0:00 Welcome to another virtual episode of Energy Bites. I'm your co-host, Bobbie Nealon, with the Raddad, John Calfan. How's it going? How's it going? How's a good man need for all these days?
0:13 Yeah, Richmond, we're seeing the cows up back. It's pretty nice. Patty has actually been built, so we're on the right track. But no, super excited. We've got Cameron Snow here, co-founder of
0:19 Denomics with us. So
0:22 he's got a,
0:29 you know, really cool software and also, I think, you know, you're, you know, not shy to share your takes on AI and technology and everything else going on. So excited to have you here and,
0:39 you know, here's some of that. Yeah, well, thanks for having me, guys. I'm an, I'm an equal opportunity offender. So, you know, I'll try not to hold back too much here. I think you're in
0:49 good company with, with, with us to don't hold back unless it's bad for your business. So we want it, we want it all. But, you know, sometimes we start with some current events and I don't know
1:00 if you got any opinions on it can, but I think interestingly, EOG bought Encino recently correct last week or two. And that's, I feel like it's usually kind of out of character for them to go buy
1:12 someone, right? It's been a while for that. Usually they try to find things organically, but interesting to see that and see what's going on. I wonder if you guys got any takes on that. Yeah. I
1:22 mean, my take on it is there's probably not a lot of organic things they can chase right now that are of material enough size, right? I mean, for it to move the needle for those guys, they need
1:35 to build a ramp, anything they're getting up to, to what a couple hundred thousand barrels a day. And, you know, there's no secrets anymore in the Permian Basin really is terms of like leasing
1:46 things organically, same for the Mountain West. You know, so I think just kind of put them in place where they had to do an acquisition if they want to, uh, they want to grow the base.
1:59 And I think someone with something that I think Encino's got a lot of takeaway capacity, I think, in that area, too. So I think that there's a big aspect of that, too. But
2:09 I found it pretty interesting. I mean, it's been a little quiet on the MA front lately, but I think if prices suppress a little bit longer, it's going to ramp up, I think, with some of the PE
2:17 folks. So it should be interesting. Get their deals where they can, right? Yeah, that's the whole point. I do think it brings up like an interesting piece about just like where we're at in the
2:29 kind of US shale production acreage tier kind of stuff, right? Because it's to your point, Bobby. They don't typically go out and buy a lot of acreage, but up in the Utica, it seems like a lot
2:43 of that acreage is already kind of leased. And so apparently they think there's something special about it enough that they went and bought it. And so I think, yeah, that's going to keep driving
2:55 the MA stuff until we start seeing more discoveries,
3:01 Where else are you going to get it? I
3:04 think some of that, like, what does it discovery these days? I assume sometimes I see discovery and it's on top of a geographic area, at least that they've been pumping oil for years, you know,
3:14 so
3:16 yeah, I think that's a good point. I mean, you know, are they going into the same acre, it's looking at the same, you know, essentially the same place, but they're just completing it better
3:24 They think there's, you know, potential to like been the, you know, the type well, you know, a little bit, a little bit higher. So, you know, a lot of different ways they can, they can
3:35 probably optimize and make money on that. So. For sure.
3:40 So, I mean, it looks like we've done a lot of work on that, that's subsurface side. So, I mean, let's talk about what dynamics is and, you know, kind of what the genesis was and, you know,
3:48 kind of what you guys do and we can dive deeper. Yeah So, what it is we do is we're a subsurface interpretation package, so we have. you know, geoscience, petro physics, and reservoir
4:01 engineering workflows. So, you know, in terms of the users, you know, the geologists are in there picking tops, you know, making structure and ISAPAC maps, calculating reservoir properties,
4:15 you know, things like porosities and saturations and oil and place.
4:20 And then we've expanded that into kind of 3D property modeling. We've also added forecasting and economics in there You can bring in your production, you know, forecast everything out, you know,
4:33 get some, get some economics. And,
4:37 you know, we've also focused on, you know, really trying to, you know, make it a scalable platform. So, you know, as we've gone from like conventional plays to unconventional, a lot of the
4:51 legacy softwares were locked into looking at, you know, especially on the petro physics side, looking at things one well at a time. And we're really focused on, you know, letting them look at
5:01 hundreds, thousands, tens of thousands as well. So you can go in and analyze something like the Permian Basin in a reasonable timeframe. So that's what our focus has been on. It's been delivering
5:10 the tools for the geologists and engineers to do their job. And why did you guys need to exist? I mean, I've been around it, so I kind of know, but I mean, thinking of someone's listening, they
5:19 don't have the background like, what are you up against and what are you trying to make better? Yeah, so what it is that kind of got us to start this was, I had been working on the petrophysics
5:32 side for like the first half of my career, and I felt we were doing a really good job of looking at one well at a time and doing a gold standard analysis on it, right? So logs that come in, maybe
5:47 you've got some core, you build the model, you calibrate the core, And then you ship that off to the geologist and, uh, you know, they go to kind of apply that model and a bunch of the kind of
6:03 nuance and sophistication gets lost. And then I went to work for an exploration group. And
6:10 so it was more on the geology side. And all our analyses were kind of one size fits all, like top to bottom. You would kind of run the same model. If you wanted to
6:25 customize things, you'd have to kind of run the whole process over and over and over. And it was a real challenge. And so we decided we wanted to bridge the gap there. So we wanted to live
6:38 somewhere in between the legacy petro physics softwares and the legacy geology softwares. So make something that still scales, but that allows you to go in and build formation by formation models.
6:56 You know customize wells as you you know as needed And and that seemed to really resonate with some of our early customers and we've just been building that concept out from there I mean so so it was
7:08 really that both of the legacy software has been built for you know one use case for the other and We needed something in the middle
7:18 Got it so it looks like you guys started around like 2017 so we're just probably getting where I really come some of that cloud maturity You know coming along, right? So is that was that something
7:27 that enabled this that maybe wasn't available before or? Yeah, I mean so when we when we started that was a conscious decision to go cloud That was kind of a thesis that we had was that cloud was
7:41 gonna be big because we were probably a little bit early On that at least you know relative to with our customers comfort levels
7:51 And yeah, yeah, you guys know what that's like. Um, yep, so you know, cloud certainly helped. And on the, like for us on like the sustainability front, like the ability to push updates and
8:05 modify the software and grow the company, a cloud was a must. 'Cause otherwise, you know, you're shipping a product like once every six months at best. And you're waiting on IT to install stuff.
8:18 And, you know, with cloud, like it made it so we could iteratively get feedback and improve And, you know, without that, we'd be, you know, years behind where we are now. No, for sure,
8:32 'cause even a lot of this, you know, typical, I guess, the petrophysics and geology stuff runs on, and people have these massive, you know, towers at their desks and everything. So it's, you
8:43 know, like, now you've got to wait for IT to install it and they want that only the last LTS version, even if you came out with a cool feature and this intermediate feature and everything.
8:52 the cloud, again, you kind of have almost a super computer at your disposal if you want it, right? And you can do some of that processing really quickly, but then you can kind of get pass on that
9:01 because it's browser based. But with that, I mean, like, can you, because like what I've always kind of heard, a lot of times the cloud or hosting some of these things in the cloud can be a
9:11 limiting factor for some people trying to do geological workflows, whether it's the, the, the screen, like pixels and be able to get things to render. So I mean, like, has that been something
9:20 you've had overcome? Not, not, not really. I mean, so in terms of screen resolution and the sharpness of stuff, that, that was not a challenge. Now, what is it? What has been a bit of a
9:34 challenge is we've recently started working with, with seismic data and 3D property volumes. And with that, you're just, you're just shipping a lot more data across. So you need to pay a little
9:47 bit of attention to like how you're how you're loading data, what are you cashing and whatnot. But those are all very manageable challenges. I mean, I think a lot of the struggles that
10:04 some of the people have had historically with this have been because the software they're using isn't actually cloud-based. It's a remote desktop that they're logging into.
10:16 And so it's like, especially now when you talk to people and they're remoting into their work desktop to remote into a cloud desktop, it's like, oh my god, what are we doing here? Yeah, and in
10:28 many of VPN on top of it and just making a round trip, multiple different directions. So yeah, no, I think that makes a lot of sense.
10:40 All right, so I mean, are you excited to go cloud first? I mean, like, as, and say as much as you're comfortable with it, I mean, like, did you guys land on one cloud over the other? Do you
10:47 try, you know, Azure AWS GCP and like, arrive on one that was sitting your needs better than another or? Yeah, we ended up going with Google Cloud and not for, okay, any super reason. The
11:00 reason we went for it is the guy that I started with, Roland, he was working for Google at the time And it was like, all right, we'll just go with GCP. So, you know. I was gonna say, I feel
11:13 like that's a
11:16 thing that people don't think about when they're considering is it's like, hey, use the one that someone there knows how to use, like that's like a critically important piece. We've been building
11:27 out all of our language model stuff on Azure because most of the industry is on Azure And a lot of -
11:36 our stuff ends up being hosted on the client's stuff, but no one here knew Azure when we started. And so it's like that adds a whole nother wrinkle to it because it's like, to always point, they
11:46 all do the same things, but they all call them things, they call those things differently and they all have their own nuances about how you deploy them or the boundary conditions and parameters that
11:58 you have to live with and for those. And so it's, yeah.
12:04 It's work with something that you know, though, is always a much easier thing than having to learn something about you. Yeah, and I think that's a really good point you make there about, they all
12:12 are for the same thing, but they all call it something different. I mean, it makes it, you know, really hard to, you know, if you want to go do a toy project on AWS and you're used to GCP,
12:24 it's like, oh, what am I looking for? What are they called, like, the small instance or something? I mean, yeah, so Yeah, so I guess to that end, though, I mean, like, is. is your
12:36 application tightly coupled with GCP or have you all built it in a way to say containerized and so on that it could be deployed elsewhere or even like, I guess it maybe brings in some data
12:45 sovereignty things where, you know, let's say I'm assuming you guys probably work with some international clients potentially too, and some of these countries require it to run in their data
12:54 centers or within their borders, just curious like if you built it, you know, with that in mind. Yeah, so, you know, it is somewhat containerized, but I will say that, you know, we are
13:06 heavily integrated with GCP, so, I mean, you know, if we were, you know, required for some reason to use Azure AWS, I mean, it would be a major effort to kind of board that over. Yeah, you
13:21 know,
13:23 but when it comes to like data sovereignty, I mean, you know, our focus has largely been on the US. and Canada, you know, a little bit of Europe or Australia, But, you know. When it comes to,
13:35 let's say, some of the Middle Eastern countries,
13:39 we have worked with a couple of our customers to get approval to use GCP, even though we're not hosting it in country. So you just work it out contractually with the ministry.
13:58 And then all the cloud providers are growing their worldwide footprint, right? So that's kind of one of the exciting things to see about data center build out is you may have looked a couple of
14:10 years ago and been like, oh, there's nothing in this region. And now all three providers stood up a big data center there. You brought up the whole data residency issue. How has that been to
14:22 navigate? Because that's something that we've talked about a handful of times on previous episodes and stuff, but it's one of those things that if you don't know about it coming into the industry,
14:32 really kind of screw yourself in certain ways or have to at least jump through a bunch of extra hoops if you're not aware of it. But how has that kind of been on y'all's end? Has it been pretty,
14:44 are they fairly reasonable to work with generally speaking? Or is it, is a lot of kind of labor? Yeah, it varies country by country, right?
14:53 You know, we obviously knew about it going in, you know, 'cause I've worked in countries before that have residency requirements on their data.
15:02 You know, and it, it depends like in some countries, like you just get a no right away and okay fine, like we understand and in others you get a, you know, a not a problem or it's up to the
15:19 operator to decide if something should be secret or not. And by secret, they don't mean like public or not, they mean like kept in country So, you know, it. berries. I mean, you know,
15:34 each country kind of has its own little nuance and policy there. So,
15:42 so I think you kind of handed that a little bit. But so what, how big, how big of a team did you start with? Was it you and your partner? And there's the two of you like moving fast and breaking
15:50 things or how does work? Yeah, yeah, exactly. It's just two of us at the beginning. That was for the first couple of years. It was just two of us. And that worked fine because we got a couple
16:01 of really customers. They were giving us a lot of feedback. I was able to handle most of the sales myself at that point. And then a couple years in, it's like, all right, well, let's expand,
16:03 let's hire some people.
16:21 And the team size is kind of,
16:27 you know, it's expanding and contracted and expanded, contracted, you know, as people like came and gone, um, you know, we're, we're still small though, right? We're trying, we're trying to
16:36 hire, but man, it's tough out there.
16:40 Yeah. To find people that fit that mold, you know, have the knowledge and the skills for what you're wanting. Yeah. That's why you had something posted a week or two ago. Um, yeah, I, well,
16:52 yeah, I, you know, I posted a job on LinkedIn and then I got mad because all the candidates, uh, were, you know, basically not eligible to work in the US, even though it was US based position.
17:05 So then I just posted it on post and I got 326 LinkedIn messages. Uh, and like, you know, a day and a half and it's like, Oh, what have I done? But it looks like we actually got, you know,
17:16 much better quality people contacting me through that than we did through like the paid advertisement. So. Oh, good. Fingers crossed.
17:27 That's a, that's a, that's. God, the internet's so broken, it's only going to get worse. I'll be like, before it gets better. But so looking at your, you know, your history and what ever
17:38 doesn't look like you were a trained software developer before that. But I imagine going all the way back to your probably masters and PhDs and geology and geophysics, I'm sure you prior to dive in
17:48 on some coding stuff. And what was that journey like and how'd you jump in on that? Yeah. So, you know, I think
17:58 I like to be fair, like when I was in high school, you know, it's when we had typing class, it was on a typewriter, not on a computer. So when I got to college, I had like zero familiarity.
18:09 Yeah. But you knew how to type what's yes, I'm an excellent typist. And I have very proper form on that home row. But yeah, I wish I had. Yeah, I wish I paid better attention I can't remember
18:23 what they called that class in high school, but. Yeah, but you know, it started off in college, like, you know, we had to take like a required programming class. I think I took like a four
18:33 train class. And then, you know, like, they're in my master's and PhD, like, you know, screwing around a little bit and like MATLAB and Python.
18:45 And then kind of dropped it for a while. Came back, did a little bit of stuff and like,
18:53 PHP MySQL back in like the, I don't know, 2010s,
18:59 like, between 2008 and 2010. So I got all the lab stack, like, yeah, that's a good one, PHP, yeah. Yeah. Mm hmm. Exactly. And then I kind of dropped it. And then when we got started back,
19:13 you know, Roland is the real software engineer on the team. And so what it is that we did to make ourselves more efficient is we kind of divided the. the roles. And so he handles all the
19:28 infrastructure side of things.
19:31 And putting, let's say, putting the rails in place. So, you know, if it's moving data around or coming up with data displays, that part lives with him. And the part that lives with me is the
19:44 kind of the technical implementation. So putting in all the petro physics methods, you know, working through like the kind of the analysis that people are doing. So we split those two pieces. And
19:59 that made it, you know, a significantly more streamlined process. Because, you know, I think we first got started, he asked me for an equation, and I gave him a saturation equation. And he
20:11 typed it in. And he's like, Oh, look, now I calculate a saturation. I gave him like a dozen more equations for the same thing. And he's like, Nope, you're doing this yourself Let me just give
20:20 you a way to do it. That's hilarious.
20:24 I welcome to our pain, right?
20:30 That's really funny. What a,
20:35 how, like,
20:37 was that just like an organic thing basically for you guys to kind of just split it there? I mean, it makes sense, right? As far as like the, you're more of the SME on the actual geology,
20:48 petri-physics side and he's more the software person, right? Yeah, I mean,
20:55 I guess, you know, one of the things that we realized early on is that like, we couldn't, you know, when we were talking to customers when we were first getting started,
21:06 you know, they came to us with a giant list of features that we wanted, and it was basically like, okay, well, you know, there's no way we're gonna be able to cover all these bases and put in
21:16 every single option. So it's like, well, we've got to build away for customers to like kind of roll their own solution in a way. And then it's like, well, if we're gonna make it so customers can
21:26 roll their own solution, we might as well make it so we can do that internally.
21:33 So it kind of came about from
21:37 one, not wanting to hardwire in a giant body of technical literature into the infrastructure layer.
21:47 And then
21:50 from the customer side, it just gave us a way where we could divide up how the work was done, where it's like, okay, while you're coding how to make, you know, crawl sections and log plots and
22:03 maps, let me go work with the customer and make sure that we're replicating how you're calculating clay and porosity and saturation and not have to come back and like deliver across technical
22:16 requirements, right? It's like, well, we're two people when we're getting started So, you know, we just needed a way that, like, uh, that we could do things in parallel and not step on each
22:26 other's toes a lot.
22:30 We didn't wanna make it where it's like, there was a giant poll request and review on every
22:39 little equation change. It's like, all right, let's just, I'll have my part, you have your part, and we were able to work that way for like a couple of years before we started to run into each
22:49 other.
22:54 That's pretty helpful. So again, so when you guys started building the app, like, did you agree, like, are you all using Python or anything in particular or certain things on the back end and
23:03 different on the front end or. Yeah, so we, when we started off, we were using Java on the back end, and we were using, I think, Angular on the front end at first, and then about a, about a
23:18 year and a half in, we decided to switch to react So there was a, you know, big rewrite there for migrating over. You know, there's a few things. I mean, it's still mostly Java on the, on the
23:31 back end, a little bit of C for, you know, some specialty things that we need a couple of solvers and libraries. So this C and not even like C is like all the way down to the bones Yeah, sorry.
23:44 No, it would be, actually, some of it is in C. Yes. I think we have a little bit of C and a little bit of C.
23:53 We added Python integration into the platform so users can write their own Python code. So we now have that, it's kind of sandbox off. And then we have the general scripting language that we put in
24:10 YAML that customers used to write their own equations and then it's all react and type script on the front end.
24:17 The main thing was is since speed was gonna be critical, we wanted to have a bit more of a performant language and also we wanted to be typed. Yeah,
24:32 Python can get out of hand. There are typing libraries and everything now, but it's definitely, it's not the same as C or Java or
24:42 any of those that are more robust. You've really tried true for years. So does like most of like all those equations that you're writing get converted into Java or are they making calls to like
24:51 specialized functions otherwise or? Yeah, yeah, it all, so, you know, the scripting language, it's all in YAML files, those will then get parsed into the Java and executed, you know, in Java
25:07 there. Okay. And there'll be a few specialized functions where it, you know, shoots off to a solver that, you know, in something else, but then it'll come back. I've got a question for you
25:18 that you brought up a little bit that I wanna maybe unpack for a second, but
25:25 you mentioned the, you know, you could do this one thing with 12 different equations and stuff, right? How has it been, in our industry, people are very particular about which equation, with
25:39 which methodology, which algorithm, et cetera, that you use or that they use, right? their company or they personally, how have you managed that with,
25:52 you know, at the same time, building a software that has to be able to, you know, account for all of those things, obviously adding Python so they can do it themselves is a big benefit. But,
26:04 you know, I feel like that's another one of those like nuance things about oil and gas that you don't really know until you get under the hood and you're like, Oh, good. Now I've got an account
26:12 for 12 different ways of doing decline curve analysis or whatever, you know, yeah, yeah, that's, I mean, you know, so on the, so on the petrophysics side, because that,
26:43 you know, in terms of equations, that's where I'm kind of most familiar with what people are using on a on a day to day basis. You know, it was, all right, I know the, you know, let's say the
26:44 five to ten usual suspects for calculating any given property that would cover 90. 899 of use cases. And then for things beyond that, we just made it where the user could customize things in a very
26:53 native way. So the way we've set it up is users, when they customize our config files that control the equations, they can build stuff in very natively. They can modify dropdown menus in the
27:08 software, checkboxes, defaults. So if they have their own proprietary method, they can have that show up in the dropdown menu for just themselves or for their company and have that fully
27:20 parameterized and run it natively. So that's help because it's a very simple scripting language and even people that aren't comfortable with code at all are able to get in there and do that. From a
27:35 practical point of view, though, whenever someone comes to me with a request and they're like, hey, I've got this new method, it's like, all right, well, is it proprietary or not? Okay, it's
27:44 not proprietary. So do you have a paper? Okay, let's look at the paper. Oh, okay, this only works, or this is some empirical calibration that only works for this carbonate play in somewhere on
28:02 the other side of the world that no one's working on. It's like, I'll just help you get that in as a custom method just for you as opposed to adding it for everyone But then a few times, customers
28:14 have came in and they're like, hey, here's the method that we're using, it works really well. We look at it, it makes sense, we just add it into the stock config and ship that out to everyone.
28:25 Yeah, that's nice that you go. There is a lot of flexibility.
28:31 Yeah, I mean, there's a lot of juggling requests though, and I will say we've made it a lot easier on the petrophysics side than we have on like the reservoir engineering side. Uh, and that,
28:46 yeah, that's because, you know, a lot of the methods there are, you know, like the concepts can just vary a lot more. And it's like an entire schema changes is kind of needed. So, you know,
28:58 for some of those, it's like, well, someone's making a request. Like, you know, they have, they have to show us that they're serious with a little right, a little bit of commitment, right?
29:06 No, I mean, I think that's an incredibly valid point, right? Like I think a lot of software companies, especially first-timers or early, you know, in their early years, right? They're just
29:16 trying to get revenue or trying to get customers, right, revenue in the tech game comes secondary, typically. But it's one of those things where it's like, hey, like we want to serve you, but
29:28 we also only have four developers or two developers or one developer. And it's like, if you want this, we have it takes away from everything else. And so I'm glad to hear that, you know, I think
29:41 people are very to ask clients to pay for custom stuff, but
29:48 that's the only way it makes sense for both sides in a lot of cases, right?
29:53 Yeah, yeah, I mean, from the business side, I mean, it's like, well, at the end of the day, like their business were business. Like, it's always about money. We shouldn't, it's not like
30:05 it's offensive to ask how much are
30:10 you willing to pay for something. You know, so I, like, I found myself, when I first circled, I'd never done sales. Like, I was a little hesitant to ask for money. You know, so it took a
30:22 while to kind of get over that. But I found, you know, when people can make requests for free, you get lots and lots of requests. But, you know, when they make a request and they're willing to
30:41 put, you know, awesome. some cash behind it, like, you know, it's a, it's a real request that way. And it's something that's actually going to get used. And I'll say like in hindsight, like
30:51 100 of the time when we, you know, when customers have paid us to build something into the software, it's been a feature that gets heavily used. And, you know, when it's someone that's like, Oh,
31:04 this would be great. You know, uh, you know, this would be really interesting It, it's like, Oh, it gets, you build it, you tell them about it, you show them how to use it and it gets used
31:15 once. And then it's like, well, great. Now it's sitting in the code base forever. Yeah. That was, that was worth it. Right.
31:22 No, I mean, I'd cheer a point. It's the whole concept though, is it's like, if it's worth it to them, they will pay for it. Like there is a tangible piece there. And they also have, like you
31:32 have that buy-in that they're probably going to use it because they just paid you for it versus, Hey, it'd be really cool to have this and I know I use it all the time, Tony, up anything for it.
31:42 I just want you to do it. And you know, there's obviously a ton of nuance to that, right? If it's a feature that everyone seemingly might need versus just a very custom one off specific thing for
31:54 a specific client, right? But it's a, it's a tricky thing. And this like, that's one of the hardest things I think about software is being able to prioritize your resources because everyone
32:04 always has. Yeah. Yeah. Yeah. I mean, there's, yeah, I mean, you know, we used Jira for tracking stuff. And it's like, whenever I log in, it's like, Oh my God, like, like, I'm
32:16 scrolling down forever. It's like, you know, it keeps loading at the bottom. It's like, man, if I like circled around or something, I mean, there's always so many requests, uh, you know,
32:25 even just like tiny little things, you know, it's like, Oh, can you make the line weight, like a little bit, you bolder, whatever. It's like, Oh, well, I guess I could, but there's gonna
32:34 be a consequence for that. death by a thousand cuts, but like, yeah, and you bring that up, it maybe left because like, in my experience working with geologists, I mean, it's like color
32:44 palettes and the way the lines, I mean, these are like all hugely important, but different between every
32:52 geologist I've worked with. Yeah, I'm glad you brought that up because, you know, geologists are like incredibly visual, you know, people like all the displays we're making, you know, maps and
33:05 crawl sections and whatnot. I mean, you can have a perfectly functional map, but it just doesn't look good. And people are like, oh, that map's terrible, you know, perfectly functioning crawl
33:15 section, you know, and it's like, oh, well, can I, you know, color my formation lines by different things and fill in between them? And it's like, well, no, but do you really need that?
33:26 Well, I need it for my PowerPoint. You know,
33:30 so that's been something that's hard for me because like, I have no Aesthetic. Sensibility right so you know for for me like when I look at how I worked like back in an operator and how I work now
33:43 It's like go in there get the best technical answer work through it and You know and like I can point on the map like hey, here's where we need to drill right? Here's the zone. We need to target it
33:57 ignore the fact that my my map looks a little dull and You know, it's and it was always for me about the quality the technical quality of the product and You know now when when people are like, oh,
34:09 yeah, I got to draw this in a PowerPoint So can you show me how to make this pop a little bit more? It's like oh boy? I'm the wrong person if you
34:18 Obviously if it's not orange, you know colored like I'm probably not the guy to talk to for it
34:26 No 100 I mean I get that too cuz even you know say my world of more of the BI and data analytics like you know But even now I'm building a spotfire tools for people. It's like I can make it look
34:37 really pretty, but like say if you were gonna pay me to do it, it's gonna take me another, probably twice as long, just to like, and I can do all the things with HTML and JavaScript, but it's
34:46 like, you're gonna get the same answer, whether I give you the one that's stock formatted or not, but if you really wanna make it look super pretty, you're gonna pay me extra, 'cause it's gonna
34:55 take that much longer to nat's ass where the button is and how these things line up. And I guess that's why some people hire UI UX folks also, but it's just, and not to understand it's importance
35:06 because a good UI and user experience is hugely important. But I can totally sympathize with what you're saying where it's like, no, I can build you something that gets you the answer. And that's
35:16 what you wanted all along anyways. So why does it matter how it looks? But
35:21 it's an interesting conundrum.
35:25 Yeah, it's one of those things too. Like there's what people want to do and the way they wanna do it. And a lot of times someone will tell me, Oh, I wanna do this. And it's like, okay, great,
35:36 like we've got a way to do that. It's like three button clicks and you're there. And they're like, oh, well, that's not how I do it in this other software. And it's like, well, you're getting
35:46 the same answer, but with half the clicks and 10 of the time, like
35:53 it sometimes it's like people just get stuck in this mode too of where they expect to do things the exact same way. And it makes switching people over sometimes really tough Yeah, no, that's a, I
36:06 mean,
36:08 I'm an engineer, right? So I come at it from, I don't care if it's pretty, I just want it to be accurate and work, right? But then of course, like my counterpart at Collide and at DW is our
36:18 designer graphics person. And so it's like, she'll put in requests where it's like, yeah, the padding on this needs to be cut down by five pixels or some shit. And I'm like, I don't know that I
36:30 would even recognize that that change was made if we did it. But she notices, and I'm like, thank God, because it looks way better, because of her. But I think you brought up a really
36:39 interesting point that I've run across that, I mean, even while we're building our product, is that people have expect, like especially if it's an existing workflow of some sort, that they're
36:50 already doing somewhere else. But people have expectations because of the prior software tools, experience, whatever, of how things should work, whether that was a good workflow or not, it's
37:04 just is what it is, right? And so like when we're building out the language model stuff, it's like, well, do we wanna, you know, mirror more of the anthropic interface? Or do we wanna mirror
37:14 more of the chat GBT interface or another one or like all this stuff? And it's like most people are using chat GBT. For example, Julie, my designer hates the anthropic interface. I'm like,
37:25 they're not that different, but they are in, you know, those minutia details and stuff And so it's like part of that though, is. Again, just understanding what the users are used to and trying
37:37 to find a way to meet them, at least with the logical elements of it, right? Like, well, first we're gonna do this, then we're gonna do that. The buttons might look different and they might be
37:47 in different places, but just the high level workflow itself is trying to meet them. It's not fun when you have an incumbent that has set all these expectations prior to you. Well, it's amazing
38:00 how quickly people get locked in to like the muscle memory and the familiarity of a tool. Like, you know, mentioning chat GPT and Gemini, like they are incredibly similar, right? It's like, you
38:12 know, the text box at the bottom, you know, it scrolls down, you know, but then, you know, it's like, oh, well, you have the canvas that pops up on one of them, not the other, or, you
38:22 know, maybe it's a slightly different button, and it's enough where, you know, people look and they're like, I can't switch. And then you translate that across to something
38:33 uh, subsurface software that does decline curves, crawl sections, mapping and log analysis. And it's like, you know, it's like, oh, well, I've done this in Petra this way for 30 years. And
38:45 I've done this part over in tech log for 30 years. And it's like, oh, we're fighting 30 years of like inertia, you know, yeah. So I think this kind of brings up, I mean, that it's always
38:57 interesting when we have folks like you on where we can talk about the tech stack that powers your application, but then you are actually part of, you know, a lot of oil and gas companies, tech
39:07 stack. Um, so I mean, like, can you talk about how dynamics like integrates, say, with a, with a customer, like whether it was with their data systems and, or even other, you know,
39:18 applications that they may have, like, is there are any things, are there any applications that you talk back and forth with fairly well or just curious how that all works? Yeah. So I mean,
39:27 there's, uh, I mean, when I look at our place in the tech stack, there's kind of two ways that we can. it in. So with some smaller companies, we're an end to end solution, right? They'll use
39:38 this for, you know, interpreting well logs. They'll use this for their mapping. They'll use for crawl sections. So we're there, their geoscience package, their petro physics package and the
39:47 reservoir engineering package all in one, you know, it's super cost efficient form that way, you know, and it's got the benefit of all the data is in one place. But, you know, to get to that
39:59 point, you can only get there if it's like you're in the door on day one. And so we've, you know, we have been that for a couple of, couple of companies. Now with the bigger operators,
40:14 typically what we find is they have their legacy mapping package that they've been using for years and years and years. And, you know, it's like Charlton Heston, you know, from my cold dead hand,
40:26 so we'll try this out of there. And, you know, and we understand that and. You know, so what we try to do in those situations is we try to slot in as a, you know, as an intermediary, right?
40:44 So it's like, you know, thinking about like working on well logs. It's, you know, they, they'll bring their well logs into dynamics. You know, we'll do some data cleanup, some interpretation,
40:56 and then they'll ship that out to their geoscience package
41:00 And, you know, in order to accommodate that, I mean, there's kind of multiple levels you can think about accommodating moving stuff out. There's, you know, there's just making stuff, making it
41:10 easy to get things out. Then there's getting it out in the right formats. And then there's, you know, actually using some type of integration
41:22 You know, we have really focused on getting stuff out in broadly acceptable format. So, you know, it's like, oh, most of our customers are taking well-long data out. So all right, let's get it
41:35 out in LAS format. So they can just kind of one-click load it into, you know, petra-patrail kingdom, whatever. You know, for production, okay, they want to take it out of our software and move
41:47 it into Aries or something similar like that. Let's make it so they can get the production streams out and the forecast streams out really easy. So get that out in Excel or CSV to them. So that's
41:59 where our focus has been You know, and this comes back down to the, like, identifying, you know, kind of what it is the customer wants, what they're willing to pay for and where the actual pain
42:13 point is. Because we find that a lot of times they tell us what they want and then they go show us what the problem is and they're two different things. It's like, oh, well, what we want is to,
42:27 we want the data to end up in our geo-science package, all nice and clean and tidy up. And we needed to come out of your software and go into that. And it's like, okay, well, show me what you're
42:36 doing. And it's like, okay, well, I come here to my company drive and I spend a week trying to scavenge data off of it to find it so I can put it into your software. And then I spend a day in
42:50 your software, get it right, and spend 10 minutes getting it out of your software and into the other software. It's like, okay, so the problem's not integrating seamlessly with some other, with
43:02 big blues like solution here, it's finding whatever's been strewn across like a dozen different network drives. So it's been one of those things where I find like the biggest challenge that most of
43:17 our customers have actually isn't necessarily with data integration, it's with data management and organization.
43:25 I can absolutely agree with that Thanks. Yeah, John, you're getting into the belly of the beast. with some of these things where you're trying to get stuff into collide and probably seeing how the
43:35 sausage is made a little more. Well, that's, I mean, that's honestly one of the big, like that's the first piece of value that a enterprise rag brings to the table is it's like, tell me what
43:47 information you've got, what documents do you have? Like you can just start asking questions about where your data is, like that's the thing, right? Like the average person in a company knows
43:60 that the document exists. They know that the information lives within that document. They just don't know where the hell the document is. And so it's like, if I can put in a search in a language
44:08 model around that, that's gonna ultimately match with the stuff that's in the document, then that saves me. I mean, we initially started, the numbers we were using was like 30 to 40 and we are
44:23 actively hearing that it's closer to 70 of
44:27 users' time is spent. engineer's operations time is spent looking for documents, which is crazy considering how much money an average salary for those people is and all of that stuff.
44:41 That's where we kind of hope
44:44 to provide a lot of value with these clients, at least initially, is just being able to have know where their stuff is, which is kind of crazy to say all out, but it is what it is Yeah, I would
44:56 say it's similar to what we're hearing is people are like, Oh, I spent two-thirds of my time trying to get the data organized and ready, and then a third of the time doing this interpretation of it.
45:09 To think that when I started 20 years ago, I remember we were debating internally. It's like, Oh, should we have a corporate
45:18 archive of this data where we store everything and I remember the VP at the time was just like well, yeah, who's going to pay for it? Like, you know, we're going to have to pay a data manager,
45:29 we're going to have to like, you know, convince everyone to work through this and we're going to have to have a process and, you know, blah, blah, blah, blah. And it's like, you know, when I
45:38 left the company, you know, 10 years later, it was the exact same conversation going on. We hadn't started anything and, you know, talked to someone there like last week and it's still the same
45:48 thing. They're still debating, oh, we should get a corporate, you know, initiative around data going And, you know, so it's, I mean, it's baffling to me because it's like, well, you know,
45:59 everyone's been talking about it for decades. And other than a small handful of companies, it seems like no one's got there yet. Well, and now they're not nearly as expensive of projects as they
46:10 used to be. I mean, like, I mean, it used to be, like, if you were going to stand up a data warehouse and you'd probably have some souped-up SQL server, or, you know, Conoco, we finally
46:18 moved over to Teradata and these are multi-million dollar projects. Now with some of this cloud stuff and again, at least on the technology side at least, I mean, there's still probably a lot of
46:27 resources needed to throw at it, you know, who have been being wise and there's a lot of the process and technology or people and processes part that needs to be fleshed out but these can be done a
46:36 lot more efficiently, you know, than they were ever able to be done 15 years ago. Yeah, I think the technologies came a lot farther than the people in process each time. 100. You know, it's,
46:49 and I, you know, I see that as kind of a, maybe a cultural thing in the industry, it's right, right? Like I think a lot of the, you know, the SMEs are like very independent, you know, people
47:01 where it's like, hey, I'm doing my interpretation, I'm doing my work, these are my prospects, you know, this is my well. And, you know, because of that, you end up, you know, it's not out
47:12 of any, you know, like, I don't know
47:18 It's not necessarily a territorial thing, but it's like. You know, it's like, oh, well, it's mine. I know where it's at. You know, it's in my email over here. Like, what do I need to like
47:28 put this in an archive? They have ownership over it, right? Like, that's ultimately what it boils down to. But I think the flip side of that though, is that like the way I look at these things
47:38 is it's, these are like, this is tech debt, right? Like at the root of it. And it's like, you're gonna pay for it at some point in time, right? Like we just kicked off a project with a client
47:47 who's sent over a bunch of bankers' boxes to get the documents digitized 'cause they don't even know what's in them. And I know that that exists at every single operator out there today. And it's
47:57 like, guys, like, if you wanna use the data, you have to have the data. And so if you want it digital, you gotta, whether you do it now or five years from now, you gotta digitize it at some
48:09 point, right? Like, or you gotta put it in a structured storage system somewhere if you're gonna end up trying to use this stuff. And so it's a weird kind of conundrum, of course, You know, of
48:21 course, all of that costs money, but it costs money on the back end to go back and do it and redo it, or you're missing out on those shit ton of opportunities in the meantime, because you can't
48:30 find anything. Yeah. Well, I think I think that's, you know, the exact thing is you're going to pay for it at one point or another. But I think psychologically within the company, if you say,
48:40 Oh, we're going to spend, you know, 2 million or5
48:47 million or whatever to create this gold standard database that holds all this information, you know, someone looks at that as a line item. They're like, Oh, you know, 5 million bucks of GNA on
48:55 this, you know, don't, don't ask me for that one. Oil prices starting with the six, you know, and, uh, but then they're perfectly, they're perfectly fine spending that equivalent amount of
49:07 money, you know, in, in lost productivity on salaries, right? Because your reservoir engineers and your geologists, you know, who are high paid technical experts Spend all their time like
49:19 searching for data. Instead of actually doing the engineering, you're paying them to do yeah, I think that's that's the big one right is it's like again You're gonna pay for it regardless Either in
49:31 people's time or for actually going and doing it But the problem is is that it ends up getting looked at by someone who doesn't Truly understand the value of what? Whatever it is brings and it's also
49:45 very hard to have a tangible value of like if we do this then this will happen Normally, it's not a revenue goes up or costs go down Which is what everyone makes their decisions off of it's a well
49:57 shit's just gonna be easier to find so people Can actually do their jobs and so it's like yeah, that's it's a hard ROI to justify Without numbers and stuff behind it so I think this leads into
50:12 Interesting kind of cams got some opinions here should
50:19 SMEs and reservoir engineers and geos learned a code or should or should you know, or should it be required? I guess it's what where some of it came from. Oh Dagger to my heart right there. Um,
50:30 you know, I've gone back and forth on this a lot, right? And I'll tell you where I've settled on this, which is I think that you should Learn to I would phrase it as you're learning to write
50:44 scripts and not learning to write software Um, because I'll I'll tell you what what I see all too often is You'll see someone they're like, oh, hey, I came up with a little python script of I
50:58 don't know Run a machine learning model to I don't know calculate shear logs Okay, great that that's valuable run those shear logs dumb amount of python throw them into your gng software and use them
51:11 Get perfectly good use of scripting uh unacceptable use is Oh, look, I wrote this little script to do the same thing. And then I wrote 5, 000 lines of code to force matplotlib to make a, you know,
51:26 a crawl section of three wells. And oh, by the way, if you wanna see a different three wells, like - Yeah, it's a hard code. It's gonna take another few hundred lines. Yeah. Yeah, I mean, so,
51:34 you know, like, that is where I think there's a big disconnect. It's like, well, okay, you were creating value with applying a machine learning method to generate this prediction, to fill in
51:46 some gaps in your data coverage. Stick with that. That's where you're creating value. You're not creating value by reinventing the will to create a basic well-logged display, especially when you
51:58 spend 90 of the time making that well-logged display, that you can do better in your commercial software that you've already paid for. So that's where I've settled on, like, the learn-to-code
52:08 thing. Yeah, I think that's totally valid. I agree I don't think enough people mention the term scripting when they talk about coding. And, you know, I completely, I consider myself a script or
52:20 none of my code ever touches prod in any sense of the word, but I use, I mean, that's the thing on the data side, you use scripts all the time, you have to write to do a lot of things. And so I
52:31 think that's the natural connection as to where people get into this, especially in our industry is it's like, oh, well, I need to pull this frack data or I need to pull, you know, this
52:40 subsurface data or whatever and it's, it's available via an API Like, get it. Okay. Well, now I've got to use a Python script to go pull it and then do some stuff with it. But then the other
52:50 piece of that is you don't want them writing software that
52:54 ends up having to be maintained and deployed and all of that stuff because that's a whole another can of worms that you as a oil and gas company have no business dealing with in all honesty in most
53:04 cases
53:07 That's a it's probably one of the most frustrating parts of the industry I feel like is you get these a lot of the super majors are pretty notorious for it but it's like yeah. bring in a bunch of
53:16 companies, learn what they're doing, learn how they're doing it, and then just go build it themselves and then hope that they can maintain it and keep doing it for however long they're going to do
53:25 it. I
53:28 mean, as I was going to say, when it comes to like scripts as well, like I don't think companies should hesitate to ask their vendors, like, hey, you know, we've got an internal script that
53:37 does this, you know, here it is, make it work, you know, and we'll deploy it by basically just giving away this script, right? I think too many companies hold like completely unproprietary
53:52 things, like way too close to the vest, and that forces them to then maintain some little widget that, you know, gets used once in a blue moon. But of course, the other thing is, and I, you
54:03 know, like to hear your guys' opinion on this, like, I don't think all geologists and reservoir engineers should have to learn how to code. I I think there's,
54:14 I think, you know, there's enough value creation for being an expert reservoir engineer and expert geologist that, you know, if you choose to focus on that, that shouldn't necessarily be to your,
54:25 to your detriment. Yeah, yeah, I 100 agree. I mean, I've worked with plenty of excellent reservoir engineers who are just really good at using Aries and or, you know, whatever they had at their
54:35 disposal and know how to make good decisions with it. And again, maybe there are probably some things that they could have been maybe a little more efficient at if they knew how to do that. But if
54:44 they had pulled the time off to learn how to code, but it have made them less effective or learn how to script rather, even that would have probably made them less effective. So yeah, I mean, and
54:53 there's plenty of extremely experienced folks that have a ton of knowledge that you'll - and they'll never going to touch an IDE or maybe not even VBA scripting. But they're going to be just fine and
55:06 add a ton of value. So it should not definitely not be a requirement. I agree with you too Coding is a scripting, coding, whatever. even the tool, the softwares that you mentioned, they're all
55:18 tools, right? Like you need to be an expert in whatever your job is. And it's like, if you can solve, it's just, you know, physics, the same thing, right? Like there's five different ways to
55:28 solve this physics problem. So which tools are you gonna use to solve it? And it's kind of the same thing as far as this is. As long as you can get, you know, the right data and good answers out
55:38 that are, you know, reasonable and sound engineering, then it does, I don't think many people should care about how the sausage gets made at the end of the day. But I also agree with Bobby that
55:49 there's a lot of times, and that's honestly how I got into scripting is where it's just like, I'm so frustrated because all I need is this one little piece of additional information that isn't in
55:60 whatever tool I'm using. So screw it, I'm gonna figure out a way to get all of that data into one place. And that ends up being, yeah, a lot of
56:11 the ways people get into it is just solving, solving problems, right? what we're trying to do. And so these are just all tools to do that. I think some of the problem is that now all of these
56:22 undergrad programs have them take some Python also, which I think is good to expose them. But at the same time, now we've got a couple hundred reservoir engineers, fresh psychology have Python on
56:34 their resume. And it's like, do you even really know Python or did you take a, you know, one semester course on it?
56:42 Yeah, yeah, I mean, you know, and it's one of those things. I mean, there's, I view there as being like different levels of comfort that add different levels of value, right? Like,
56:54 you know, if you're comfortable enough that you can write a tool that's broadly useful, then like, that's great. You're adding a lot of value. But, you know, just knowing just enough that like
57:03 someone can hand you their script that's written to do some task and you can go in and point it at the right directory and make it run. You know, I think that's, you know, for a lot of people,
57:14 that's probably good enough. It's like they can look at the code and say, all right, let me update this one variable and run it. So, you know, I think there's levels to like the coding thing,
57:26 scripting, coding things yourself. I also think part of that as a function of where you're at and like the maturity of the technology, right? Like Bobbi and I, I mean, we grew up with the
57:37 internet, like as the internet evolved as well, right? And so like, Every millennial knows how to tweak things in HTML because of MySpace and all these old platforms that we all used to use. No
57:49 one coming up now has to learn HTML because all the software is good enough that you don't have to. And so that's where I see, it's a weird thing. And of course you see this with GPT and stuff
57:60 where it's like, you're gonna atrophy in some areas or not have to learn things because technology is just better in UI or the UX, all of that ends up being better and so you just don't have to
58:11 learn it. Defragging hard drives is another personal favorite of mine. Like, no one anymore does that because the software just does it automatically in the background, I think I don't even know.
58:21 But it's just like, as technology evolves and matures, they get better and they automate things away from the user that they don't have to know how to do those things. And so there's a very real
58:32 future where they don't have to know any of this stuff 'cause the software just does it for them or they can ask a language model to write the script that. goes into the software that immediately
58:42 works. Yeah, it's
58:45 yeah. I mean, you know, when it comes to the software just doing stuff for you, so you don't have to learn anymore, you know, that's one of those things where, you know, I think as an industry,
58:55 we have to be really cautious to make sure that people understand the fundamentals of the tools that they're using. Because, you know, some tools I'll admit are just so easy to use. It's like,
59:07 right, you know, it's like, oh, but they didn't get an answer out. And there's just a couple of checkboxes in between. And, you know, it's like, well, I got a great answer. I don't know how
59:17 it got there. I don't know if it's, you know, like, I assume it's right. Like, software told me it's right. So it's got to be, you know, that I think when I look at, you know, a lot of new
59:27 hires in the industry, I think there is a danger that it's like, there's not enough, like, mentorship and early career, you know development to to make sure that when they put data in and get an
59:40 answer out so they know, you know, what's inside the box, but uh.
59:47 So I think one last thing before I hear we're running up on the hour actually already, which is always crazy to us, but um, I know you're fairly opinion on AI. So like, how is AI going to
59:58 revolutionize dynamics or even SAS in general and how is it not going to? Yeah, um, well, like, so, you know, for dynamics, we try to sprinkle in machine learning where it makes sense. So I'm
1:00:11 not even going to call it AI. It's just, you know, it's like, hey, you know, we can make our customers like do this way faster if we just sprinkle in a little, you know, K means algorithm or
1:00:22 random forest or linear regression here that's just automatically dumb for them. So they don't have to consciously do it. Like the answer just shows up and they're like, Oh, well, that's great.
1:00:32 So, you know, we view it internally as a productivity thing. like we've recently been working on like an assisted tops picker for geologists and you know it's like you just sprinkle in a little bit
1:00:44 of like AI and machine learning and it can make you significantly faster and I think that's you know we'll probably continue I think it's an industry on kind of a glide path that way you know where
1:00:58 things will just get easier. Now I think in terms of in AI and like the stuff that you guys are dealing with where we've got these big LLMs I'm I don't yet know or I don't yet have a good feeling for
1:01:15 where that's going to go because it looks like there's almost endless potential but it always seems like just out of reach right like uh you know I saw someone on LinkedIn they were talking about like
1:01:29 oh hey can I get you know my LLM to digitize a a well-logged for me. And it's like, okay, yeah, okay, it can read the header. All right, well, that's cool. You know, you can put it in nice
1:01:40 JSON structure for you or something like that. But it doesn't actually digitize the logs, right? Like it doesn't know what to do. And it's like, well, you could see a future where like you just
1:01:52 tell it what you want it to do specifically enough and it's gonna do it. But it just right now, it just seems slightly out of reach And I think until there's like, you know, I think what would be
1:02:09 revolutionary is if you had an AI that could essentially stare over your shoulder as you're doing the job. So you're kind of like, you know, imagine having your AI watch a, you know, a 20 year
1:02:20 experience reservoir engineer or geologist go about and do their job and, you know, watch it, you know, recording the screen and the motions and the interpretation and, you know, And then after
1:02:31 doing that for, you know, across the workforce for a couple of years, you know, it being able to, you know, then being able to spin up and, you know, agent a Bob the geologist agent, right?
1:02:44 And have it, you know, pick tops for you and make maps and do a lot of the stuff and, you know, make your, you know, one geologist be able to then do the job of 10 or 15
1:02:57 But, you know,
1:02:60 my guess is that we're probably a lot farther away from that than I would, you know, than most people would think. I would say, you know, a decade or more. Yeah, I think people, again, people,
1:03:14 as with most new technology, especially when you have the fang suite of folks pushing it,
1:03:22 they over-hype the hell out of it But then of course the reality is somewhere below that. But the
1:03:31 you know, again, a language model to me is a tool, right? Just like a machine learning model is a tool, and big data is a tool, and stats are a tool, and all of those things, right? So it's
1:03:40 like the language model can help kind of orchestrate these things, but it's still gonna be going in and writing Python, or using a machine learning model, or using all of these things that we're
1:03:49 already doing. And so, but it is, to that same point, it's this kind of compounding effect of, oh, well, we've got machine learning models, and now we've got a language model, and if we can
1:04:00 figure out how to get the language model to understand how to run machine learning models, then that makes it better, because now I don't have to do the machine learning model. Those things kind of
1:04:10 build up on top of each other. And so, I think it is, I agree with you. You know, I don't know a lot of engineering workloads that we want, agentic, or even if you could do it agentically,
1:04:21 automatically today, that you would have any confidence in as a user or as a customer.
1:04:29 you know, that's where the human and loop, I think we're going to see years of human in the loop type of things that end up ultimately becoming these, you know, automatic things. But, you know,
1:04:39 going and searching a website is very different than going in and picking tops. Yeah, which even going in and picking tops is like far more simpler than like the operation out of rig, right?
1:04:52 Exactly. You know, I think we're, I think we're a long, long, long ways off from having humanoid robots out there as a rough neck, you know? But, you know, I'm hoping someone will prove me
1:05:04 wrong, you know? Yeah, but Cam, they don't have to drive and they don't, they can't do drugs and they can work 247 and, you know, all this fun stuff that they're going to be pitching us, right?
1:05:14 It's a, yeah, the
1:05:17 future's going to be crazy. They could also, they could also turn on us and
1:05:22 be end of days, but,
1:05:25 well, awesome. We want to get into quick little speed rounds.
1:05:29 Yeah. So to camera, are you coming in town for her tech, it looks like? Or. Yeah. I'll be flying in on Saturday. So. Nice. Sorry. So when you, when you, when you get into Houston, you
1:05:40 know, a couple of times a year, where do you need to go eat? Uh, Lupe's Lupe tortillas. That's a good one. You know how many of that good Mexican food out in Dublin? Uh, believe it or not,
1:05:52 not really. I'm a big facetious. They don't, they don't have scotch margaritas out there No, it's certainly not terrible, by the way. Uh, you seem pretty well travel cam. What's your, what's
1:06:05 your favorite place that you've, you know, ever been or lived?
1:06:10 Uh, boy, that's a good question. Um, you know, for living, like, I will say, I had a really good time. Uh, when I was living in Egypt, like great lifestyle there, you know, I was working
1:06:20 for a good company, uh, with a good group of people. I mean, that was awesome.
1:06:26 So in terms of favorite places to live, like that was high up on my list. Yeah, that was before I had kids though. So they'd be a bit different now. In terms of favorite places to visit, I've
1:06:39 always been really partial to the mountains. So I think the Canadian Rockies are great for hiking and for skiing, like the Alps are definitely the place to be. Yeah, we went up to Calgary for my
1:06:51 first time up in Calgary and Banff area last year for a event we did. And I was like, Oh, I get it, it is beautiful. It is crazy up there. All right, so if someone's coming to visit you there,
1:07:05 what kind of food do you have to take them out for? What's your favorite food in Ireland? If there is one? Blood sausage, obviously, right? Yeah, if someone's coming to visit, like, we'll
1:07:16 just cook at the house, be honest. But no, I mean, I guess like pub fare, like you know fish and chips and again us you know. That's kind of what you'd say it's best known for a year. Yeah, my
1:07:32 son loved the Chicken Gujans, which for anyone who's listening are just chicken tenders. We're fighting nuggets. Yeah, but he loved them. Whatever they were doing there, we need to find them
1:07:41 here. But no, good times.
1:07:46 On the technology side, what's your favorite open source package, do you have anything that you guys will use that's relative to even though we're on gas or?
1:07:57 I mean, in terms of open source, I mean, I guess like, I think scikit-learn is, it's like for a Python library, I think that's incredible. Same for NumPy, to be honest.
1:08:12 In terms of open source software, not really sure. I mean, there's so many packages out there. No, scikit-learn is a good answer Yeah, I agree.
1:08:26 We'll wrap this up with the last one. What's your favorite kind of Azure service that,
1:08:33 or GCP or, sorry, we're Azure. GCP,
1:08:37 what's your favorite thing about GCP? Let me just, what do you think?
1:08:42 I really like Datastore and Firebase. I think,
1:08:48 just the fact that you can like, set up auth with Firebase, tie it into a Datastore, you can stand up like a flask website and like an hour. Yeah, so it's like pretty wild, yeah. Yeah, that's
1:08:53 a good one. We looked at Firebase quite a bit until we decided to go on the Azure
1:09:09 side, but
1:09:12 this has been great, man. We really appreciate you. I know there's also a time change, so I appreciate you taking time in the end of your data. to hang out with us here in chat. But where can
1:09:22 people get in touch with you? Where can they find you if they wanna reach out and learn more? Yeah, yeah, so you can go to dynamicscom, send us a message to their website. You can also reach me
1:09:35 on LinkedIn, just search for Cameron Snow, Denomics, and I'll show up there. So those are probably the two best places to get me. Awesome Okay, I appreciate it, man. Yeah, well, thanks guys.
1:09:53 It's great talking to you. And, you know, I don't think I said anything I'm embarrassed of, so hopefully it's not too dull. Good podcast. No, I think it's great. Yeah, it blew by. That's
1:10:03 normally a good sign. All right. Appreciate it. Bob's, I'll see you at the happy hard night. Yeah, man, sounds great. See you guys. Appreciate it.