Welcome to Energy Bytes with John Kalfayan and Bobby Neelon. Your essential guide to the intersection of data and energy. This podcast dives deep into the world of energy, shedding light on how data, AI, and technology are revolutionizing this sector. Each episode equips listeners with insights into the most efficient tools and resources, paving the way for a data-driven future in energy. From technical nuances to broader industry trends, Energy Bytes offers an unparalleled perspective on the evolution of the energy industry. Join us as we decode the algorithms of energy, one byte at a time.
All right, guys. Welcome back uh to another wonderful, not exhilarating, but wonderful uh and exciting episode of uh of Energy Bites. John Calfan here with my trusted partner Bobby Nean. What's going on? How you doing, man? Doing great. Doing good. We got [snorts] Ted Cross today. How are you doing? Happy to be here. Doing great. Ted's uh the VP of product over at Nobby Labs. That's right. Got that right. I remembered. Um, sorry. I was We went to the rodeo cookoff kicked off last night and that was the first Sarah and I It was like the We haven't been since Collins was born, which was seven years ago now. And I was like, how has it been seven years [laughter] already? I went the one year and your dad was cooking. Yeah. Well, that's every year he's on a team. Shout out to Buck [clears throat] and Wing and those guys at the cookoff. [laughter] Thanks for letting us come hang out with y'all last night. But yeah, I don't normally go to bed at midnight on a [laughter] Thursday. We haven't the data in whiskey chat was up till 12:30 if you were if you were on the real ones. Yeah. Shout out to eight of us hanging out. Everybody uh on the data whiskey chat. We uh we started kind of sponsoring that. So for those of you who aren't aware, check it out. Was it once a month? Uh they try for a month probably at least once a quarter. Once a quarter. Yeah. But really follow Jeff Hughes and he'll make sure you know. Yeah. Bobby and I will repost and and share all that stuff as well. But [snorts] we're gonna mix it up a little bit. Instead of jumping right in, we're going to kind of bring up some uh current tech events real quick. Uh the first one I want to just point out because I appreciate a good troll as as uh good as the next guy. Um shout out to all of the government services who took down the uh [snorts] I think it's Bit Locker or Bit Lock 3.0, know, but it was a hacker group that would go in and they would, you know, essentially lock people out of their enterprise accounts and hold them for ransom. You know, pay them in Bitcoin and then you'll get your stuff back. But this group took it a step further where they had a public site where they would actually post the company's names with their logos and they had a timer and if they didn't pay by the timer, they would release all of the documents that they had. Well, so the uh it came out I think this morning. Um, I saw [laughter] on Instagram the uh a number of different federal agencies took them down. They locked them locked out of all their crypto accounts. They've now taken over their website as well and done the exact same. They've just completely trolled the [laughter] hackers. They've put uh they actually released their encryption key on the website. They have uh they now have a countdown that's counting down the time before they release the hackers personal information. [laughter] It's like it's a it's a really great a little spice coming from the government there. Yeah. I mean, it's scary, but it's also kind of funny at the same time for just the troll aspect. You figure the people that the government is probably hiring for that though. Oh, yeah. You know, they're [snorts] they used to do that. Yeah. They used to do that. [laughter] Like that's all these You have people That's what penetration testing. You have hackers that are just trying to do ethical hacking, right? You know, like Yeah. No, I'm sure most of them have been caught hacking and now instead of going to jail, they work for the Yeah, it's a it's a good game. catch catch me if you can when they you got to flip them and like hey [snorts] help us game theory is a a crazy thing prisoners dilemma when it comes down to that but uh I I just I love that we're getting some sauce out of out of those guys that's fantastic and then of course the the really exciting one or funny one I guess in the the tech space this week is the uh the Gemini what do we know is it the same name for their vision model or I don't know I'm not sure wrapped up in Gemini yeah so the Gemini uh vision update came out this week and it's apparently [laughter] just like overly woke and uh if you ask it like a picture of the founding fathers, it returns a bunch of people that are definitely not what [laughter] the founding fathers look like. I was watching too much Hamilton. Yeah. Right. [laughter] But I just we you know, like I said, we're trying to bring up some some uh relevant kind of tech topics at the beginning of the show, but I know that stirred some things up for for Ted, you know, historically. I love all these AI image tools online and love to find, you know, poke in their find the corners of them and places where they'll run into issues. But I mean, when I saw this, I thought this was crazy. But I always leap to like, okay, well, what can we do for like oil and gas related things, you know? So, um, I I got it. I asked it, you know, draw me an image of a protester in France with a just stop oil, you know, spit that out, including like it did like an oil chic like holding up like someone from the Middle East like holding [laughter] up holding up a just stop oil, which I thought was creative, right? But then I asked it, okay, well, do the same thing but for more LG. And it's like, sorry, I can't do that. You know, I don't advocate for fossil fuel usage. I was like, okay, do it for nuclear. It's like, no, I can't do that. I was like, all right, how about more lithium mining? goes, "No, mining is bad and potentially dangerous and I can't [laughter] render that." So, I thought all that was crazy. And then, you know, not to get too political, but you could ask it the same thing, like have more children and said, "No, overpopulation is dangerous." Like, I can't advocate that. It's like pretty crazy the level of instructions that they put into it. Um, and you know, it's it's it's just worrying to have an institution as important as Google like be putting clear like bureaucratic [laughter] thinking over over this. Yeah, you gave me uh I think it was yesterday when I I saw that thread come up and uh you gave me a nice dose of anxiety considering pretty much most of our my data storage is on on Google Cloud. I was like, God bless. This is why we won't be on a a major cloud platform at some point in but I mean they came out it was what two or three years ago they came out and said they weren't gonna help with any more models to help with extraction. But then like two months later signed a contract with Saudi. Yeah. Yeah. You know it's like money talks guys. Uh shock shocker. They're still in business to make money. No it's it's crazy. I mean, I think I think this for whatever reason, the folks working on the image side of the Gemini model put in like much stronger limitations. Like in Google's defense, you could go ask the Gemini text model to like give an give a defense of the use of LG and it would give you like yes, it can help you switch from coal and it's like a reliable base load power source like all the stuff that you would like expect. Yeah. Um but the the the image instructions were just absolutely crazy. Yeah. It's probably like overcorrecting because I know it's it's always been kind of a with with any of the image training I think they were like well again it's it's biased towards what you've trained it on. So if you've only trained it on Americans or only trained it on like certain races obviously it's going to be biased that way. So they feel like they just swung way too hard the other way. Yeah. I mean, they I don't know if it was an error in their testing to not run it through because it's it's not crazy to want to, you know, hey, make me an image of people working in a coffee shop or, you know, a boardroom or something. You want to have like not just white dudes in that like that's that's like that's pretty solid. But, you know, to put in draw me a picture of the medieval king of England and then like literally a black woman, you know, [laughter] well, maybe it's uh not exactly accurate. Yeah. Yeah. Yeah. Well, and you you did bring up a point earlier before we started. Is it's like is may maybe or somebody did in the office. I don't know who exactly, but is this maybe like an intentional thing to drive more traffic so that pe more people use it so that the model gets trained faster? Could be. Um yeah, I don't know. That's a weird That's a a kind of a brilliant or completely disastrous like that could go one of two really good ways or really bad ways. This one feels like it went too far. [laughter] Well, just like no such thing as bad publicity. Yeah, this would not count as bad publicity. It's uh No, it's it's it's crazy all of this stuff. But yeah, it does highlight the point that these models are only as good as how they're trained, right? Like they're not magic. They're not uh but it's also a thing to talk about that people continuously need to keep talking about. That's why open source is such a good thing is because it doesn't give power to an individual. Um, [snorts] and yeah, you know, as we're all Bitcoin maxis in the office and stuff, too, for many reasons uh around that stuff. So, it's it's it's interesting to kind of talk about that and see how that evolves. And so, you guys like that we'll keep kind of talking about some of the current events on this stuff. But, uh, it was just relevant because of course your your thread yesterday blew up on the internet yesterday. So, um, well, [snorts] tell us tell us about yourself. How uh kind of how did you you know get into energy and tech? Were you know where'd you grow up? How did you get were you taking things apart messing around on computers or were you you know it's it's the the the theme across all the episodes is there is gen just genuinely not a traditional background which is really cool which is one of the reasons one path. Yeah. Yeah. [clears throat] Well, I definitely probably have come into this from not a traditional background of like software engineering or anything like that, but um I'm from San Antonio and I um studied geology in college. Um so, you know, I grew up like there's a little bit of oil industry floating around. You know, my grandfather worked in the business. He was a chemist for Texico actually. So, I still have his patents on some loops. That's awesome. Yeah. Pretty sweet. Um, and my my dad in invested in some bad oil and gas wells, [laughter] but I have an amazing picture of me and my brother out on some some drill site. It was a a dry hole when we were uh, you know, six years old. Pretty sweet. Yeah. You didn't know any better, which is great. Yeah. Yeah. Exactly. My first dry hole. That was the [laughter] Yeah. Um so you know I studied geology and and went into the went into the business after getting my master's degree and went to Kico Phillips where I did five years there and um you wasn't really at all on the digital or tech side like other than just the fact that geologists are looking at logs or interpreting seismic or making maps. Um, and I was really interested in decision making and in forecasting and I had I had like a uh you know what's the the red pill moment or you know more from from the matrix where um somebody showed me a crossplot. This is the only only podcast where I can wax poetic about a single crossplot that like changed my life. But somebody showed me a crossplot of um globally the productive area of field on the x-axis like oil fields like it was two square kilometers or 10 square kilometers or 200 square kilometers and on the y- axis was the um reserves of the field. Um, and so you could you could say, you know, for a field that's five square kilometers, like we'd expect it to have between, you know, whatever it was, like three and 10 million barrels, you know. I'm just don't remember what the numbers were, but, you know, so I was like, "Wow, this is incredible." You know, so like I was working in exploration at the time and we'd have prospects we were looking at and, you know, we'd get numbers back. I'd be like, "I don't know. feels like a lot, [laughter] you know, or a little, you know, but it's like so um this was just wow, we can actually ground like our exploration expectations in this like global database. This is fantastic at how we can understand and and make better decisions ultimately as a company for where to where to invest our our precious exploration capital. Yeah. Well, I think that's one of those things that like so many people outside of the industry just don't think about is it's like you can go bankrupt real quick in the oil and gas business because it's super expensive. It's super risky. You know, to your point, you can go drill a dry hole and especially exploration, right? Yeah. Oh, yeah. And no one's hardly I mean, I'm sure some people are, but there's not a ton of true exploration going on now. It doesn't feel like relative to what it was limited. Yeah. Definitely not nearly as much onshore anymore. I mean, you've got the, you know, the big offshore stuff that is basically always going to be kind of a staple baseline, I feel like, especially internationally. But yeah, you know, compound [clears throat and cough] compound, you know, in my opinion, bad energy policy along with depressed energy prices for long enough and then guess what? Energy companies quit spending money on exploration and then guess what? Oh, we now production. I mean, I remember at Kico, I mean, I think what there was like right when everything went to crap, then there was like a $400 million dry hole. Like, you can't afford that. Like, that's crazy. Yeah. I mean, yeah. And when you can when you can basically manufacturing mode and all your, you know, Entro basins and at least have some certainty as to what you're going to produce. Well, even the early days too, right? Like just thinking about how expensive some of the wells were that we fracked back in like the early 20110s even, right? Like, you know, we were selling resin coat or ceramic prop for a dollar a [ __ ] pound, you know? Like, and no one uses we didn't even need that, right? Like, we were washing out all of our frack fluid ends and stuff, pumping Hannesville jobs with ceramic and cross-link gel and just like none of that happens anymore. It's like changed a lot. Yeah. You know, it's it's just such a different uh thing. But even now, it's like, okay, we know where a lot of these shale plays are at least. And you know, there's a lot of those plays that we know of that either are were too expensive or are still uneconomic at current prices. This is like this is my issue with any kind of peak commodity, not just oil, but any kind of commodity is the minute. Yes, you at a certain price there is a peak. [snorts] Y but that's not how commodities work. When supply goes down, demand or when demand goes up or supply goes down, the price adjusts. And so if price gets above a certain point, those uneconomic things now become economical. And so now you have a whole new swath of reserves that that's like shale anyways, right? I mean like right. No, that's exactly what they always knew there was oil there. Just like matter getting it out, right? The technology and then just the money to spend. They've known those were source rocks for years or decades. There's just so many examples of that of you know just like I love the uh the comtock western Hannesville um results coming back. just like absolute monster wells in that cost like what 30 plus million to DNC you know and I mean there's other examples of I mean the whole western US like almost every basin like you can produce gas from horizontal wells it's just how are you going to compete with right Marcellis it's tough but eventually you know and you need infrastructure and all that yeah they need infrastructure and all that but in 30 years you know that 6c well that you can drill in the mangus in in almost any basin like might might be looking really pretty, right? Or at $5 people MCF, you know, like that's the thing. That's what I you know, gas. It's like enough people start buying up and sitting on the gas and they just finally let that associated gas peter out over time, which not going to happen if you keep drilling new new oil wells in the Perian. But uh like yeah, how do you keep operating at $2 gas for like that's that that to me also actually just shows like how resilient our industry is. The fact that Nat Gas has been what an average of maybe $3 over the last 10 15 years. much since I've been in the industry. I came out like I came in at the tail end like on the downside of gas prices when the Favville because I was working in the Fateville where people actually wanted to work in the Fateville and we're exploring in the Fateville, right? Like no one is exploring in the Favville now because it's you know but it's a super easy play to like operate. It's shallow, low pressure. Yeah. And so it's Yeah. All that stuff's fascinating. But so let's [snorts] kind of dive in because I think there's kind of an important thing and I don't know if we talked about it much but like you said your kind of red pill moment or whatever. Yeah. Um, I feel like, and correct me if I'm wrong, but it's almost like kind of that marrying of like kind of a statistical world with a physical world because you're I mean, I imagine as a geologist, you're looking at logs and a lot of things are probably more physical type parameters to Yeah. to, you know, come up with the evaluation and then it's like, wait, we can do some of this statistically too. And at the very least them together, they're great. But, you know, Yeah, for sure. And that's I think a lot of the reason that was of interest to me is you know you coming from like a conventional background and then moving into unconventionals especially from a geology perspective I just there's so much all these assumptions and techniques that have been kind of built up for the conventional world you know let's just take a simple example like a play fairway map you know you might if you're analyzing for a conventional prospect you might go well you know I'm going to need mature source and I'm going to need a certain amount of sand presence and you know these depths I can drill in whatever you come up with like stop lights like red, yellow, green, you know then you go we'll go lease up the green and go drill the prospects there. It's like pretty straightforward but you know you go into an unconventional play what thickness do you need of reservoir there's so many variable like just thinking about it purely like mathematically from an engineering perspective one shale well has so many knobs that you can tweak on it and that's just okay assuming you got the geology part right which that's a whole another can right and so it's like one fract job okay how many purse what phasing how many what's the distance between the purse how much sand what pounds per gallon how do you ramp it up [laughter] what rate what fluids there's so much that goes into and the heterogeneity I mean one that's what I'm saying like that's not even including the rock part of it right like there's so much more on that and like Just to give you one concrete example of that, like a lot of times like here at Novi, we analyze all sorts of geological data to try and find trends for shale production, right? And one thing that often surprises people is our models will say like actually you want thinner reservoir, not thicker reservoir. Like that's that goes to your intuition. But the fact is in many systems, not always the case. If it's thinner, it means like the organics maybe have been concentrated smaller area. And so you're draining, you know, 60 feet of good stuff rather than 300 feet of marginal marginal um marginal rocks. Yeah. And it's just it's counterintuitive. You know, you explain it to someone, eventually they'll probably understand it, but you have to look at the data to come up with your, you know, new models of thinking about. Well, and that's the other thing too, right? is like, you know, if you have an unlimited budget, you can generally speaking do most anything you want. But it's one of those things, too. It's like at the end of the day, we are still a business, right? And so like, yeah, I could spend 30 million and make, you know, a whatever 5,000 barrel a day. Well, but at $50 oil price, that doesn't [laughter] work, right? Like especially with the decline curves and all of that with all the shell stuff. And so that's a whole another like a thinner reservoir with better quality rock means that I don't have to put as big of a frack on it. I don't have to have as many stages. I don't have much sand or as much horsepower. Like there's so many other components of that because it's like okay the frack design you can optimize the frack design for production but then you have that cost element too where it's like oh yeah also the economics have to work. Yeah. And well in the industry is some people are still in the whole like bigger is better type thing but the you know last last five years after the big crash in 2014 and then going through another round in COVID has been trying to figure out like how to optimize my economics. Yeah. No, I remember selling frack jobs uh in I was in Chesapeake's office and they had just had like the uh there was like a press release and it I asked them how it happened and they had like you know Jeffrey's analyst or whatever in the office and in their hallways they had like pro kind of like at a school you've got projects up on like boards in the hall and you can see and there was a there was a pad they were doing I think it was in the Hannesville it was either Hanesville or South Texas but it was a like mega pad or something and it was like I I can't remember 5,000 lbs a foot or something stupid, right? And it was just like this [clears throat] big deal and they, you know, had a poster of it. Well, the analyst saw it and then wrote a full wrote up this whole article on it and put it out and all this stuff and but like that those early days where it's like, yeah, we just kept increasing the pounds per foot. But I mean, but there's always a diminishing return. It's something. But that's the whole like the prito cur you know like engineering you've got to push the limits and see you know I know I think we did like a kitchen sink test in the eagle herd where it was like let's do ceramic proper and this and like throw as much as we can and see what we get. Yeah. Yeah. Well that's the other part of it too right is it's like well do you throw everything at one or do you try and like scientifically change one control? And then we went to that though. Then we did some really high water, you know, ratios, you know, and but it's Yeah, there's a lot like Well, and all of that is valuable, but there's also I think a lot of folks underestimate like the natural variability of well results. And it's just it's so hard to make a decision off of like two wells where you just varied one one little thing and in the end doesn't matter you like I mean obviously it's at some level it matters. I mean, no, I mean, completions have gotten better and more targeted and well, it's and they've gotten better, but they've also gotten a lot simpler in a lot of aspects, too, right? Like, you're not designing, you know, based off crush or your closure stress and then your crush strength of your sand and then your chemical like everybody pretty much pumps slick water. Everybody's pumping Yeah. dust. Uh because dust is infinitely more permeable than no permeability, right? like it, right? [laughter] Yeah, exactly. It really when you put it into that perspective or whatever, but it's uh Yeah, it's it's crazy how much all of that has kind of just evolved in just the last 10, 15 years. Um Yeah. Okay. So, you you had your kind of red pill moment like so what where did it take you after that? Like was that within Kico and then ultimately to you went from Kaneko to Novi or Yeah, that's right. Yeah. So while I was within Kaneko, I I naturally glombmed on to any projects that could have like a a data bin to a data bin to them and I I um Kico was great. Gave some data science training, you know, shout out citizen data scientists program. Um and uh you know that was fantastic and uh but eventually I was kind of looking for something new to do, looking for a change. Did you work on any of the frack hit stuff? I didn't work on the frack stuff. No. Yeah. I mean what million projects going on there? Yeah. But yeah, um you know, I would do I would do stuff like with more of a geologic bin like where like landing zone it came and things like that. Um but yeah, so jumped over to Novi back in 2019, like five years ago now. Nice. So been there ever since. That was how how mature was Novi? Yeah. How mature was Novi at that point? Were they company was like four years old at that point? Yeah. [snorts] So we uh the founder came out of Hess. Um he did data analytics for Hess like I mean he was involved in IT like high level guy but um saw a real opportunity you know hey data driven methods cloud compute machine learning all this data being generated in shale probably had some of the same like um awakenings like I did [laughter] um but whenever started Novi realized this is something that operators should not be doing themselves. Yeah. Um there should be a software solution for this. Um and then in in 2019 we raised a series A and I came on board along with you know various others at that time. Nice. That's awesome. So break uh break Novi down for for people who don't know kind of what what's the the focus? What's the the benefits that sort of thing? Who are you [snorts] targeting? Yes. High level first and then we'll start getting into the tech and all that stuff. Fantastic. Yeah. At a high level we have really two offerings. One is machine learning software for production forecasting and development optimization. So how many wells should I drill in a section? What's my how much proper should I pump? Those types of questions. And that's mostly used by large EMPs like um you know the biggest of the big. And then there's um we also offer data kind of just like IS or invaris you know we go out to state agencies and collect it. Um but because we have all these relationships with operators like we can license proprietary data and we have all this technology that we use to you know do better job at production [clears throat] allocation or EDP forecasting and things like that. Cool. So what does that mean on the tech side? Now let's let's kind of start slowly unpacking that to to the extent that you can of course. Um so yeah just like workflow I'm a Novi Labs user. I want to use one of your let's say the ML tool for yeah production. What [snorts] does that kind of look like high level? I'm pulling in I assume a bunch of production data, right? Yeah, exactly. I mean we we'll have everything preloaded in there if it's from our public data sources which is part of the reason we have that that whole offering is because believe it or not it can take operators a long time to get their data together. Um no I know never. Um, but you can go in if you're if you're an operator at a a big perian company or wherever, you can load in your production data, load in your completions designs. If if you have your own geology you want to use, you can load that in. Um, and first step, we'll build a data set. That's always the most important place to start. You guys, you guys cover this, but it's a huge pain to get everything cleaned up and normalized and in the same, you know, the same format, right? Tables, etc. So we'll we'll do all that and then we have data that's ready for modeling. So we have it's all web- based tools but web- based tools for training a machine learning model. So you'll say all right well I'm going to go grab this [snorts] data set that I just built but let me just filter down to the northern Delaware basin as my focus area wherever it is. Um you know choose the variables you want to use. um few other little configuration items model will train all all in AWS and you'll get some results back on you know feature importance and error scores or things like that. you can say, "Okay, yeah, this looks good. Let me now go use it to make a decision about my lateral length or whatever business questions they have going on." And so that can they do like it doesn't necessarily have to be true multivariate, but this, you know, tweak the knobs, turn the knobs, change the thing, you know, that's that's the whole how much does it abstract, but then also may how much control does it give? Like so I'm sure you've got different uh types of users. You may have a you know a reservoir engineer who just you know kind of adjacent and understands enough and you your platform needs to handle more of that but maybe there's more data sciency type that may want to get in and tweak tweak the knobs a little more like it always happens in engineers I want to come in and be like all right let's let's really see what we can do here you know I'm I'm going to be playing with lateral links and per sand [snorts] yeah I mean we try to set up our models in a way that they can be I mean there's never going to have one size fits all, but they can be pretty well multi-purpose. Like if you made something that was purely just for optimizing completions, like that would be great. But to your point, that's going to have interactions with, let's say, your spacing and your stacking design. And so it's it can be sometimes hard to answer those questions in a vacuum. So we have it set up to be able to answer answer those. Yeah. No, that's awesome because that's I think that's historically been one of the big issue, one of the things that actually just almost forced you we structurally companies tend to have kind of their their walls and their barriers between departments or groups or whatever. But then it gets even worse when you have these siloed data products that it's just like okay this is just for this one thing then this is just for this other thing but those things are directly related to each other and they you know they directly cause and affect each other. There's no linearity, right? Yeah. But [snorts] then like it depends on some of the models and I was gonna ask like I mean do they get to pick some of them like the actual models that are being used in a hood like can I say I want to do a random force in an XG boost and stuff like that because like XG boost handles linearity you know kind of but other ones you have to minmax normalize and all kind of stuff like yeah so we've selected algorithms that are like pretty robust to these types of problems and like we generally want the users to um not have too many ways to mess mess things up. Um now we I mean we do have some more advanced algorithms that like colinearity like would give you a bigger problem for it like let's say versus you know random random forest that would be easier. We do use a lot of random forest. Um, but like we try to we try to as much as we can like have chosen maybe not like our our algorithms are more accurate than reservoir engineers at forecasting their wells and we've done you know dozens and dozens of competitions where we've beat them. But um maybe there there could maybe be a little extra benefit on the accuracy you could get at the cost for either risk of the algorithms like messing up if certain things are done incorrectly or loss of explanability and how they work. Yeah, that's that's just that's one of the trickiest things I feel like that a lot of pe not a lot but people just need to consider anytime they're trying to introduce or talking about AI or M any kind of potential blackbox product in the energy space specifically the oil field because as engineers we all want to know how the thing works right and so if you can't generally explain how it works which is [clears throat] not always easy even just with a normal like it's not because it's my proprietary model. It's just that's how they work. It's not the that's why computers do it because we as humans can't comprehend and understand the data just as it is. So that's why we have to have a computer to do it. But then the user wants to know well how did you do you like there's that dichotomy there all the time, right? And so it's always this like fine balance of like okay well here's what it is and then here's some variables and constraints that you can play with. But yeah, also from just software design, you know, it's like good software is easy to use. It's hard to mess up. It's easy to use from a user perspective. And so there's always that balance of like Yeah. And that that's been our goal is to have something which is like easy to use and reliable. I mean, where, you know, there are some places where you can, you know, pop open the advanced section and like edit a few additional things, but those are kind of with with additional caution. Yeah. Well, and that's the thing, too, right? It's like, yeah, you could get potentially a few percentage points better on whatever stat that you're trying to get, but then yeah, the repeatability drops. And so there's that or you're overfitting at that point, right? That's also another very big [laughter] point, too. Um, so let's get into the tech stack. And I don't know how deep you get into it personally at Novi, but I'm sure you at least have I'll do what I can. Yeah. Yeah. Um so is that like AWS um using a top I mean so are you all probably using like dockerized type things underneath the hood or is it you using Sage Maker or I mean uh not so much on Sage Maker like too expensive. Yeah. Yeah. Dockerized things lots of step functions lambdas depend on what we're doing. Um and your stuff's all cloud-based right? All cloud based. Yeah. Yeah. I mean, are do you feel equipped to talk about like say step functions or lamb like are explain them to people or I'm not going to No, that's that's beyond my No, I mean I mean because like step functions are literally like you can just define like literally like step because and that that came up too. I think I was talking to Derek Turk one time and he's like why do they use step functions because it's a totally different thing in engineering like thing than what they're talking about in AWS. Yeah, exactly. Yeah. It's like just like a nice way of scheduling out different Yeah. You just say once this is done then do this then that you know. It's just a logical work. Um and then yeah for people that may not be aware lambdas are basically like you can write functions like serverless functions that run any you can run in Python y probably basically almost any language at this point for the most you know within AWS like they're pretty sweet you know JavaScript C you know Java um all of it but um but definitely can be affordable way I mean once you get a certain scale you start paying and like um but I think you get like a free million lambdas a month I think or something like that. Yeah, I mean AWS is is great, but yeah, they definitely get you with like the low initial cost and Yeah. And then as you scale Yeah. All great software companies I found have that. It's not a It's not quite premium, but it's pretty close, right? It's like good model to have. Yeah. Well, I mean, it's ultimately it's the drug dealer model. I'll get you hooked on this little thing and then you'll come back. You know, it's the Philip Morris model. We get you addicted to it and then you just keep coming back, right? Uh, no that's so I've got a kind of a interest well on that point maybe not as far deep in the tech side but from a from a client perspective what are some of the you know you don't have to mention specific names or anything like that but just from like a case study benefits perspective what are some of the really some of the more interesting things that your clients have kind of some wins or whatever yeah come up come out that have come out of Novi with your product yeah uh yeah we have I'm sure there's we have a bunch them to share. But um a few of them that we few of them that we've uh gone over before we had um this is like going to be a little bit of a different take here for a second. But yeah um in addition to just like development optimization that' be one win and we come back to that in a second. We also have customers who use us for like asset valuation A&D type workflows relevant right now with very all the acquisitions happening. Yeah. So we had a customer Primex resources who sold to Ken and they used our software to like let me lay out some sticks on on our acreage and forecast those and them to you know what Ken had and okay it's going to be a creative to your quality to like make this deal. Um that's just like and then they sold for hundreds of millions of dollars. Yeah. And they could do that because you've got all the public data there exactly the box. It's it's the advantage for them is it's a consistent model. It's accurate. It's data driven. you know, it's almost kind of like a fair opinion to have on this. Um, and I mean, there are plenty of other uses for like buying acreage and things like that that we've had customers use um use it for. Um, and then lots of other customers just for like we had a company up in Canada who, you know, okay, well, we have stage design, propent loading, fluid loading, you know, lots of different things that could vary on their completions design. and you know, oh well, if we keep this one constant and move that one down, you know, boom, you save $600,000 on a well. Yeah. You know, and no change to your production. Great. So, that was that was another nice win, you know, times times uh however many dozen wells that you're drilling up there. It really adds up. No, that's uh I've been preaching that ever since I got out of the field and into the office and had like a perspective looking because before I got in the oil field, I was in manufacturing and it was such I've said it many times, but it it's just such a dichotomy there where in manufacturing you are monitoring every cent every second literally of every process and everything's just in time, right? Everything Yeah. And everything is optimized. not only the actual manufacturing of whatever you're making, but also the inventory for the things that you're making on a daily basis, right? You're not overordering inventory. So, you have a 100 pumps just sitting there waiting to be manufactured even though they can't be manufactured for six more months. That's literally a waste of your capital. Yep. Because it's not doing anything. And so then going [laughter] sitting on a frack job and like all the [ __ ] that happens on frack sites and all the NPT and all the random stuff and then it's like because it is it's just a giant logistics operation at the end of the day. It is the problem the hard part about the oil field operation side besides the fact that you know you're typically in the middle of nowhere and you have no connectivity and limited resources um is the fact that it's a literal mobile factory. Like it it's essentially what it is, right? If you had a plant that you were just trucking sand and chemicals and water in and out of, it's essentially a frack site, right? Like except now it moves every two weeks, right? Every two weeks, right? Yeah. So you have you don't have internet all the time. You have so it makes all these logistics, you know, much harder. Yeah. Then you have weather. There's all these other things, right? And so it's a uh there's just so much room for optimization there though. Like and that's the thing, the scale of it, right? If 50,000 10% whatever on a single well does may not or 1% on a single well may not sound like a lot but 1% on $5 million AF on a thousand wells is a lot of money. You know [snorts] it's uh but yeah I feel like the the low prices have really finally forced people to really optimize for that cost side. Yeah it's it both cost side and well performance. I mean, it just forces you to optimize whatever you can to stay in business. Yeah. Well, and then, of course, everyone finally debunked the uh the Aubrey [laughter] methodology of just ripping the chokes for IPs and not worrying about what those decline curves look like in the [laughter] the production. But it uh yeah, that's uh do you like So, as far as that goes, I mean, was it something as simple as like was it something kind of stupid simple that ended up saving them all that per Oh, yeah. something like make our stage lengths like 20 meters longer or something something on that order. Yeah, because a lot I mean for almost any of completions variables there will be a point of diminishing returns where your capital efficiency changes but there may also be a point of like not diminishing but like no more returns at all. You can't squeeze any more out of that. Um and like being past that point is is uh you don't want [laughter] Yeah. And that of course varies depending on the geology and all the other the other things that are going on which is the like the complexity of of the challenge. Yeah. How does that work together as far as like you know I'm an operator I'm developing a field so I've got my existing production now I'm trying to plan my infill and whatever area right I think one of the things the industry finally is kind of all in consensus about is that you know a field is a system. Mhm. It's not a well and then another well and then another well. They all interact. They're all pulling from essentially the same rock. They're all the same rock, right? And so, you know, I vividly remember, y'all both worked at Kico. Shout out to Dave Kramer. Um, I vividly remember watching going to like a SP lunch and learn with Dave Kramer where he was talking about some Barnett use cases where, you know, just infill barnett like there was no reason they shouldn't be able to frack these wells and they could not get into them. like they were pressuring out every single time. And uh well, you know, they it took I don't know how long, but anyway, net of it was well, they had an offset that had that had been on production for, you know, two years. On the other side of the pad, they had a well that had just been fracked [snorts] um and not bled off or like just fracked, not even producing. So, super high pressure zone over here, low pressure sink over here, and you're fracking in the middle of them. So, guess where your pressure is going to go? It's going to go to the low pressure. Fluid flows from high to low. Like that's basic physics. We learn that. Even geologists know that right decades [laughter] ago. But you know, it took this full case study for like them to be like, "Oh yeah, if we" and the result was they shut in the well that had been on production and let it build up. They turned on the well that they had fracked and drew that pressure down a little bit and then guess what? They had no problems. And it's like, oh, and then you get all the frackit stuff that we've been doing for the last 10 years now, too. And it's like, hey guys, how much value have we just completely destroyed in these fields? Well, I think that's one of the advantages of the new plays that are coming online is they can take it they can take all those learnings and apply them. Um, yeah. I mean, I think you hear a lot about like the Vakcoma. I would love to have a empty Vakamua block right now and just be able to take all the learnings from North America. Yeah. Yeah. I mean, that's what um, you know, you can go look at some companies down there like, well, hold on. What if we do like 2,800 lb per foot and drill like a three mile lateral? It's like lo and behold like you get great results, you know. Um that's uh better than the perian actually. But it's it's to to be able to go in there. Plus down there they also like typically it's not you know you're not drilling sections. It's it's like lease blocks like traditional governmental licensing. So you have more freedom for like program level development than you do if you're you know you got a HD. Oh yeah, there's a whole l I mean we're in the Bakan right now and it's like you know because again the direction of stress doesn't matter and like literally drilling calling them hashtag wells where like you had wells that were north and then drilling across them like in the same zone and just like not fracking. Yeah. But like but then there's but there's a ton of value that you're unlocking. Yeah. You know that because you're able to do that there. Fascinating. I'm I'm very convinced. Then there's all this whole land like Tetris thing of like and it's it's worse than Tetris because they can overlap and then if they overlap and like it gets so crazy in the legal land side of it like where you could actually maybe steal acorage be much easier if you just had one [clears throat] government big thing [laughter] you're like yeah let's just stack them in you know line them up in here and yeah I've never envied uh whenever I remember the first time I found out that you know certain states you can split by depth too. Oh yeah. I will never be a land man because that sounds so mis so complex the land side and even like the accounting and the operations and you know like all the interests and stuff. It just that's no one no one really truly I'm getting better at it. Like when I hear people talking fluently about it I'm just like oh yeah amazed like it's cuz just it doesn't click for me cuz I'm a visual person anyways. Like if I see it then I can put my hands on I get it. But like people are just talking about it out loud. Like there's a lot. It's crazy complex and a lot of um kind of risk or you know like if you don't understand it, you know. Yeah. [laughter] Like Yeah. Who doesn't love litigation risk? Yeah. Yeah. Yeah. No kidding. Uh I'm curious since we were talking about Vakam Morta. Cuz for those who don't know, it's in Argentina, right? Yeah. Argentina. And what's the other one? It bleeds over somewhere else too, maybe. It's basically all in Argentina. It's It's kind of in western Argentina, like northern fringe of Patagonia. Um but it's like how many meters thick? It's huge. Well, yeah. I mean, it's like the Eagleford but three times thicker. So, [laughter] I think the the right analogy for it, um you know, it has some characteristics in common with the Eagleford and some in common with the Perian. But, I mean, extreme pressure, the fracks seem to it just shatters. Um yeah, so I mean the well the well results there modern well results it's like 30% better than the perian for oil production and that's just from one one well do you well that would be even for like full like pad drilling full develop say like you mean come in and drill a version well I think we saw that yeah I mean I think over time just like in any play you know they'll start to figure out like which zones are on the you know are actually economic which ones are on the fringes like there's definitely stack pay and multiple targets. Um, but like the Eagleford, there's kind of like one pockets super like organic rich zone that's like very well-developed. I mean, it'll change all over the basin depending on where you are, but for the most part, there's like a one banger zone and then a few others that look good, right? Well, how many meters or feet thick is it? It's like it's something ridiculous. Yeah, I think it varies depending on where you are in the basin, but I think it's what I don't know like 200 to 1,000 feet thick. [laughter] Yeah, it's thousand feet thick on because you get certain things to your point you were saying earlier where like it's tight but it's there's a good there's oiler but like I mean you're have to land your well boore again three miles out through this very narrow or very thin Oh yeah, we're talking about 20 feet. Yeah, they they they try to get I think pretty precise in their targeting there um just to kind of nail down the zones that they're hitting and their stagger pattern or whatever else. Well, like stuff like that makes me, you know, the uh like the horseshoe wells that they're playing around with, right? Like that actually to me makes a [ __ ] ton of sense in these giant, especially on the exploration side where you're, you know, you're going to spend more money anyway, right? But like, hey, let's zigzag basically through this stacked pay and see if we can figure out from a single Well, of course, as long as the cost interesting idea for a for a uh for a test. I have a lot of questions about it from an operational perspective just being from the field. That's uh like I can't I mean we we always had you know issues with you oh wire lines running in too fast or you're pumping them down too fast all that stuff which I know that a lot of that has changed but even then it's like okay you now have to do that through two a dog leg down and then another around the corner. You could probably just draw you could probably just drill something through multiple zones and if you could fiber fiber the whole thing and do the production allocation like you could you could test them without having to do a horseshoe. That makes a lot of sense. I guess in those places where you don't, you know, you got you got a gigantic block to deal with, you know, no need to pack it in. Yeah. No, that's uh but yeah, like going back to the the peak oil thing, right? It's like there's all these things that we already know about that are just not economic. Vakam is not one of those. But it's, you know, it's like how but it's not tapped into yet, right? That's the thing. It's not even close to being like exploited to the point that Eagleford is or global demand potential. Yeah. Well, you mean like what I mean? Venezuela has horses and anyone, right? Like if they ever figured out the next 50, you got a whole another If you if you transplanted the Oronokco oil belt up into Alberta, it would be like a world changing prolific field. Not to mention down there, there's also I mean it's not just that. It's like the greatest conventional reservoirs in the Western Hemisphere are also in Venezuela. Like 50 billion barrel fields. there's uh unconventional prospect prospectivity there. Like it's probably the greatest source rock in the Western Hemisphere. Yeah. Um so it just it just piles on like over and over again just the richness of of the of the play opportunities there. Yeah. What uh you got anything else or uh was I think maybe on the tech side where where is Novi going? I mean like obviously you've got you know it's like we have to talk about it and I like talking about it but like LLM's AI I mean all that kind of stuff but I mean what on the tech side is kind of exciting you guys and where you trying to steer the product? Yeah I mean we've been doing a couple things as our focus areas recently for I mean obviously we're always working on a bunch of things like we're doing a big research project of feature engineering on spacing parent child with great results but check back in six months on that one. Um, but we're we're trying to build out proprietary data sets. Um, so like in Texas there's production allocation, you know, so the um the state publishes production data at the least level and in order to figure out what the wells are producing, you have to estimate their production based on well tests. We're like, well, we're pretty good at that. Probably the best, but like what if we didn't have to do that at all? So we've been going out and getting proprietary production data from minerals companies or operators being like hey will you let us just publish your well-level data right like sure. So it's been surprisingly helpful and that's like as far as the tech side the data that you use to train your model or make your forecast that's like more important than the algorithm. Yeah. And so junk in, junk out, right? Like that's what they taught us in math, [clears throat] every math class I've ever done. So we've really been focused on like, well, what can we do to make the data the best possible like throwing away assumptions that public data has to be bad? Like what if [snorts] we could go out and build amazing public data sets. So that's one interesting angle there. and then doing what we can to leverage all the technology we've already built for our, you know, EMPs going out there to optimize a development, make a, you know, $500 million capital decision, you know, on what if you could use that to improve public data for, you know, forecasting PDP wells or like undrilled inventories, you know, that are just as sensitive as, you know, making those big decisions just, you know, ready to go for people as like a data product. Sure. So that's like kind of turning our ML platform into kind of advanced data products. It's like a big effort initiative for us right now. Sweet. Yeah. Where do you Chuck want to meet? We were talking about this earlier, but uh he's like you just need to ask this question to anybody you're talking to about AI and so you're going to be the guinea pig. But where you know how do you see if you're thinking of this from the perspective of you know on the manager or seuite on the tech side like where do you see how do you see uh because Chuck got this question so it it's relevant but how do you see or where do you see kind of the AI the machine learning stuff um impacting you know an operator's daily or weekly monthly you know lives. Yeah. Well, [snorts] I mean, I'll say, you know, there's the whole generative AI revolution, chat, GPT, and all that, which I think there's like plenty of things to do, but in our industry, we still haven't even remotely come close to implementing, you know, these other machine learning or other datadriven methods that can impact the other business processes. Like to your factory factory analogies for all those optimizations, like simple stuff that's like how much are my wells going to produce next quarter is like under underdeveloped with like the potential that it could be like with datadriven techniques, a IML um and there's hundreds of hundreds of versions of that across the whole stack. But just things like production forecasting, risk, budget and planning, um you know, asset valuation, like all these processes that are like still done by engineers like by hand, right? Or siloed or in Excel or siloed in different Yeah. siloed in different um structures throughout the through through organization. There's still like the job is not done there. Yeah. Nohere nowhere close. And at the end of the day, that means like faster time to value, less less time wasted, lower risk, lower risk, you know, better well performance, lower cost, like all the stuff that you want to do in your business. It's like all the business goals. Yeah. You know, everybody everybody wants to achieve. So the answer is lower risk. So lower risk. Yeah. Start spending money on it. [laughter] It's not it's not a risk from a monetary perspective. It's a risk of if you don't do it, you're way it's way lower risk than you know um doing it by hand doing it by hand and like relying on some you know there's a lot of great I whatever whatever function you're looking at let's say reservoir engineers. A lot of great reservoir engineers out there but you know it's there's there's risk associated with having them do everything by hand you know. Yeah. or even just like you said in a siloed way, right? Like it's like, oh well, they're doing they can be the best reservoir engineer and doing great things in that one software, but it still doesn't tie into how that affects things upstream and downstream of that consistency. I I think you talked about it earlier and it was actually a theme we talked a little bit last night too is just like you know again production forecasting in a consistent way. I mean if you have four different you know production or reservoir engineers forecasting at a time like then they've got their different areas some of them just by nature are more conservative or more aggressive and like so then you're going to get this variability where it's like if you can codify you know and do it in a consistent manner. Yeah that's so huge. I mean the number of times like like every now and then we'll like run like maybe somebody has like a type curve area and like as like they'll put that into our models and you know it's like we'll see that oh somebody is tends to be more optimistic or somebody tends to be more pessimistic or like hey why why is everybody doing so poorly over here and it's like often because there's some engineer there who has a different choice about what to do one tiny thing. Yeah. But yeah, that even comes back then too maybe to some of this where like maybe the AI isn't just doing it for you though or maybe it is like you have where it's help helping them saying hey you know a lot of your stuff's coming out a little aggressive you know like I mean like almost like a co-il the whole idea of co-pilot kind of thing where it's like just you know making suggestions augmenting the workflow it's not doing the I I'm a big proponent that that's how most AI ends up getting kind of rolled into the market right initially is it's not going to be it's going to do everything for you. It's it's going to help you. It's going to help guide you. And then while you're doing that, it's going to be learning from how you respond to it. And that's going to be this nice federated learning loop that just makes it better. Yeah. I mean, we'll have customers who every AF they'll have their reservoir engineers forecast and they'll have Novi's forecast. And if there's a large gap between them, they might say like, "Okay, well, hold on. Let's not right let's not send this out yet, you know, let's let's try and figure out what what's causing the discrepancy." But if they're on top of each other, you know, all right. Yeah, send it off. Sweet. No, I mean I think that's I' I've had conversations internally about using language models for under the same guise of like, okay, here's a transcript, right? Here's the different categories or tags that I want you to use. Which one of how would you classify this transcript? Right? And it's like, hey, if I use the same ML model and the same prompt every time, theoretically at least, yeah, as long as I should get consistent results, I should get consistent results. And that's way better than me saying, okay, uh, Sydney, you're going to take these 25, Bobby, you're going to take these 25 files and just distributing it out because everybody's going to have a different way of doing it. And people get tired. I mean, like your results yourself like I don't by the 20th like I mean, no. Well, that's what computers are made for. some monotonous, boring, redundant, repetitive things that humans don't like doing and are not good at to begin with. We once did a comparison between our spacing calculations and some stuff that were done by hand by um a GIS person at one of uh our customers. And you know for the first 40 um we were the same and then the next 20 you see like we started to be a little bit better and then like you know by the end the the GIS person was like totally messing things up. Yeah. It was like 1:00 in the morning and they were trying to get Yeah. Exactly. Yeah. It's just you know getting their signs wrong. Anyways all sorts of but just it's just human nature. I mean human nature. I'm I'd be the same thing. Yeah. I probably wouldn't have lasted through 40. [laughter] No. That's very true. You wanna Man, it's already an hour. That's Yeah, it's crazy how quick stuff goes. It does when you're having fun. When you're having fun. For sure, man. That's what we like to do here. So, uh, jump into the speed round. Speed round. Speed round. We're just going to pepper you with some questions. Who's your favorite uh kind of social, Twitter, LinkedIn, follow it. It doesn't have to be one specific platform, but yes. Uh, I got two favorites right now. And I feel like always when you're looking for this, they got to be funny and insightful. That's your That's your combo. That's the sweet spot right there. Two favorites on Twitter. WTI Realist and uh Joe Merchant. He's like Oh, yeah. The Padre something. Yeah. Yeah. Padre 13, I think. Or Yeah. Yeah. Shout out Shout out to them. Yeah, I tweeted him this morning. [laughter] Tweeted over to Joe. I laugh I laugh so much. Laugh and learn. That is uh I love that about our community. I love that about the I'm glad that X is around and still exists and allows people to maintain anonymity and all of that stuff because we wouldn't get such great not just interesting insights but also actually you know fun and uh entertaining as well. So shout out to those guys. What's what's a piece of tech that you use daily to like help be a better product manager? I mean that helps that you really enjoy using I have that question for you as a product owner myself. I'm very curious on some fun tools that you might use. Well, this is not going to be a fun tools answer, Jira. Um, yeah. Yeah. All right, moving on. [laughter]
Actually, their product discovery tool is pretty slick. It's a thing they rolled out for product management specifically. It's been been pretty nice for like like high level roadmap planning. So, it's like a kind of a template or workflow that you kind of work through. It has a lot of similarity to Jira in that like for those who don't know Jira is like a ticketing and a planning system that's used extensively like in the tech industry for like what are we going to work on this week or like this quarter things like that and all the detailed you know this then that etc. But this lets you kind of you know lay out timelines and move things around boards like kind of like almost a little bit of notion style flexibility but like within the Jira ecosystem. Nice. Yeah, that's good. No, I mean Jurro is a standard, right? As far as the can't live with it, can't live without it. Yeah. [laughter] It's like anytime I'm I'm working with a contractor, consulting dev shop, if they say Jira, it's just like, oh my heart, I don't want to have to open Jira. And as long as you said it, like that's that's my only gripe with Jira is it's it's like any other big enterprise software. It's like it's only as good as it is structured. Yeah. initially, right? But it's so flexible that it's very easy to be like, "What the hell am I doing?" [laughter] Yeah, exactly. Um, that was a good question. What What are some of your favorite product kind of books or blogs or any kind of written content? Yeah, let's see. I'll I'll shout out like a couple um books here uh for folks to to learn from. Like I mean I've got I've got few things on the other ones but favorite um you've got like the I mean this came up a little bit earlier. You're talking about all your lean manufacturing lean startup Eric Ree like classic one. Every product manager should read that. Yep. Um innovator's dilemma. Yeah. Clayton Christensen. Um uh another another kind of classic one right there. And I'm blanking on the third one I wanted to recommend. Crossing the chasm. Yeah, dude. That's like your That's like your triangle, right? Yeah. I could see it in my I could see the holy trinity of see the holy trinity right there. Yeah. Which uh you know, hey, maybe that's not like innovative things here, but like we got a actually a lot of folks in in oil and gas or geologists like don't know about like product management like that. It exists. Yeah. I didn't read these until I kind of started. Same. I was like, whoa. Yeah. Great books. No. Yeah. I was in engineering and then sales and technical sales and then I got into product and it's one of those things like until I read that book. you're just it's almost like you're just floating and you don't you don't know because there's just like it's all new and you've never done it before and like yeah you can have training and stuff on it but that doesn't necessarily translate always in the real world and so after reading crossing the chasm for the first time and I was reading that as we were working on rolling out new products I was like oh my god like this is what I've been looking for like clearly defines each part each segment of your audience throughout an entire product's life cycle what their personalities are what their interests are how to connect with like Yeah. Everything. I read that. I was like, "Speak to me." [laughter] Well, because it's just like it's scary. It's unknown. You're like, you just don't know what Oh, yeah. to do like how there's so many variables and things that you're just like you don't know. Yeah. I love reading in those books, too. Like innovators dilemma. He'll talk about like stuff from Yeah. a long time ago, you know, of like certain excavator technology, you know, hydraulics versus, you know, pulley system. And it's like, yeah, it it actually is the same like repeatable motifs no matter what technological like generation you're in. And it's it's the applying those concepts is is so powerful. Yeah. Even crossing the chasm is like what late 90s or Oh, yeah. I mean, I don't know. They're talking about I don't know early internet does psychology change. I mean, like we're still dealing with humans, right? The same exact. Yeah. Great point. Yeah. Well, and it's it's just such a good, you know, because they take it from the crossing the chasm specifically is talking about technology, right? But it's like most products are driven by some kind of technology. So, it's literally every market at some point they had to get someone to use the wheel, right? Right. I mean, I assure you this is better. Oh, yeah. [laughter] Yeah. Yeah. You don't want to drag your Yeah, there's a dragard instead of a lagard. Right. Great. Um, Bobby, they didn't they didn't roll the pyramids into place anyway. acoustics conspiracy. I need my tinfoil. Let's get Jeff using here. Yeah. No, Jake actually made a request to to make some tinfoil hats for the the Bitcoin podcast. Oh, yeah. Jeff got into the moonlander last [laughter] night and yeah, I think it's hard because I brought up, you know, just that most people from Oklahoma I know are kind of tinfoil hat and really like they're [laughter] like kind of the things that they can control and and Jeff confirmed that. Yeah. He's like he's like, "You're absolutely right." [snorts] Um [laughter] but uh h I'll go the last one just where's your favorite favorite place to go on vacation. Oh man. Um tough one. I'm probably more of like a mountains guy. Okay. You know I always dream about like going back geologist too. You get to like see rocks and stuff. Um so I'll I'll shout out let's say Wind River Range in Wyoming. Okay. Yeah. Amazing. It's like a mini Sierra Nevada. Same geology topography kind of glaciated granite, you know. Little geology shout out right there. Um, but it's nowhere nowhere near as crowded as a Sierra and it's like over a lifetime you could hike the whole range if you wanted to. Be tough to do that in Sierra. Quick followup to that and then we'll wrap up. What are some of your favorite places to check out like outcrops? Just geology question. Oh, geology question. I mean tough here in Houston. Well, not much here in Houston. [laughter] It's it's at sea level and flat. Yeah, there's some good spots in Austin, like great road cuts you can look at. Um, that's it's almost dangerous. Like I drive It's almost dangerous. I'm actually literally [laughter] going on a field trip tomorrow where they were like, "Wear wear your orange safety vests and like don't step off and die the," you know, [snorts] but Well, even just driving though, right? Like I my in-laws uh he's got a ranch south of Ozona and just driving once you get west of San Antonio going to Ozona, you know, you've got the road literally cut through the rock and you've got you can see all the nice planer Oh, yeah. It's really sweet. But it's like, oh, I'm going I'm also going 85 miles hour. I need [laughter] to like focus, but a little less dangerously. You can just go in into the Austin green belts and you have a lot of the same exposure. Um see the same rocks. Pretty sweet. Um my wife always has to deal with me even though [snorts] we hike on the same trails like almost every weekend. Yeah. Like I'll be like, "Oh, look at these mud cracks, [laughter] you know, some great burrows. Look at the church nodules." Oh, yeah. My wife makes fun of me for bringing rocks and stuff back from different vacations and stuff. And I'll show you my rock collection on my desk when we're done. But yeah, excellent. Where can uh Appreciate you coming on. Where can people where can people find you? How do they get in touch if they want to learn more? Get in touch. Novillabs.com. n ov labs.com. And you're on LinkedIn and Twitter? I'm on LinkedIn and Twitter. Yeah, Twitter, Tedcross. LinkedIn. Just search for Ted Cross. Um, you know, you you probably find me there. Yeah. Perfect, man. Appreciate you coming on the show with us. Yeah, pleasure. Thanks, guys. Sorry.