to begin with, and so. Yeah, and there's also the, when you're talking about the hacking perspective, there's the issue of package
22:43 managers too. Like I always think about that for certain applications, like if you're gonna build your missile defense system, do you really wanna build it in Node, or even C, where you're
22:53 basically pulling in how many different packages could be 100, could be 20. I'm sure that level has to be like, like local package managers with vetted packages. Like you can't just go to PIP and
23:06 pull it out. I hope so. But that was the whole thing too, wasn't it? I mean, maybe, I think it was when we were at RDS, like I think scikit learned, they realized that there was like something
23:15 program, like, in, in correctly programmed, like, on the, uh, on some of the interpolation methods or whatever it was. And it basically invalidated a ton of - Scientifications, yeah, that
23:27 worked. Oh, wow. Or that they had to go back and validate that those were still true, because like the way it was calculating something in scikit-learn was wrong for years, right? For years,
23:36 yeah. So then people were using that package to do all their research and - Yeah. The pros and cons of packages and third party repositories. Yeah, apply that to a missile And you know, the cost
23:51 of being wrong, get a little higher. Yeah, big time. Well, even like on the, you know, talking about the OT space and the hardware side, right? Like that's the last job I had was an edge
23:59 computing startup and like all every single chip had to be validated and like they had, they were all serialized. And like if that chip serial number didn't match the board, didn't match the CPU,
24:13 didn't match the GPU, we couldn't use it because they were super worried about, Whoever manufactured it originally might have put some malware on it. And then that ends up on a, you know,
24:23 government application. There's all kinds of fascinating security just around the hardware going into that. No, no less the password. Yeah, like it's pretty wild, but so you started, started in
24:36 that. And then how did you, how did you turn that into actually in being in software for most of your career? Uh, I just, I, I picked up software, uh, development and, uh, started working in
24:53 the industry, uh, during the dot-com boom, did a couple of startups. Uh, and then came out the other side of that and, uh, became, I, I actually wound up at an event, uh, with a guy who's
25:07 startup famous, uh, at the Air Force Academy and, uh, oh my gosh, what is his name? steep like steep like he does the uh. He wrote like four steps to the epiphany and he's, you know, if you
25:22 talk to like the dot com boom Silicon Valley crowd, they all know Steve Blank. And anyway, he was talking at the Air Force Academy and there were a bunch of us who went down there from the Denver
25:34 startup community and listened and we had a round table discussion with him afterwards. And as often happens with startups, everybody started saying, When I can't get any money, no one will invest
25:45 in me And so Steve Blank being a very tactful man, instead of responding the way I would have, he said, Well, you know, part of the problem is you guys are trying to build a social networking
25:60 startup in Denver, or you're trying to build an advertising or a finance startup in Denver, and you really need to look at where you have an unfair advantage. And if I was in Denver, I would be
26:10 building an oil and gas startup
26:14 And so I heard that and thought, that's a really good idea. That's awesome. So that's pretty cool that that was just like, you didn't have any ties to the industry per se. 'Cause everything up to
26:27 that point that no oil and gas experience previous to that or there were. No oil and gas experience previous to that. But I knew a guy named Josh Sherlock, and we kind of known each other off and
26:38 on for years. And he was in that room too, and he heard the same thing. And a few months later, we actually, I think it was
26:47 a few months It might have been a little longer than that, but yeah, a little while later, we got together. And there was an event that was going to be energy focused. I think it was Jim
26:57 Thorson's first, they called it energy tech mashup or something like that. And Josh and I got together and said, okay, I can do software development. Josh, you can do oil and gas. He was
27:09 working for, I think, PDC at the time.
27:13 Yeah, let's go and I need you to talk, I don't talk oil and gas, I'm gonna bring a couple of developers, let's go talk to these oil and gas companies and start a company. That's awesome. And
27:23 that became Well Data Labs, by the way. No, I mean, and yeah, y'all have been around for quite a while, have a good, great reputation. And I mean, we've worked with them extensively in the
27:33 past at different companies and stuff. And so I always, I love hearing the stories of like how people got into the space because there is not, like people think that it's like, oh, well first you
27:44 do this. And then, you know, you build up and then you move over here and then you start your own company. And it's like most of the people we've had on do not have a linear path when they started
27:54 to where they are now. And I think it's just comforting for most people to realize that most people don't have it figured out and there isn't a grant, they might have a grand plan, but life never
28:03 works out to be what you're thinking it's going to you. All right, so around whatever time that was it, you got, you know, exposed to oil and gas I will say what was like the most? like shocking,
28:13 shocking thing. You're like, holy cow, what's going on here? And then also what was maybe like the most endearing thing, if you will, or I thought you weren't your favorite things about oil and
28:21 gas that maybe you had an experience in your career.
28:25 I don't know that anything was necessarily shocking. I would maybe say the most endearing and the most surprising thing was actually turned out to be the same thing, where I had worked in a lot of
28:37 different industries. I'd done software in pay-per-click advertising.
28:43 Defense, Homeland Security, all kinds of stuff. Done stuff for the energy, the Department of Energy, not the energy industry. Were you working, did you do anything with NREL while you were? No,
28:54 Rocky Flats.
28:58 People who aren't, for anyone who's not aware, Rocky Flats is notorious for, they used to manufacture the plutonium triggers for nuclear warheads there and they would take like the waste from the
29:09 process, radioactive waste and. Um, Rockwell, I think was the contractor at the time and the FBI raided the place because they were like taking, uh, radioactive ways, sticking it in barrels and
29:21 putting it out in a field with a tarp over it. Wow. Yeah. Can't do that. Yeah. So that, that made Rocky flats famous, that made Rocky flats, uh, you know, a clean upside for the next, I
29:34 think it was 25 years before they finally got it cleaned up But, uh, yeah, I was the part of DOE that was involved with the Rocky flats clean up is where I worked. But anyway, I, uh, getting
29:45 back to your question. I, uh, I had worked with a bunch of different industries. Um, and the only thing that I knew about energy and oil and gas was really, you know, what you see in movies and,
29:56 you know, the common perception of the old boys club. And, uh, the only thing I knew about fracking was I didn't want it done in my backyard And so, um, when we started Well Data Labs, I really
30:07 had to dig in and educate myself. about the industry and about fracking in particular, since that's what Well Day Labs was all about. And the surprising thing that I realized is, number one,
30:20 fracking was not the boogeyman that I thought it was. And number two, which was the really the most surprising and endearing thing, is that this industry is full of some of the
30:35 the smartest people with the most integrity of any industry that I've worked with. 'Cause people are people, you run into good folks and bad folks in any industry, but oil and gas has a
30:46 dramatically higher percentage of scientists, engineers whose beliefs are based on truth, who care about people, who care about the environment. And that is not what you see when you do
30:58 advertising. No, not at all. That's my favorite part is it's like, have you ever met a geologist? They love the environment, like we make jokes about them licking rocks because they like rocks
31:08 so much. And it's like they really want to wear a pedagonia right in our face. So exactly that's right up their alley. They just didn't hate us so much. But what's the what would you say because
31:19 fracking data is something Bobby and I have both dealt with in the past. And I'm still dealing with it today. And it still drives me absolutely crazy. But what what's one of the most like, I don't
31:32 know, shocking things for people who aren't familiar with frack data that that that you discovered as a kind of entrant into the frack side of things.
31:42 It's probably the channel mapping problem. I think is the biggest thing for everybody. Yep. By the way, I probably should have brought this up, you know, off podcast. But if you ever if you
31:52 guys ever want to talk, I got some ideas. I think spark plug B has kind of shown us the way of how we could actually solve that channel mapping problem. You know, you get some folks from why I'm
32:03 in AMI in here and have an afternoon discussion, we might be able to figure that out. See, that to me is ultimately the problem, right? Is it's like, if the manufacturers would have started at
32:13 that, I mean, even most of them do now, where you can just have a template file, and you just open that file, and the naming convention is exactly the same, but none of them use it in the field.
32:24 And so that's ultimately the problem to me, is that there are humans involved in the process. And because of that, there becomes variance in how they name things, or what order they put things in,
32:33 and all of that fun stuff, and then switch it from one stage. Blows it up. Yeah, yeah, that's exactly it. They've got the template file, but number one, who has access to that template file,
32:44 it's definitely not other systems. And then like you pointed out, it changes, 'cause things happen during a frack job. You know, what was your treating pressure on stage two, is no longer your
32:54 treating pressure on stage four. And so frustrating. Or just like, I mean, one of the things, The most basic things that Bobby and I would - push for it, RDS was just like, pull in every
33:05 channel, don't try and just pick and choose which channels are active at the beginning of the stage, because by the end of the stage, you're probably going to have different channels active
33:12 depending on what, you know, how much chemical is left and one of your totes or whatever. And it's, or you, you know, there's just so much there. And so just the way that people, you know,
33:21 we're thinking, they weren't thinking of it from a data perspective when it happened. That's ultimately, I think, one of the why we run into some of these issues and the, and the oil and gas
33:30 space, at least on the data side, is that it's like the software existed before anyone ever thought about using the data from the software. And then now it's like, oh, well, of course, we're
33:41 going to use the data from the software, but the software itself hasn't really changed that much in 20 plus years. And so it's like, it does more stuff, but the root of how you set up the job and
33:52 all of that stuff never changed. And so we're just, it's basically tech debt that I mean, before 2010, probably like, what do I want? No, I mean, absolutely. It made no sense. No one was
34:02 asking me to stream the raw data anywhere in 2010 when I started working on frack jobs. No one gave a shit. Like, I gave them the ASCII file at the end of the job. And that was it. You know, in
34:15 the PDF. There we go. And then five years later, everyone's like, Oh, no, we need digit. Like that data needs to be digital. We need to use it. And what the fuck? What's more? Frank,
34:24 what's the most? Oh, it's still there. There's a new, there's another, there's another iteration. So the guy was telling me about it. Early in a couple weeks, it's probably only coming from
34:31 Halliburton trucks. I'm sure it's Halliburton only or, you know, two service companies that run the consortium. And yeah, we'll go on a whole diet drive about that. But yeah, you know, it's
34:43 funny The, the frack stuff is, is not really related to what I'm doing today in petro bricks. I mean, it could. We, petro bricks would be a great way to get data out of well day labs or Corva
34:54 and get it into snowflake or something like that. But that's not what people are really using it for. It's all the production use cases. But with that said, spend a lot of time thinking about FRAC.
35:06 And you never stop thinking about that problem. And I'm not kidding. I think that what you said, where you were encouraging people to track every channel, I think that's it. Pulled 160 channels
35:17 we've got. Publish that to an MQTT broker. And then similar to the way that we layer state information for devices with Spark plug, the channel mappings are just more state information Yeah, it's
35:31 just one of the things. If you have all the channels, and then you see whatever, your biosides start going down in the middle of it, and then on another channel that's not named, but it's mapped,
35:41 and it's there, you start seeing it come back up, and they happen to be in the same range that the channel that fell off was in. There's a very high correlation that that was just a different tote
35:51 of the same chemical. And so, but if you didn't have that channel because it wasn't originally mapped, then it just looks like it fell off and nothing happened And so it's just, it's little stuff
35:60 and it's even annoying though that you'd have to do that because then it's like, you still don't have the confidence in that data or that channel being the same data from the one that fell off. But
36:11 generally speaking, it should be, right? Like you just have your domain expertise and it's. Well, if anyone from Lime or AMI is out there. Yeah. It's called on the right or on the corner. Yeah,
36:23 well, but that's the thing too. It's like on top of that. Then the service companies are all now trying to push out their own thing because they buy the equipment from the manufacturer, and so
36:32 it's just a business. Yeah, we need some standards. And
36:38 seriously, I've been thinking about this for, I would say, a good six years now, maybe seven. And it occurred to me while I'm building this other stuff and seeing how other problems are solved.
36:51 It's like, this is the solution right here. Like it sparkle of B isn't the answer, but it's basically MQTT for the base data. and then state information using a spark plug, be like standard would
37:05 actually solve that problem. And suddenly everything in the field would be able to talk and share that channel mapping data. Even when we were at RDS, I pushed so hard for a couple of different
37:15 people and even us. I was like, it doesn't have to be a format, but we can just publish a naming convention. And just like, let's see if anybody decides to use it, piggyback off of it, we never
37:26 did But it's like, people get so hung up on, is it Whits-A-Mill, or is it X-A-Mill, or is it JSON, or what, it's like, it doesn't fucking matter what structure you're left in. I can pull it
37:38 right around. I can do whatever I want with it once I get it. But if the naming conventions are different every single stage, then it doesn't matter what format it's in, because it's still junk
37:47 and it requires all this extra work, where it's just like, hey, if it's FR, let's name it FR. Even though our proprietary chemical name is something else, we know that FR is FR. Like, the
37:57 chemical will be on the - bomb, it'll be in the report, like you don't need that type of level of detail, but just having a standard naming convention would literally solve this problem. I feel
38:10 like in the industry, but we're never going to have zero confidence that that will ever happen. For anyone out there who has not had the joy of dealing with frack data, it's hard. Yeah, that's
38:20 definitely not true. Three row
38:25 headers are my favorite. We're doing a project right now for this. Yeah, it's like chemical name, proprietary chemical name internally, the general name, fr bio side, whatever, and then the
38:33 units. And that's one that's the header for each column for all the chemicals. And then on the next, next file for the exact same operator, same format and everything, there's only one header.
38:44 And then on the next one, there's only two. And it's like, or the headers run together. So when you do the extraction, it shifts the columns and screws up your entire data set. It's just, it's
38:54 so anyway, we'll get off of that. We've been on that for a minute.
39:00 What are some, these are one of the big things where we like to ask people, but what are some tools that you kind of use on a daily, weekly, monthly basis that you find valuable, you find helpful,
39:10 whether that's libraries or softwares or anything like that? Yeah, I'm boring.
39:17 I do. I kind of take the party like it's 1999 approach to software development where I'm using Visual Studio I probably will switch over to Rider sometime soon for a development environment and - Was
39:33 Rider, I don't think I've heard that one. Oh, Rider's the, it's so - Visual Studio used to be awesome, but man, it's - You're talking about like old school Visual Studio, not VS Code, but most
39:43 of us - Yeah, not VS Code, Visual Studio, like full-on Visual Studio. They still make that.
39:51 But they keep adding more and more stuff to it, where now it crashes on me. It's very heavy, right?
39:56 So Ryder's a lighter weight version done by a company called JetBrains. They make great
40:01 stuff. They make some good stuff. What was there?
40:05 PyCharm, yeah.
40:07 Yeah, PyCharm, quite a bit. DataGrip. Yeah, DataGrip's a really good ID for SQL stuff. Yeah, but yeah,
40:15 DataGrip, PyCharm, all the stuff that JetBrains makes is pretty great. So I do use JetBrains, but I use basically Visual Studio with JetBrains ReSharper plugin And a lot of, you know, one of the
40:28 things that I did intentionally was I wanted to reduce the number of external dependencies.
40:39 So I minimize the number of libraries and, you know,
40:41 external, I don't know, code that I might bring into petro bricks because the idea is to make it as secure as possible, but also as simple as possible. 'Cause I, you know, in other environments,
40:51 I've done a lot of consulting around with different big distributed systems. Consistently, you would see stuff where, oh, we just pulled in the latest version of a package and now prod is broken.
41:01 Yeah, everything broke. And the crazy thing is, a lot of times, those errors, they break in a way where you don't find it during testing, where all the automated tests work, and then you deploy
41:11 to prod and suddenly, oh, after it gets under some load, it blows up. What changed? Yeah. So anyway, I really try and minimize that kind of risk and as a result, I write very boring stuff
41:24 where there's a lot of, you know, write the simplest version of it into the product itself. I'm a big fan of the Aka framework, which I don't know if you guys have - I think you told me about it.
41:36 That's through a Java framework or? Yeah, it came up through Java. There's an Akanet that's written for C. But, you know, for anyone out there who's writing complex systems where you have lots
41:51 of multi-threading or even if you just - and it can be multi-threading on a single box or it can be multi-threading across like a whole server farm.
42:00 That Akka model, the actor model, is a really good way to handle concurrency. And it's kind of, it's not the way that mostnet folks do it. Yeah, I was gonna ask you, can you explain kind of the
42:12 difference in its methodology versus something else? Yeah, it's a, so instead of basically, Akka allows you to write multi-threaded code in a way where you don't have to lock anything. And for
42:28 folks who aren't programmers, when you lock stuff, like if you have multiple threads who are trying to access the same piece of memory at the same time, if they both change memory or one piece of
42:40 code, write something to that piece of memory and then it reads it back, but another piece of code has changed that block of memory in the meantime, it causes major, major problems. And that's
42:50 how you get bad crashes and security vulnerabilities. So the way you get around that is locking, but locking creates a whole host of other problems. And so Aka is a way where it uses functional
43:01 programming
43:04 principles to use
43:07 non data that doesn't change immutable state data. And it basically forces you to write a big system instead of writing it like one big piece of code It's like it's, you know, 20, 50. It's kind
43:22 of ties almost. Right, many, many pieces of code, like many programs. And then each part of your system talks to the rest of the system only through message passing. And all of that's a lot of,
43:32 that means a lot more to the programmers out there than the average folks, but the benefit that that gives you is it just allows you to make, it's a much safer way to structure highly concurrence,
43:45 you know, highly concurrent systems where you get a bunch of different things you're doing at the same time. Oh, that's cool. What have you used any AI coding tools or anything yet? Oh, yeah.
43:58 Keeping that up. What are your kind of favorites, or which ones have you been playing with, and what's your impressions on them? So I don't wind up using AI coding the way I think a lot of folks
44:09 do. I almost use AI. I use Grok almost exclusively. I've tried Claude And I wind up going back to Grok. And I did the Pepsi Challenge thing, where I was like, hey, go to four different AIs and
44:24 say, write a piece of code
44:28 that does this, and compare the results back. But I actually don't use AI to really write code so much as with Petrobrics, it connects to a bunch of stuff that I'm not familiar with. And I might
44:44 have to deal with a new encryption standard or a new. Like with Spark, it was actually really helpful with Spark would be when I was first getting into Spark would be because I could go and I could,
44:55 you know, have it in just the entire Spark would be specification and then ask questions. So it's fantastic for stuff like that. So I will usually do something where I'm trying to learn something
45:07 or answer a question about something that I'm not familiar with and I'll show it the resources to go pull in and then I'll kind of have a conversation and then I'll ask it to write sample code and
45:17 I'll tell it, Well, here's where that's a problem. And it's almost like it allows me to have a conversation with a knowledgeable colleague. Yeah, yeah, no, that's exactly. People don't, in my
45:28 opinion, They're programming a way or whatever. They don't think of it as like a partner in it. A lot of the, like, okay, yeah, it's a partner because it can write the code. But let's take a
45:38 step back and help me think through all of the different things that I'm not thinking about or if it's a new package or a new service ask nine long like error message and like, right? Right, tell
45:49 me still is down what the hell's going on. Oh, it's fantastic for that. You can just drop an error message in there and it tells you, oh yeah, this is what it's talking about. Yeah. Yep.
45:57 Really? I didn't get that when I first read that error. Yeah, no, it's great for that. Well, these are the conversations we like having here too 'cause it's, you know, I think we were saying
46:07 it before we came on, but it's like, you know, it's good to know what's in your stack, but at the same time, like, Patrick Brix could or should be part of a lot of, our listeners stack too,
46:18 where they don't even have to know, they don't even have to know it's the Akafre Merck and, you know, maybe in five years, something better comes around and you rip it out and they'll be none the
46:20 wiser, right? And that's the whole point, like - Yeah, no one wants to know about Akaf who's using Patrick Brix. There's like
46:34 five programmers out there who are like, Oh, yeah, that's interesting. Yeah. Yeah, Josh Grohl's a bit like that. Okay, well, that's probably listening, so Hopefully. Well, so given that,
46:43 what is in your stack on? on for petro bricks. What are y'all running on? You have any cloud? I guess you said you're - Is it containerized in that way? But you can run it, say, edge here or
46:56 there, every work kind of thing. Yeah, so the idea is to make it as simple as possible. The app itself is a running instance of petro bricks. We call a petro bricks node. And so you've got the
47:10 core kind of data switch, which is the node. And then you've got a web configurator, which you use to create all your data pipelines. But the interesting thing is, the way that it's structured is
47:21 you're essentially creating a big config file. And the idea is that a petro bricks node could be deployed headless to a piece of equipment out in the field. So anyway, you've got your petro bricks
47:34 node, you've got your configurator, and all of that stuff is wrapped up in a single Docker container that you can basically run anywhere you run Docker Yeah, that's awesome, which is Linux or
47:44 whatever. Yeah, windows, windows. Yeah, I did. Interestingly enough, you know, I thought I'd run into more windows. Everybody is doing Linux. Yeah. Especially - They are and they aren't.
47:55 It's like, well - Well, for Docker. Oh, oh yeah, I mean, no, I think running Docker in production, I don't even know, Windows is like, I think you can run, I have Docker desktop, Bobby,
48:07 what do you mean? Yeah, you have defined production.
48:11 Yeah, well, it's production. It's on a laptop, I'm gonna close it JD Long, we head on, you've got a tweet or something, but I was like, you know, asking about production and then, yeah,
48:20 Excel spreadsheets. And he's like, well, by that definition, all your Excel spreadsheets in the company are in production. You know, like, production's a bullshit, you know,
48:32 excuse me. Probably the fact that they're making multi-millionaire billion dollar decisions off of whatever Excel spreadsheets someone's got, but not it. It happens in every industry every day,
48:41 it's crazy. But I think there's also a lot of IT folks
48:46 what's a container and I'm not touching Linux. Actually, that's a good for us to just, I don't know how deep we've ever gotten into containers and all of that stuff before, but I think it's
48:56 probably a good thing to just highlight what they are and why they're so useful because yeah, they really are when you compare it to, I mean, it's kind of almost going back to what like the Java
49:07 virtual machine was in a way, but write once and deploy anywhere kind of thing, right? Like, you know, you can write your code. It's almost like a whatever VM, what is, but you don't need to
49:15 partition the hard drive for everything, right? It just, the best description I've ever been given is it's a recipe for the computer as to exactly how to set up the environment when you start,
49:26 right? Like, it's pre-configured, you don't have to go and install packages manually, the environment, the operating system, everything is structured and ready to be. Well, I guess it solves
49:35 the whole, well, it worked on my computer. Yeah. I'll send you the local host link to share with you. Yeah, I always tell folks, it's a. You'll start with something that they maybe have heard
49:47 of, a virtual machine. And then you say, it's basically a virtual machine, but it's actually not a virtual machine. Yeah, pretty much, yeah. But functionally, for all intents and purposes,
49:58 functionally it's like a virtual machine. Just better. Just better, exactly. But again, because you can take it and deploy it to an edge device, or all major cloud providers have container
50:10 Kubernetes services that you can just send the container there, or again, you can run it, 'cause they even just the deployment of it, right? Like, I can go into Azure and drop in a GitHub link
50:18 and deploy a Docker file or whatever else, and they just, like I said, it's a recipe, and it just works, it just works. As a software developer, the big advantage from my perspective is you
50:32 have control of the environment, you have a consistent environment, and you can avoid, you ever heard of DLL hell? Yeah. So you're installing software on bare metal, You've got five different
50:43 pieces of software and you got one of them that decides to change some DLL that everybody else is using. None of that happens in a container 'cause you have, it's just its own self-contained
50:54 environment in a container and you know exactly what the environment's gonna look like, what DLLs you're gonna have and so it's a much more stable way to deploy software. Yeah, no and it's just so,
51:07 it's so much easier than the traditional ways too. But I do think like when I was at Hivesall and I was talking to a lot of OT folks, I do think in the SCADA industrial manufacturing world, VMs
51:20 have been the standard for decades. And so even whenever that was three, four years ago, it was like, what's a container? I'm like, oh man, this is gonna be a long conversation. Can I say a
51:30 container on a VM? Yeah, well no, and like most of
51:34 the pilots that we did ended up being that. It's like, oh, well, we've already got a VM, so we'll just put it in there. And I'm like, well, you're kind of missing the part. I mean, I guess
51:41 we can do that. You know, one thing that's fantastic about containers for any oil and gas folks out there, I've seen this where there's operators who aren't currently using containers. And I'll
51:51 talk to those folks. And, you know, we have a little conversation around Docker and it's like, yeah, not only are you gonna use this for petrol bricks, but let's say that you wanna have an MQTT
52:02 broker that you're using for a test system. You just pull down a Docker image and you edit your config file and you start that sucker up You've got it in like, I don't know, 20 minutes. It's so
52:14 easy to pull down and deploy software. We're like, you know, installing a database and configuring it used to be like an all-day affair. Yeah,
52:25 it's wild. And it's just wild to me how many people still, I mean, you know, you've got all this legacy stuff built in VMs and so they've tried to get out of that, but man, it's worth it and I
52:36 feel like
52:38 in a long run. And for you guys, 'cause I'm like, You said like there's akanet so are you running. net core? Yeah. NET 4. Which I think, that in itself has been a huge, you know, what in the
52:48 last five, 10 years, a big, not even 10 years, like probably last five, six, seven years or so. Yeah, Microsoft, I didn't know what to think when they first were changing. NET and they're
53:00 gonna, you know, they're gonna port it out so that it runs anywhere, but specifically on Linux, because really it wasn't about running anywhere. It was just about running on Linux, 'cause that's
53:10 what everyone's using for enterprise Yeah, enterprise anywhere, right? Yeah. That's everywhere. They've done a really
53:18 good job. Yeah. It's a great way. I, like, I, and as a developer, you know, I did, you know, Windows, Windows deployments, and, you know, it was basically building Windows software for
53:30 most of my career. And I really didn't notice a significant difference when, you know, switching over and running to, running on Linux and building to Linux. It's basically a build option at this
53:43 point. No, it is really cool. And then, I mean, I think it must be similar to things going on under the hood, but where we had John Archer on way, we've actually started with, he's here with
53:53 Red Hat over at the Canon, are they still there? I think so, but I remember him saying actually that SQL Server actually runs faster on Red Hat than it does on Windows. Hmm, I've heard stuff
54:05 about that, Postgres, you hear that about a lot too. And apparently the Linux file system does things a little differently than what is it, NTFS, on Windows. And it just works better for
54:20 databases. Yeah, but it's just funny because, yeah, you'd think, Microsoft. Yeah, but again, Windows is usually so much more bloated than a Linux distribution, too. There's a lot going on
54:31 there, too. 100. So I guess with that, we can get in a little speed round. Speed round, man, that's, yeah, hour flies by. To that point, I'll start with one, just give 'em off the top of
54:43 my head, but like, Windows or Linux, what do you prefer? For a desktop, I do still prefer - Or Apple, I guess we can either Apple if we're gonna talk about desktop too, but - Oh, desktop. I
54:54 would say Windows, I use Windows, and you know, it's probably something more like, that's what I'm familiar with, more than it's superior. It's just, yeah, I know how to do stuff, so I keep
55:03 using it. Yeah, I mean, some of it may be familiarity, but like, I've tried to use Apple, and like, now getting the folders and stuff on there's like, why is this so obtuse? Yeah, it should
55:14 be pretty similar. It's just such a different experience if you're doing it as part of your day-to-day like business, like if you just have normal shit that you're trying to do, and you've been on
55:24 Windows for your entire life, it's very easy to stay on Windows, even with development. It's like, yeah, you've got some funky stuff with WSL that you got to deal with and all that, but it's
55:34 just like, if I'm not coding, then it's infinitely more productive. for me to use Windows as Apple, especially in oil and gas or where I need the office suite, never uses that look. Oh, yes,
55:44 it's the other thing. It's just all the different things, like they all run on Windows. So you are making me think of something, though. Have you, I do use Windows and yeah, I agree with all
55:54 the reasons stated here. But have you ever been in the situation where you got a meeting and Windows forces a reboot and then it's 10 freaking minutes to apply whatever the Windows update is? Do you
56:06 ever think about using Linux as a desktop in those moments? For sure. Yeah. But I got myself out real quick because actually the XPS laptop I have, I bought from Dell as a Linux laptop because I
56:19 thought it was going to be cool and I was going to do some consulting work on that back with 2019 and then I was like, oh, how long did the Linux instance last for a couple of weeks? I was like,
56:29 oh, I need a spot fire and I don't want to, again, I don't want to create a Windows VM because now I can only allocate half the resources to it or whatever else. But I mean, with that Windows
56:41 subsystem for Linux has come a long way too. I mean, I gave it was janky early, you know, probably five, six, seven year, but like it's pretty, pretty, I'm good now. Yeah. WSL two. Yeah.
56:51 I haven't run into nearly as many issues in the last couple of years as I, which I think that was enabled by containerization to, I think, doesn't need Docker technically running underneath the
56:60 hood. That would make sense for it to work. I think it's been a virtual, I mean, spinning up all that stuff. Yeah. It makes it a lot easier I think you're right. And, but even with that, they
57:11 made some massive improvements with WSL2. Like I noticed a huge difference. Yeah. I mean, I think a lot of true developers like to use that. I mean, because it used to be, you had to have
57:23 basically almost a Mac, you know, if you want to do any Unix based development. Um, it was just kind of like, we have to use this because then I can port this over to a Linux server more easily,
57:33 but now it's not the case. So it's BSD under the covers. So yeah But like you're saying, I mean, I think across the board. Microsoft has done a fantastic job of leaning into Linux. And again, I
57:44 think when they bought GitHub, they were also the largest contributor on GitHub. And people always thought they are very enterprisey and probably the whole embrace, extend, extinguish thing still
57:53 lives on a little bit. But they've, you know, they were the largest contributor open source as well. Yeah. I feel like if you just give Microsoft enough time, they'll eventually get it right.
58:03 Kind of like with Power BI, like it's so bad when it first came for fabric. I don't know if that's ever going to happen I've been waiting for Synapse and they'll just change your day to warehouse.
58:12 Well, they do. Every time they change the name, they get a two-year reset when period, right? They do have some really smart people working on the. No, I mean, they brought in some people. I
58:21 think a couple of guys are mother duck, but we're also big query folks and everything. They brought them in on the fabric team. Yeah. So, to your point, give it two or three years and that could
58:30 be really interesting. Who knows? Right. But yeah, normally I don't like to mess with their stuff early on because it's just like, I know it's not going to be great. Yeah. Um, what's your go
58:41 to, if I'm coming to, are you in Denver or are you some, like outside Denver or even Boulder or anywhere or are you actually in Denver? A little bit outside Denver, closer to, closer to Golden.
58:51 Okay. So what's your go to recommendation? If I'm coming into town for a day, what should I, what should I go do? Uh, you should call me. I will. I haven't been
59:06 a Denver in long enough. You should call me. And then, uh, I would, uh, highly recommend you go walk along clear, go to Golden walk along clear Creek, they've got a, they've got this amazing
59:10 path that they're building where, uh, it goes like miles up into the canyon for, you know, riding bikes and walking and all that. And, uh, you can walk all the way back down into Golden and you
59:21 go to the mountain toad for a beer. That's cool. But yeah, sounds like a good day. It sure does.
59:27 Um, what's your favorite vacation destination? I think the black hills. I really like the black hills these days What's
59:36 your go to LLM today? And this obviously changes, but which one, if you've got a problem that you're going to let an LLM solve, are you going to go to first? It's got to be Grock. Yeah. I don't
59:48 know if they're the best. Grock with a K or with a Q? With a K. What's Grock with a Q? It's their enterprise version, so they've got, but it's all, Grock with a K was built on Twitter, but the
1:00:04 tech and the infrastructure that Grock with a K runs on, I believe, is Grock with a Q And so it's just incredibly, like if you want an API key and you want to do dev and stuff with it, they've got
1:00:15 all these open source models that they host on their whatever TPU servers or whatever the ridiculous servers they've got. And it is insane how fast it is. I did a test like, I don't know, nine
1:00:27 months ago, just comparing like time to first token, tokens per second, and just overall completion time of like Azure OpenAI, Jim and I, Claude. and grok and grok just murdered. I mean, like
1:00:41 10 times faster for time to first response and tokens per second. I might have to check that out. So I'm definitely, by the way, I'm also not the guy to take any kind of opinion from on AI. But
1:00:57 yeah, like we talked about, I do use it and I keep going back to grok, but my sampling size is very small. Yeah, that's fair. I mean, everyone uses it differently. So I'm not sure how often
1:01:10 you come to Houston, but is there any one anywhere when you come to Houston, like I need to go there, like, but as far as restaurants or destinations or anything, I keep saying I need to get
1:01:19 barbecue and I never have. Oh, every time I've been here, I never went up going to get barbecue, but it's going to happen. Yeah, it's not going to happen this time either.
1:01:29 What about either favorite book or a book you would recommend to someone they're trying to get into techsoftware or energy, you know? I don't know, for someone looking to get into tech or energy,
1:01:41 I don't know what I would recommend. Most of my reading is, so much of my work life is tied up in - Oh, that's totally fair.
1:01:52 Yeah, so when I read stuff, I actually do a lot of reading. And I've gotten into reading old
1:02:02 hardback science fiction novels. And then I do a lot of science reading and stuff like that But I don't know that I have a good - I've got like, nerdy programmer book recommendations. And I've got
1:02:14 old sci-fi book recommendations. I mean, I think our audience falls into both of those categories and all of these. You want me to be one of the things, at least, yeah.
1:02:24 So on the sci-fi
1:02:28 book recommendation side, I would say
1:02:32 use of weapons by Ian Banks. is a fantastic book, his whole series, I don't know if you've ever heard the culture series, but it's a sci-fi series. You might like it. It's about a very distant
1:02:46 future where nobody really remembers Earth and humanity spread out through the universe. And they basically have partnered with sentient AIs. And
1:02:56 the AIs really run everything, but it's not like Terminator AI. They're like, they're like useful, nice AI, but at the same time, they're kind of
1:03:07 jerks.
1:03:09 And they act like people and it's very interesting. And Ian Banks is a smart guy. That's cool. No, my wife makes fun of me because I don't read any. All I read, I listen to mostly are like,
1:03:24 either business entrepreneurship or like history or biographies, right? I don't listen to, or I don't read or listen to any fiction. So she's like, you're so weird I think it's weird that you
1:03:33 listen to made up stories all day. You know here nor there I would I need to get get some friction or some fiction going to let my Creativity have some fun a little bit. Yeah, but I think that is a
1:03:45 side of it We're like you can get locked in for sure not expand your horizons or think of think of what's possible. Yeah a little more so Cool, man. Well, thanks for coming. This has been very
1:03:53 well out of worked out. Yeah, thank you for having me I appreciate it Where can people find you get in touch with you if they want to learn more or reach out to you guys for some work? so Rudy are
1:04:04 you dy at petro bricks calm where you know petro bricks calm on the web and Yeah, anyone feel free to especially if you're an operator. You're looking to solve some of these problems Send me an
1:04:16 email hit me up on LinkedIn happy to talk awesome. Thanks so much Rudy. Thanks. Thanks guys. Appreciate everybody. We'll see y'all next time