00;00;01;21 - 00;00;23;13
Unknown
Wake up. Welcome. Welcome to a new episode of Energy Bytes. I'm Bobby Neil, and I got the rad dad John coffee. And how's it going? Doing great. How's it going? And, super excited and, you know, trying to get these these two on for a while. But we got some run Vinny from, well worth, might be our first husband wife duo.
00;00;23;19 - 00;00;51;12
Unknown
Yeah. Is it good? Yeah. Well, have you noticed those sisters? Have you not had Colin and Julian time? No. No, we avoid avoid having it on our fucking heads. But, But. Yeah. No, I mean, thanks for thanks for being here. So just, I don't know. Real quick, if you want to give, you know, kind of a quick intro on yourselves and then just, you know, what well worth is and.
00;00;51;14 - 00;01;18;10
Unknown
Sure. Yeah, we have we we use it at XML. We were early early adopters I think so, yes. Yeah. One of our favorite customers. So I have a background in upstream oil and gas equity research and investment banking. And so I was building a lot of Excel based models. And then it has a tech background. And so I was complaining to him once, saying, like, engineers have all these solutions for their workflow and finance teams are still building Excel based models.
00;01;18;13 - 00;01;38;16
Unknown
Why does it. So my baby. Yes. Yeah. That was the limit of our coding. So I was complaining to him. Why doesn't someone build something for finance? And he said, well, why don't we do it? Yeah. So that's where the idea for well was, came from. And a brief intro and well worth. So well worth is, a SaaS platform for upstream oil and gas finance teams.
00;01;38;20 - 00;02;04;15
Unknown
Yes. It basically helps, finance decision makers, make better decisions around raising, managing and deploying capital. And our customers use well worth alongside reserves, tools like comma curve areas and PhD in. And it's used for both a and and A workflows. Great. My background is in enterprise technology. I've been in that space for about 7 or 8 years before.
00;02;04;22 - 00;02;21;04
Unknown
But apparently I got started because I was looking to invest and so like I said, she was complaining about this, and I said, okay, the easiest way for someone to solve a problem for them, right? Yeah, that's how I got into it. Nice. And then I went to rice from MBA. They have a really good startup program, entrepreneurship program.
00;02;21;06 - 00;02;38;22
Unknown
So we were part of the startup accelerator called Augsburg. Okay. That's where we did our customer research to figure out, you know, is this problem solving or is it just us imagining that we can make money from this or whatever? And the research was great of that. So I'm like with that investment banking job. And then right after I graduated, I got into this full time.
00;02;38;28 - 00;03;04;25
Unknown
Okay. It's been a good journey since then. Nice. That's awesome. Yeah, I think that's probably one of the biggest thing is people go into it and they like and they think that their problem is, you know, everyone else's problem. But it's, you know, like you said, you have to validate that. I think that that's one thing that was great about the Spark program and that they force you to not start building until you've spoken to a lot of customers to understand, you know, is there really value in this?
00;03;04;26 - 00;03;23;15
Unknown
Yeah. Or more importantly, is there value for potential customers? Sure. And then just for you. So that was a good, habit building exercise. It's an important step if you don't want to waste a bunch of time and money and fail fast. Yeah. Yeah. For sure. Now, because you can build something truly amazing and it solves your problem really well.
00;03;23;22 - 00;03;47;15
Unknown
Yeah. No, I was like, I didn't ask for this. Yeah. Yeah. But yeah, I mean, I think, I mean, if there's anywhere I, I've been in plenty of oil and gas companies, like, literally everyone's big and small was doing their finance, you know, and all that work in Excel and like they don't run fast. No. And you just access really scalable though too, right?
00;03;47;16 - 00;04;07;11
Unknown
Yeah. I mean, cause when you. Yeah, you it's not a database, so you can't it's always going to be based on aggregated data whatever you build in Excel. And so if you want to change any one level assumptions you can't. It's not an Excel. It's in the reserves app. Yeah. So you have to go back every time to get that.
00;04;07;11 - 00;04;30;25
Unknown
If you want your model to be that dynamic or analysis to be dynamic, if otherwise it's going to be static. You can build as many data tables as you want in Excel, but since it doesn't really have access to the available data, those data tables are going to be a little bit static. Yeah. No, I mean, and I think I mean, was being at the center of a oil and gas company, the finance folks are going out to everyone.
00;04;30;25 - 00;04;49;26
Unknown
Hey, give me your assumptions. Like, you know, and they're and again, it's always give it to me in a way that I can then plop into Excel and again, like it's it's like that. It's as it's already out of date probably by the time it gets there lands there. And again, there are ways to connect Excel back to say, database systems.
00;04;49;26 - 00;05;06;05
Unknown
But again, most people aren't doing that for one and two. It's not the most efficient way of of doing it. But also then like, I mean, a lot of the things I've seen is then they're taking aggregated data that they because they, you know, have enough rows to bring in every well at a, you know, daily or even monthly basis.
00;05;06;06 - 00;05;25;17
Unknown
So they're bringing in, you know, rolled up to a certain, area. And it's like, but what if they wanted to go back and say, you know, there's just no way to drill back down into the into the data? It's almost like it wasn't designed for that. Yeah, I don't know. But maybe I just love yeah, how much our industry likes to square peg round hole itself.
00;05;25;17 - 00;05;46;14
Unknown
But these but I mean these finance people like that model is their baby. Yes. Like am I right. Like, I don't I just don't know. I mean, I've spent lots and lots of time on Excel workbooks and stuff, but the minute I discovered a database, I was like, oh yeah, this is way better. This makes way more sense to doing anything on this data from now and in the future.
00;05;46;16 - 00;06;17;12
Unknown
And so is there. I mean, it, I assume this is kind of going to sound like a stupid question, but I'm a completions engineer, so I'm happy to ask it. But, like I would assume historically speaking, that was just the way everyone did it, right? Yeah. No one was looking at like, hey, what if we just put this in a database and we move things away from an Excel file that lives on one person's desk, and then has to get passed around a million times?
00;06;17;14 - 00;06;36;29
Unknown
Yeah. What? We empathize because that's what I was doing as well. And the most that I could think of doing was write VBA code, but that was also then Excel dependent. Nothing to do with the database. So as somebody who is in finance, you're not really thinking about the technical side of hiring somebody to build software for you.
00;06;36;29 - 00;07;07;25
Unknown
And there has been nothing before. Well, words that could solve that problem. Yeah. And so it had to take somebody in investment banking and somebody in tech to come together and build something. And as an oil and gas company, you're not you're not thinking about having a tech team that will build something for you. And we have heard about a couple of people on the tech side trying to build something, because we've spoken to so many people by this point, and consensus is there is nothing like, well, what's out there right now that can do reserves as well as integrate them into corporate finance.
00;07;07;29 - 00;07;25;27
Unknown
Yeah. And we've heard that a couple of people have tried it and it hasn't been a true corporate solution like, well, what it is, it's interesting that you said discovered because so the way I look at it to 30 years ago, had you asked someone that you're not while you're walking down the street, you're not able to make a phone call?
00;07;26;00 - 00;07;41;18
Unknown
Is that a problem? Right. They would say no because they can't imagine. It can't be done right. So that's the way with most of our customers, right? They can't imagine that there is a better way of doing this. And unless that happens, unless they see something, they're still going to believe that Excel is the best way for right now.
00;07;41;18 - 00;08;01;11
Unknown
That's, I feel like it's a very common and it's not like an issue, but it's just it's this dynamic that happens, especially in a big company when you're within reserves are never like, you know, not stressful. Right. And the economics and it's so you have all this external stuff. But then at the end of the day it's like you just have to get it done right.
00;08;01;11 - 00;08;15;22
Unknown
So you're not going to spend any time being like, hey, how can I do this a different way? How can I don't have the time? I don't have the time for it. But so budget season comes around and right know this time of year and everyone's just hair on fire trying to figure out. Right. And so it's just gets it done.
00;08;15;24 - 00;08;34;04
Unknown
Exactly that doesn't mean the most efficient. It's hard to get people to look at something new. Yeah. Thinking of building something new is like a full time job for someone. Yeah. Even just using something new. Right. Yeah, exactly. And then from a sales perspective, it sounds it works both ways. As in, it's good for us that these people are busy.
00;08;34;04 - 00;08;50;27
Unknown
So the value efficiency more than right. But on the downside, because they're busy, it's difficult to get them to look at your stuff. Yeah. To even to maybe even do like a proof of concept or even just try it for one cycle. It's like I don't have time to do that or they don't have the resources. And I think they're I'm sure I'm pretty sure like what?
00;08;51;00 - 00;09;02;26
Unknown
I was like kind of a Phillips and I'm not sure exactly, but I, I'm pretty sure some of the bigger ones have probably tried to do things internally. Yeah. But again, then it becomes more of a corporate thing. And then I think a lot of people on the edge feel like they're forced to use it. Right? Yeah.
00;09;02;26 - 00;09;20;29
Unknown
That's true. But I think, I mean, obviously why almost everyone uses Excel is for the flexibility. Yeah. So I mean, like and also you being the original user, like you had to value that. Like so like when you guys were designing you're building the application. Like how did you take into account like the how can we design this for flexible ability?
00;09;20;29 - 00;09;41;06
Unknown
Because like, you know, you can you could guess what they want, but then there's someone's always gonna want something different, you know, the slightly different way. Yeah. So we built it in a modular way. So when we were building Excel based models and I was an investment banking, I had built a lot of models already, and I knew the pain points of, okay, I have this template, but that template always needs to be reworked for every deal.
00;09;41;07 - 00;09;59;15
Unknown
Yeah, even though I've made it super modular. But every time, maybe I have five columns for a cost right now. And then they asked me to add one more column. And so if I'm doing these, if I have ten different sheets for type curves in my Excel file, I have to manually go in and add an extra column in each one of those sheets and add formula there.
00;09;59;17 - 00;10;22;13
Unknown
So since I knew the basic structure of how you build an Excel based model, the way we've built well works is super modular with everything like you can add as many costs as you want and do well where they're all dynamic. You don't have to go in and, you know, manually add manually on the back of your cost, add formulas or refresh formulas, or be worried that something's going to break.
00;10;22;16 - 00;10;38;23
Unknown
Because now by this point, it's been tested. Like one of the first models that we actually built was the client curve analysis, which now we do we don't even show. We don't talk about it. Yeah. It's like back then we built it because a convenient topic to test it. So there is an entire UI built out which lets you put in apps, equations, blah, blah, blah, all of that.
00;10;38;28 - 00;10;56;24
Unknown
But we realize no one cares about it, right? Because right now there's so many people in, doing the same stuff, it doesn't make sense to just focus on it. So for us, it was easy that we made that decision early to not focus on the production side, production forecasting side of it, because Austin's like in economics and financials, because all your parties are gonna want to use areas common.
00;10;56;24 - 00;11;19;29
Unknown
You're already using it. And we just say bring your data in. We focused on making it. So we focused on making it easier to bring that data in. And then then from there after, the parties would love it if their finance people could go draw their own curves. So yeah, they would. So right now, I think the only thing that well, doesn't do is production forecasting.
00;11;20;03 - 00;11;41;27
Unknown
Yeah. And so you can import data from reserve tools really easily, into Alberta. And then if you want, you can do, cost modeling, pricing differentials and the reserve level economics in Alberta as well. Yeah. But if you don't want to do that, you don't need to. If you've done all the work in a reserve that you can essentially import everything within two minutes.
00;11;41;29 - 00;12;01;26
Unknown
Nice. So you already have your production costs, differentials and available economics. So with that, like say connecting to the back end, you know obviously people have like areas SQL databases or a common curves API. Or is it like hey, export these reports out of the we do that on sort of a demand basis because internally we use API's within the app itself.
00;12;01;26 - 00;12;21;11
Unknown
Right. Our customers have not really ask for direct connections as such. For us it would be a close to trivial task of connecting to the databases. Right now we have something called a data import wizard. So that way they can export this data out of their tools for forecasting in any format, some sort of tabular format, any column names.
00;12;21;14 - 00;12;43;15
Unknown
We have a data import wizard which is which lets them say, okay, my oil production is in a column called a okay. Gas production is map of them, not just create that mapping one time, and then every time they need to update or refresh the data, just use the same mapping, update the data. Nice. Yeah. It's like all of the tools like even comma curve can export data into a standard format, which is a one line file and a monthly file.
00;12;43;16 - 00;13;15;17
Unknown
Yeah. So we just need those two files. Nice. There's so many side tables in these applications. It makes for some sort of an overarching, connector right now. Yeah. Eventually we'll just do it. If some if we find enough patterns, we'll just build it up. Now that's I think that's a big thing that a lot of people don't pay enough attention to is like, for example, you have a bunch of incumbent softwares that people use in some capacity or other, as well as the myriad of Excel workbooks that are just floating around.
00;13;15;17 - 00;13;30;11
Unknown
Some of them have 2 or 3 that we just mentioned. They all they have them in-house at those two places, right. Like, yeah, right. And so but you know, a lot of people complain that it's like, oh, well, they, you know, they won't beta test the app or they won't test it. And we gave it to them for free.
00;13;30;11 - 00;13;58;26
Unknown
And it's like, well, yeah. Did you think about how they were able to get the data to make to actually use your app? Like did you get to that step of the of the process? And you guys have clearly done a very good job of one, you know, the problem. So that's a huge help to understanding. Like hey these people aren't just going to drop, you know, our potential clients aren't just going to drop what I've what they've been using for 20 years and just turn on the switch and use us tomorrow.
00;13;58;29 - 00;14;31;06
Unknown
Y'all made it as easy as you possibly could essentially to import, to let that data move across, which, historically speaking, in our industry has been incredibly hard. With incumbent software providers, they don't like their data being moved around and stuff. And so it's, a long winded way of saying, I'm happy to see other software companies doing that, because I truly think that's one of the reasons the industry is so far behind is because we have this built in data siloing from the vendors.
00;14;31;08 - 00;14;51;28
Unknown
And so, yeah, I mean, cheers to you guys on that. Plus, we didn't want to force people to do something that they did not want to do. So in our application, every table that you have, you have a way to edit that through the interface. But you can always download all the tables into Excel. Yeah. And edit them in Excel and upload because we know like this is something that I was doing a lot.
00;14;51;28 - 00;15;12;07
Unknown
When you have a lot of data, you don't want to go into a cell and change values. You want still want that to be exported into Excel, you know, and to some extent was almost like working against ourselves and making it easier for customers to pull data out. But we realized that it was also a good selling point in that if you are able to convince them that you are not going to be stuck in, well, what if you don't like it?
00;15;12;09 - 00;15;30;03
Unknown
It had them adopt it faster. Yeah, yeah. The end of day accounting or in finance folks are going to it's going to end up in Excel. That's that point right. Yeah. You need to put it into a slide deck. And you want instead of us trying to build a Spotfire power BI. We have dashboards, in other words, but you can't really go in and change colors and edit etc..
00;15;30;03 - 00;15;54;18
Unknown
So we have an Excel add in. You can export your data into Excel and just have dashboards in Excel which have which have charts that have your own formatting. So if you want to put that into a slide deck, it automatically refreshes. Yeah, but everybody can build their own color schemes. And instead of us trying to. Yeah. And I remember right from like when you all done it like you have some experts that even come out, they even have like the formulas in and yes, we do excel at just getting a raw like dump to a CSV.
00;15;54;18 - 00;16;11;20
Unknown
It's like it's actually, you know, I'm still embarrassed that I pushed back so hard when someone came up with the idea of building the added, I said, no, we're trying to replace Excel. We've built something for Excel, right? But then now it's one of our most, appreciated features that they are able to export a fully built out model with formula.
00;16;11;27 - 00;16;27;09
Unknown
So they are not moving as away from Excel as, as an they have a choice, right? There are customers who don't use the add in at all. Their customers, so use it primarily for the formatting side of it. Let's give some sort of an Auditability in that they can see how the numbers are derived, right? Rather than just seeing some sort of a black box.
00;16;27;09 - 00;16;47;29
Unknown
Absolutely. Yeah. I mean, again, meeting them where they're at. Right. Like people think that again. Oh well, that's a competitor because it's a data product or whatever it is. And it's like, hey, but give the client incentive to use your product, which means also allowing them to talk to each other like it's it's, a shockingly uncommon thing still.
00;16;47;29 - 00;17;08;06
Unknown
But it again, the value is there because the users are using it. Right. And so I, I applaud you guys for that because it's unusual and still. Yeah, it also doesn't really replace well right. That's a good thing as well. Right. Like it's not something where when you have it outside in Excel, it can't still be a dynamic model the way it is inside.
00;17;08;06 - 00;17;30;05
Unknown
Well us because we have so, so much complex logic like the circular logic that people, very few people, let's say in Excel actually end up building. Yeah. Where like if you as an example, if you have a whole capital structure built out and you have an RBL which is going to be a revolver, and then you want to make decisions around, okay, if I need cash, where do I draw it from.
00;17;30;05 - 00;17;51;08
Unknown
And if I have excess cash, where do I distribute it. Yeah. And then if you say I want to make these decisions while maintaining a minimum cash balance of $1 million on my balance sheet. So flatline cash balance and excess payout. And if I need it, then draw. And then if you want to draw it like you prioritize your sources, you.
00;17;51;08 - 00;18;18;17
Unknown
Well first of all I want to draw on the revolver. Then like a term loan then an equity. And then you also you said priority and payout as well. Okay. So let's say first of all I want to prioritize paying out the revolver. And then you add the additional complexity of covenants on top of it and say, okay, I only want to draw on the revolver if my debt to EBITDA is less than two times and I only want to pay out on the revolver if my debt to EBITDA is greater than one time.
00;18;18;19 - 00;18;38;07
Unknown
Yeah. So you put all of that into Excel and it's a circular it's circular logic. And so if you had if you had your research data in lot of words as well, and you put in this logic, it's going to break okay. Because it's such a big model, that it can't handle it, but in value it's all of that logic is now automated.
00;18;38;09 - 00;18;53;28
Unknown
And so it if you go back and you change your drilling schedule and you say, okay, oh, actually this model is going to be delayed by a month, you go back and you change it and, well, where's the rest of the model you built as a one time activity? And you put in all that logic that I just talked about as a one time activity.
00;18;53;28 - 00;19;31;28
Unknown
Well, words will automatically make all the adjustments and maintain that $1 million, cash on the balance sheet. And it will just give you the new returns analysis for the investor. It'll just give you the new dashboards. Yeah, that's pretty sweet. That's really awesome. Yeah. You can tweak the knobs on both sides. Yes. Exactly. Yeah. We're just we're just starting to get back into the other circular reference of, this has been in the pipeline for five years now, but when once we like, fine tune our, scheduler a little bit more, we're going to get into stuff like, okay, this is my reserves model, and then this is my current capital structure.
00;19;31;28 - 00;19;53;04
Unknown
And this is what the forward looking estimated borrowing base might be in the future. And so, actually, how about I go back and have this scheduler auto adjust my schedule so that it dwells within cash flow? Okay. So stuff like that can now be automated because we're at a point where we can we can have it do that automatically.
00;19;53;06 - 00;20;18;02
Unknown
You mean Excel can't do that. Exactly. I mean I would just push back. I mean, why can't someone just build this in, with cloud tomorrow, right? Yeah, right. Just go to court. It'll it'll figure it out. No, I mean, like, I mean, people will try. Yeah. No. So I mean, it's a but like I mean, like I've heard, you know, you hear a lot about SAS is dead and all these things, but like, people are so grossly underestimating what it takes to productionize even, like a fraction of what you're doing.
00;20;18;02 - 00;20;40;24
Unknown
Yes. Anything. Truly. I mean, we have a conflict of interest when we talk about stuff like that, but it's still worth talking about. People will try, building stuff like this, but we've been doing it a while, so we know the pitfalls of, you know, being an oil and gas company and having a core business and then trying to have a tech team to build something as complex as.
00;20;40;24 - 00;20;59;02
Unknown
I mean, even if you don't say, well, I'll say something like carmaker or are you going to have as an oil and gas company, are you going to have an in-house, reserves tool, accounting tool and a finance tool, and then someone to maintain, someone to maintain or whatever, are you going to build it and then on it and it's going to be a cost center.
00;20;59;04 - 00;21;17;04
Unknown
Not really. You don't even intend to commercialize it. Just building is easy. It's more of how do you make it production? Great. Yeah. How do you make it maintainable? Yeah. The first version is easy. Yeah. Oh yeah. Then you've got to implement tests, user management and privacy. And then you have to keep adding stuff to it. Right.
00;21;17;09 - 00;21;42;08
Unknown
The interesting one and like, and I've talked to well, it's a single company now we just went through a merger. One company was very big build versus buy and they built a lot. And yeah, I think everything was very tightly coupled in the back end, which is great when you're building for one company, like you're just seeing their database in this, you know, the well app talks to the reserves and like planning all these things, but now they're trying to integrate with the other company.
00;21;42;10 - 00;22;03;22
Unknown
Oh yeah. And they're trying to pick a best in breed. And people over here love their baby. But like but oh we want the well app but we don't want the other thing as well. If you take the one thing out it's like the Jenga block and everything else, everything falls, crumbles. So it's like. And again, I'm not I guess people don't build that in-house thinking, oh, we're going to get bought one day or we're going to merge.
00;22;03;24 - 00;22;19;27
Unknown
But then that also becomes something where like, you know, looking to the future. Yeah. It works great for that company. But then is this something that's going to continue on with them? They're just like an oil and gas, I guess ultimately upstream oil and gas companies are not usually software companies. Yeah. Or should I be? I should not be.
00;22;19;27 - 00;22;25;14
Unknown
That's a better statement. Yeah.
00;22;25;16 - 00;22;55;14
Unknown
So, I mean, I guess that getting in, actually, before I get to, under the AI side, can you just walk through, you mentioned like, some, like, M&A type workflows, but then also like, like, can you talk about the differences between those and but like that, well worth can you know there's built to handle that and then like what does it look like on a day to day year to year basis within only like gas company up with like scenarios and budgets and all that versus BD and just kind of that's a really, really good question.
00;22;55;16 - 00;23;22;01
Unknown
Explain for people who don't yes I well so people ask us this question all the time. Is it and or is it for financial planning and analysis like for corporate modeling. So, the two workflows would be if you have a company that is doesn't have an asset base right now and is looking at, acquiring assets right now, it can be used by such a company as well right now, because you can essentially import, as we mentioned, you can import valuable data from your Reserve's tools.
00;23;22;04 - 00;23;44;00
Unknown
And then you can, layer on corporate side like, Jenna Hedges RBL, that equity investment. So for a first someone is looking to do a and well worth can help with acquisition financing analysis okay. So right now with the reserves app all you can do is you can say okay, this is what this asset is worth. And so this is how much we need to acquire it.
00;23;44;00 - 00;23;59;18
Unknown
But you still need to get into Excel to do the analysis. Okay. How much money do we actually need. You probably don't need you know, you probably need more than you think because there will be financing costs associated with it. So it will well worth you can within a couple of minutes just quickly layer on that investment.
00;23;59;18 - 00;24;18;27
Unknown
Equity investment, build out a whole equity waterfall if you want to run, if you're on a private equity side. Okay. Within a few minutes and you can see the returns to the investor as well as returns to the project, nice. So it helps with acquisition financing that way. So this is a really like our float, seemingly off the wall question.
00;24;18;27 - 00;24;46;19
Unknown
But it comes from when I was getting my masters and I was doing these my finance class and doing these, big, big financial plans and stuff. How much, how much do people want to pull in the actual, you know, the equivalent of real time financial data for these analyzes, whether it be loans or rates or different things, right, like versus how many of them are just going to input the number.
00;24;46;19 - 00;25;02;29
Unknown
This was my favorite part of the finance class because it's like, oh, well, you just use this constant for the the cost of money. And it's like, well that's well, wait a minute. That's not how my world works, right? Like it's not, it's like doing physics in a vacuum, right? Like, oh yeah, this theoretically works, but the real world is in a vacuum.
00;25;02;29 - 00;25;27;11
Unknown
So yeah. Are there any again, super random question, but I just thought about that from building those weird bugs from. But I don't think that people are looking to, like, import real time data like you're talking about maybe Ford cars or something like that. They just they can easily input static data into. Well, but then if you wanted to change that, maybe on a weekly basis, that again, there's an easy bulk uploading value, right, for you to go in and change it.
00;25;27;13 - 00;25;44;16
Unknown
And we do have the data streams feature where you can create streams of these interesting or surprising or whatever. Right? So it doesn't have to be a fixed single number. Right? You can do that. And then you can build scenarios using these variants. And you can compare it so that that's one of the most recent features that we built.
00;25;44;16 - 00;26;04;08
Unknown
Scenario analysis. You can have like stacks of different inputs. Right. So each stack is a scenario. And run this 1015 scenarios in one go. And you are actually able to see the returns. You can see how is your URL changing, how is your production code trends looking both on the asset side. On the finance side. So yeah, you can use streams of data or you can just use single numbers.
00;26;04;08 - 00;26;25;26
Unknown
Yeah that's cool. Yeah. And most of the time people are building based on existing term sheets. Right. And so they know what terms they need to enter into the model anyway. That's fair. Yeah, I can say because I mean, I it's like it's the same with price tags or anything, but it's like if I guess if you could predict when oil is going to go to 160, then you shouldn't even be the one doing this.
00;26;25;26 - 00;26;48;13
Unknown
You should be playing the model. Yes. Yeah. Thanks for going back to your contracts. To your. Yeah. Like that's just all. But I also like is pretty much all this is deterministic or are there any probabilistic, you know. Yeah. Yes. Right now it is deterministic. Let me just finish talking about it. So for FNA it could be so you have an existing asset base and obviously everybody's still looking to do it and as well.
00;26;48;15 - 00;27;14;16
Unknown
So for FBA parts similarly you can import your reserves data once the reserve team is has done finalizing the reserves you can import it into. Well, which is a one time activity. And then you can again layer on corporate items like Jenna Hedges, RB, all that equity. And you can do your budget modeling. You can do, you know you can estimate your forward looking borrowing base, using Bellwethers right now whenever people do that analysis, it takes them a week to do.
00;27;14;18 - 00;27;33;03
Unknown
And they get one number for what their borrowing base might be in six months, or, and then maybe you spend two weeks and say, okay, this is what it would be in two, like for the next two cycles, but well worth within a minute gives you the full forward looking estimate for the end of life. Well, for one scenario.
00;27;33;05 - 00;27;52;23
Unknown
So then when it makes the forward looking decisions for you for draws and payouts, it takes into account how your borrowing base is going to increase or decrease. In the future, your GDP base is going to be different based on that linkage. Right. So then what automatically takes your training schedule? It updates the, well, type or the development status of those wells in the future.
00;27;52;23 - 00;28;07;13
Unknown
So three years from now, if it's going to be a PDP, it's quite, exact. So you get the higher you get a better value on that. And compared to the pads at that point. Right. So over time, as your training schedule changes, your, reserve type changes, your borrowing base changes, and we are able to estimate.
00;28;07;13 - 00;28;27;07
Unknown
Yeah. That's awesome. That's that's one of the more, I think, exciting things that I'm noticing about our industry now that we've got a lot more technology focus on it is this concept of over time. Yeah. Because like, you know, you're doing something in Excel with whether you're in accounting or completions or drilling, it is a static thing.
00;28;27;07 - 00;28;51;19
Unknown
Right. And so that's also part of a lot of the issues we have across the data is that it's most of it is all static screenshots essentially snapshots. That's just been compiled together. But there's nothing looking at it from a time series. You know there's time series data of course. But you're not looking at how things are changing over time and how that one thing impacts these 20 other things downstream of that.
00;28;51;19 - 00;29;09;23
Unknown
Right. And so having these things I hate using the word digital twin because I feel like that's been kind of killed, because everyone's been talking about it for so long. But it's this concept, concept of basically like, yeah, this I do something over here and there's this Rube Goldberg machine known as an ENP, and something else changes.
00;29;09;29 - 00;29;29;04
Unknown
Or ten things down here also change. Right? And now that's multiplied by a thousand because I've got a thousand wells or whatever it is. And so, yeah, the whole living system thing I think is going to be tremendously valuable, which I never even thought about from a finance perspective. Like, yeah, it's like setting this off, like setting people off on a week long, two week long exercise.
00;29;29;04 - 00;29;44;28
Unknown
I mean, now with like again, like I think just everything was seen was like these feedback loops are there. So you guys are you know, I mentioned, you know, we had Matt and Josh on early from pod two, but like, yeah. So some of the things like you can iterate so fast on things that there used to be a software or a human and or human bottleneck.
00;29;45;00 - 00;29;58;07
Unknown
Yes. You know, where it's like, what do I really want to ask them to do it? Because this is going to eat up another week and we can't get to this answer. But if we can be like, but what does it look like if we change X, y, or Z? And we get that answer almost instantly? I mean, like, what's that worth?
00;29;58;07 - 00;30;17;25
Unknown
Like when we're making millions and billions of dollars of decisions? Exactly. Yeah. That's when we started out. That was one of the, sort of real world outcomes that we are visualizing as an because of some experience. We know that, okay, in her job, it used to be like the VP or the MD would ask a question and then sell, would go back, modify the model and come back with the outcomes in a PowerPoint or whatever.
00;30;17;25 - 00;30;34;14
Unknown
Right? To what we wanted to happen was in real time, all of them are sitting in a meeting room, someone dmd ask the question the analyst or associate or the MD themselves can just play around quickly and then you see the results in real time, right? Yeah, that actually helps decision making rather than just going to and fro between teams.
00;30;34;16 - 00;30;56;03
Unknown
Yeah. You're stuck in endless meeting loops. And I'm like, are you seeing are people using this a couple times a year for budget or are people literally, you know, obviously they'll usually have a they're one plus 11 to us. Well like are all they using. All. Are all those getting found in the well worth or does it. So it is replacing Excel based models for people right now for the corporate model.
00;30;56;03 - 00;31;12;24
Unknown
So they're using it for budgets. But as frequently as you would need to use an Excel based model, you're using Excel. So if you have an active drilling schedule, you might be using it on a daily basis because you can actually easily go in and change that schedule and see the immediate impact on the corporate side. So people are using it on a daily basis.
00;31;12;24 - 00;31;26;18
Unknown
Okay. Let's whenever they're negotiating a new term sheet, if they have to look, trying to renegotiate the terms change different rates or repayment schedules, right. They're able to plug that in well worth in real time and then say, okay, these are the five outcomes that you're looking at. This is a bull case is a bear case.
00;31;26;21 - 00;31;57;24
Unknown
Let's see how we can negotiate terms. Because these are outcomes. Contracts are going to be so different in the future. Yes we're going to talk about the future as well. Yeah. We're we're seeing our clients also using the LMS to help renegotiate or identify bad contracts, identify good contracts, help renegotiate them and stuff like that. But they're not even thinking about it at that level of like, let's not just think about it from a land perspective, but also, hey, let's look at everything else under the hood and how they can they can play out.
00;31;57;24 - 00;32;18;22
Unknown
And so, yeah, the future's going to be crazy. Yeah. So even if I just complete that thought in the part, so the companies that already have existing asset bases are still obviously looking to acquire assets as well. So. Well, but also so it helps bring together both the A and B and B and apart. So you could in the same project build out your existing company.
00;32;18;25 - 00;32;42;23
Unknown
And then all the in the opportunities that you have, you model them into the same project. Okay. So you have existing asset asset and one and two and three. And then you if you wanted to you could look at all the in the opportunities standalone. Or you could have a pro forma off base plus and D1 or base plus and D2 or base plus A and B one plus and D two.
00;32;42;23 - 00;33;01;17
Unknown
You know, you can mix and match, which is super powerful because right now when that gets done, it's different for everybody, right? Sometimes people are doing it in D in Excel and the existing asset is in the reserves up. And when you end up acquiring an asset, you have to go through that whole process of integrating everything. But if it's in Valorant, it's already integrated.
00;33;01;20 - 00;33;22;12
Unknown
All you need to do is go to the table and say, okay, this instead of being a and D1 is now. And then also on the base finance side, it's also about you have investors on your cap table, right. How do you how do you deal with them in terms of I have an equity investor, okay, maybe I can hold back on the distributions this year and with that into an accelerated rolling schedule, how does that change my returns?
00;33;22;12 - 00;33;40;20
Unknown
How does that change the returns for the investors. So if I repay my low repay my term loan and then, use, use the equity draws so they are able to play around with these capital structure scenarios as well, which, which as you can see is very different from whatever you're doing on the and sector. Yeah. So we have built all these tools in there and then various custom.
00;33;40;20 - 00;34;03;23
Unknown
We've seen various customers use it in different ways. And the variants are because of a their asset base that in their capital structure. So it changes from customer to customer on how they're using it. Yeah. That's I wrote a blog. Yes, in the last two weeks. But I was talking about, it was all about some of my testing that I've been doing on some of the open source models.
00;34;03;23 - 00;34;27;20
Unknown
And ultimately the gist of it is, you know, iterations per hour right now or what I'm focused on because when what I've found is the less the less detailed the problem, the more iterations you're going to like, the less known the problem is, the more iterations you're going to need to end up finding what the actual end result, is.
00;34;27;20 - 00;34;55;14
Unknown
And so when you think about it from the finance perspective, or just operators in general, right, it's a risk game. Like they look at everything from a risk perspective, whether it's safety or financial. And so again, huge potential number of outcomes, right. When you're thinking about this holistically. And so it does make sense to have tools that can literally let you do a bunch of different iterations so that you know what your bounds are, right?
00;34;55;14 - 00;35;14;27
Unknown
Like show me the absolute worst case scenario and show me the absolute best. Yeah. And show me everything in between. Show me the range. Right. And now I know, okay. Regardless of what we what happens with oil, which is the only thing we can't control in this. Yes. Yeah. We at least know what what to do. And then you've forecasted and played out all those scenarios.
00;35;14;29 - 00;35;28;23
Unknown
Yeah. And you know it's ready to go. That's what the scenario analysis feature that we recently built. It was one of the most painful. Yeah. But one of the most fun features that we have because when you once you look at the outcome, it's it it sort of makes you proud that, okay, I spent so much effort building this.
00;35;29;00 - 00;35;46;08
Unknown
So it makes the effort worth it. Yeah. Yeah. Those are the best ones. Like like you beat your head. These are ones. The ones you can build over a weekend that's like, yeah, yeah it's done. But then these features, you look back and think very carefully. Yeah. Because it that feature touches upon the whole. Right. The whole code base.
00;35;46;08 - 00;36;04;08
Unknown
Right. It's like it takes everything from start to finish and it creates, it lets you create like 50 different scenarios all in one go. Okay. Yeah. And then it gives you the bands that John was talking about just now. It gives you the bands of like this is my upper limit, lower limit. It gives you a line chart for literally everything.
00;36;04;08 - 00;36;26;28
Unknown
Yeah. Well I mean that was the the promise of cloud compute, right. Like you're going to be able to run infinitely more things because we have an infinite number of things in the cloud. Right. Like and you can just spin them up and then that, that hasn't I don't feel like at least in our industry, I mean, maybe with some things, but a lot of the companies that have advantages in the software space, in the oil field, it's because they moved to the cloud, because you can.
00;36;26;28 - 00;36;44;23
Unknown
When we started out, we used to get so many questions about, oh, so my data is going in the cloud. I can't host it myself. It's like, you know, we were lucky that when we started out, this was still maybe five years at least down the line where people were fairly comfortable. Yeah, we still used to get that questions with those questions, but it's it's been a while since nobody asked said no.
00;36;44;26 - 00;37;03;25
Unknown
So that's good for the so what about like for like are people able to do like any kind of look back type analysis like in or because even like say like obviously were assumptions that go into it, you know, are they able to bring in actuals and do kind of like, you know, all right. Dismissed. Because like that was a big thing, you know.
00;37;03;27 - 00;37;18;17
Unknown
All right. So we missed our budget. You know, why is it the model or was it the inputs. Yeah I can I go other ways like that. That's like an after action. Yeah. That's that's manually possible right now. That's I think one of the next 1 or 2 features that we are looking at in the pipeline. I made it.
00;37;18;17 - 00;37;35;21
Unknown
Yeah. But right now it's manually possible in that you can have one scenario with the historical or the forecast. One scenario with the actual. So you can compare them with the scenario analysis. Okay. I'm going to automate that as well. So you can build sort of a profile of what has actually happened. Yeah. And make that some sort of assumption for this area.
00;37;35;22 - 00;37;54;10
Unknown
Correct. Or. Yeah. Yeah. Everybody's accounting system is different. So it wasn't easy. Like we're not there now that we have to work out, you know. Exactly. So we're going to maybe build a map for for the giving you being able to upload the loss statements, maybe like a similar exercise to how you bring bring in forecast the present forecast.
00;37;54;14 - 00;38;08;07
Unknown
Yeah. It's going to be a similar exercise to that. How do you bring in accounting data of various formats. Yeah. And create a mapping of my my data looks like this. And then every time it is because anything about like charts of accounts and then people can compare okay. This is the database that we got from the seller.
00;38;08;14 - 00;38;25;05
Unknown
And this is the loss. And so you can see whether the forecasts make sense in relation to like your unit costs, whether they make sense for the forecast versus what actually happened. Yeah that would be pretty powerful to. Yeah. Because again just that is what like finance just centralizes everything from pretty much every corner of the business. Right.
00;38;25;05 - 00;38;46;04
Unknown
Like yeah I need your all the error, the marginal errors across every department compounded into one. So I mean I've been there and done so like oh we missed budget or you know, everyone else is like, well their model sucks. And then they're like, well no, your your assumptions were terrible, you know. So like you've got to be able to call out, you know, look, if the assumptions are all within a reasonable band, now go back and say, we need to tweak the model.
00;38;46;08 - 00;39;10;16
Unknown
Yeah. But exactly. It goes both ways. Yeah. And then you ask the question about whether it's deterministic or probabilistic. Right now it is deterministic. We have thought about probably it was in our future probably in to add probabilistic analysis in that too. But we are more of a pull than a push company right now. So if any of our if we start hearing from our customers that that's something that they want added, we will add it.
00;39;10;16 - 00;39;27;27
Unknown
Yeah. But right now they're it's a different direction and stuff. You know, people want you know, it goes into the machine and it comes out and I can understand how it. Yes. Because you then you have to you have to educate people on that as well. And then we, we proposed it to a few of our potential clients and a few of our existing clients.
00;39;27;27 - 00;39;44;02
Unknown
And, you know, sometimes you send out polls like asking them what they would want to build. And that was pretty low on the priority. So we didn't know that, right? Yeah. Because right now the scenario analysis feature lets you run 50 scenarios. And yeah. So you get that. Yeah, you're probably in the band. You get the band. Yeah.
00;39;44;04 - 00;40;11;22
Unknown
Right. Yeah. Yes. It's much easier to work with clients when you can explain to them exactly how how we got the answer from A to B, speaking from truth on that one. But so I guess we can kind of move more into the. Yeah. How has AI impacted you all as a company from I mean, like obviously like I'm assuming making you all more productive internally and being able to, you know, get features out and be more responsive to customers.
00;40;11;25 - 00;40;25;05
Unknown
But then also, you know, maybe some of the stuff you're seeing in a while where you have to educate people, like, look, you probably don't want to do this yourself because, you know, we've had to go through all this already, you know, like so. So the other part, the second part, you mentioned that asking them not to do it.
00;40;25;05 - 00;40;40;05
Unknown
Right. We we prefer not to do it, not to tell them not to do it or in that way we are happy to let customers, not not our customers, said if they ask a question, will say, okay, this is how we can do it. This is possibly the cost you will incur if you are trying to do it yourselves, then you decide, yeah, right.
00;40;40;08 - 00;40;55;24
Unknown
We won't explicitly go out and say don't do it because there is a conflict of interest, right? Sure. So we don't want to undermine the trust that they have in us. Right? So and right now, so far, our customers have not really asked about whether and they want. Right. They won't ask, should I replace you with something? I'm sure.
00;40;55;28 - 00;41;11;20
Unknown
Exactly. Right. So the way we look at it, and it was a good branching that you mentioned, right. Internally we look at first of how do we do things better. Right. Building applications for building the application faster. That's one side of it. The other is the customer facing side of it, where again, there are sort of two elements.
00;41;11;20 - 00;41;34;27
Unknown
One is how do we make it easier for them to use? Well, what as it is. So to give you an example, if I want to run scenario analysis, right. So right now they have to create is what about 2550 scenarios? How do we, make it easier for them so that instead of clicking through an application, maybe they can just type in and say, okay, these are this is a list of the 20 scenarios I want to look at risking -10%, -5%.
00;41;35;04 - 00;41;53;13
Unknown
And then the system automatically creates this okay. So that's one side of it right. That's the presentation layer where we make it easier for them to use. Well. And the other side is the reasoning layer where we help them interpret and reason through the results that the data is generating. So those are the two elements we see on how do we use AI on the customer facing side of it?
00;41;53;17 - 00;42;19;08
Unknown
Okay. Where what are you using on just like a day to day basis? Are you using any grok bot or cloud or anything like that? We use a mix of cloud. So on the US design in particular. Right. So you use bot a lot, right? Yeah. So for coding, plot is really, really good. But for anything on the marketing side or other reasoning side chat, you would've found this way better.
00;42;19;14 - 00;42;38;10
Unknown
If you want to generate an image, it would be so funny. Like try to do that on Claude and check to out what does like a clip art where you love it. Yeah, and then jet GPT generates a whole like really beautiful image. So we use we use both of those for various kinds of. And then on the code generation side, it's like we, we, we sort of feel like there's a responsibility that we cannot just have.
00;42;38;13 - 00;42;52;11
Unknown
I write the code and review the code. So that has to be a human in the loop on at least one of those sites. So and for me, it's at least right now it's easier to write code. And rather than get AI to write code and then review it, I don't want to say too long review likes reviewing.
00;42;52;13 - 00;43;09;13
Unknown
Exactly. So the way we do it as I write it, and then I'll have some review cycle which says, okay, if I made a mistake, I'll be able to easily pick it out. Yeah, but for now, it's easier to, because like you said, it takes multiple iterations, depending on how difficult the problem is and how well you have defined it.
00;43;09;13 - 00;43;30;21
Unknown
Correct. So it's almost like I'm in a prehistoric age where I'm doing a lot not all of the code by hand, but, the way we are using it. So I use Gwen for example. So it helps, generate the, framework and then I can fill it, fill in stuff easily so that that's helped us, shorten the development cycles by big margin.
00;43;30;24 - 00;43;53;29
Unknown
Yeah, yeah. And it also depends on what you're doing. So when it does the really complex stuff, like, he knows the architecture, he knows how to edit the engine in the most optimized way. So even if we use yeah, right now we mostly use it for the, you know, the easier tasks of like, okay, the easier to sort of front end algorithms and then say, okay, this is the algorithm, this is how I want you to go through it and then write a function, and then I'll review it.
00;43;54;01 - 00;44;11;16
Unknown
That's another way. Okay. Yeah. But I'm still quite a distance away from, having a fully automated system where you have out running and auto mode. Yeah, I'm not comfortable with that because and part of it is, selfish reason that I know eventually, if there are bugs, I'm going to have to sit through all of it. Sure.
00;44;11;19 - 00;44;29;13
Unknown
Fix it. So I just wanted to. Yeah. It's hard enough to go back to your old code that you wrote, except debug it. Much less like, I don't even write this. But now when I look at code that I had written three years ago, I sort of feel, a bit of embarrassment. Oh, yeah. Yeah. So I don't want to get in that cycle of reviewing code that I has written two years down the line and looking at, I'm trying to figure out what's happened.
00;44;29;13 - 00;44;42;25
Unknown
That's the thing. It's a lot easier for Claude to find a bug in your code and be like, oh yeah, here's how to fix it. Do you want me to or do you want, do you want to? And you're like, oh, that's easy, go for it. Right. Versus it writing all of the code. And then you have to figure out the bug.
00;44;42;28 - 00;45;07;08
Unknown
Yeah. You figure out the rewrite and figure out all the problems. It's, Yeah. No, it's a man. It's a fascinating one. I would challenge both of you, like, just over a weekend side project, like, whatever. It's something to solve your own personal problem. Try it with plan mode. Try it. Actually, I'll send you the paper, try it with a coding harness in plan mode, and you'll be amazed.
00;45;07;08 - 00;45;34;15
Unknown
You'll sit there and talk. I did a 300 question interview with GPT on a plan mode. Or a plan for app. The other day. And I like how I was like, you've been doing this for two hours. You realize that? Like, yeah, but this is how it works. You sit here and you tell it and it's going to ask me all the nuance stuff that I am not thinking about when I'm actually coding or when I'm architecting that I need to think about along the way.
00;45;34;17 - 00;45;56;01
Unknown
And then it generates this giant plan, and then you say, okay, and it goes, but then you put that within a harness where it's watching itself and it's like, oh yeah, there's this bug. Let me go ahead and fix that. And it's it's wild. It's, it's a, it's just a crazy exercise of that is what they're proposing as the future of, of code development.
00;45;56;01 - 00;46;18;02
Unknown
But it's also fascinating when you see it and you're like, oh shit, this worked. But what about even like, I mean, because I mean, and I you too are like the core team. Have you had other developers or or is that, you know, really I think, what a blessing for us in, in terms of planning because we because it helps us decide, you know, when do we want to raise capital or do we want to raise capital that.
00;46;18;02 - 00;46;31;00
Unknown
And then where are we going to spend the capital on. Right. So we know that on tech right now, it doesn't make sense to hire external developers when we have all this help. Yeah. So that that's been a, good part. Yeah. So it's been able to. Yeah. Allow you to not have to bring on existing. Exactly.
00;46;31;00 - 00;46;47;19
Unknown
It's like, have we rushed in and then hired a bigger team? It would be like, okay, now what do we do with this? Yeah. Yeah, that is what we added capacity. Yeah. Now having it's weird because they like the way it's moving is they want you to think like grok bot. They want it to be like an interactive.
00;46;47;21 - 00;47;10;02
Unknown
It's like slack but with bots. Right. And so it's like, oh, hey, go. I need you to, you know, research this or fine design this or whatever. And then you go do whatever you're doing. And then they come back and like, hey, we're done. Or hey, it's time to review this. And you're like, oh, oh, okay. It just reframes how you think about how you could possibly do things.
00;47;10;08 - 00;47;30;06
Unknown
Yeah. Like I've got enemies just doing random stuff for me all the time. Yeah. But it's fascinating how much, you know, whether it's, hey, I want to create this thing or, hey, there's a known bug, and I just need you to figure out what it is and report back to me so that I can go fix it.
00;47;30;08 - 00;47;46;06
Unknown
Have you tried to get it to sort your folders yet? Oh, yeah, that was done. That. Yeah, it's on my to do list right now. I haven't done it. But you know, I had Doug Ford coworker do that or it was either a coworker GTI, but, yeah, on my desktop in my downloads folders are very organized.
00;47;46;06 - 00;48;07;01
Unknown
And now it's on a cron job and it just runs every week or whatever it does. Yeah, we just finished building two really big features. So that's like, I know now my to do list is going to start becoming shorter because I have a bunch of stuff that I need to get it to do. I really like raring to sort of start experimenting because like, these are features I cannot spend time on researching more until we get these things done.
00;48;07;04 - 00;48;27;08
Unknown
Now it's now. Now we are excited about okay, now I can spend more time on thinking about what else can I do? Yeah, it's it's crazy man. Yeah. I mean, we were talked about before we got on, but it's like, again, we're all very technical people and again, probably in the top 10% of AI users. But like, it's so easy to feel like you're so far behind because like, you see people that you know, in that, that 10%, there's a whole nother ranges.
00;48;27;08 - 00;48;48;13
Unknown
Oh yeah. People in like, it's, clinical life is automated. Yeah. No, I mean, that's where it's going, and it's it's crazy. But it's also interesting. Like, I find myself like, I. What? My car wouldn't start the other day because the starter finally went, went on it and I was like, oh, well, I've got this. Got my agents running and they're they've got their plans.
00;48;48;13 - 00;49;03;17
Unknown
I'm going to go replace the starter. And my laptop was just sitting there, you know, and I'm like, oh, okay, this might be a really cool future where it's like they're just working for us all the time and we get way more free time to do shit that we want to do. Yeah. I hope that's the that's the route or down.
00;49;03;17 - 00;49;23;19
Unknown
Not the alternative, end of the world route, but. Well, there but also then it comes down to even say like how how we manage people and do things. It's like, do we just expect more from one person? And just like like, I mean, that's what it's absolutely getting to. And are you still working or are is there 20% of people are going to work eight hours a day, right?
00;49;23;20 - 00;49;41;26
Unknown
No, because they're the most productive ones. Like there's more pressure. There's more stress because you feel like you could be doing so much more. That's a lot of your I have. Yeah. Your your laptop is not does not have a sleep mode anymore. It's going to it's it's got to remain on once it's done with something I need to give it the next half a laptop.
00;49;41;26 - 00;50;00;03
Unknown
I mean you gotta you gotta have a GPU at home that's running all the time. Yeah. No, that's another real thing. I mean, for what is it, 4500 bucks? You can buy a DGX spark and just sit there and run whatever the hell you want. Any time. It's crazy. Yeah. Is it still 4500? I'm surprised. Yeah. I'm prices, even GPU prices.
00;50;00;03 - 00;50;19;16
Unknown
I think they were about 4500 a month or two ago. Now the 5090 is 7500. Oh my goodness. 2000 micro center keeps hitting me in the in their weekly email ads is like every every other week. There's one with spark in it and it's a 4500 like let me get me a good looking price. I'm thinking depreciate.
00;50;19;17 - 00;50;37;21
Unknown
Well, but again, when you think about it in this concept of like, hey, Bobby, you're meeting with clients, you're coding your core stuff that you need to be working on. But hey, I need someone doing, you know, marketing or managing my inbox and managing my calendar or, you know, like I just spun up our SMEs agent to do all my insurance.
00;50;37;23 - 00;51;05;04
Unknown
It's like my, my, our company insurance renewal was up. So I dumped the policies in there and told it about the HSA that we use. And I was like, hey, which one makes the most sense? And I was like, well, actually, if you contributed an extra 100 bucks a month to your HSA, you could go with the cheaper plan, because the gap between the cap of the HSA, whatever it is now, I think it's like $9,000 for families is only like $1,000 less than what the out-of-pocket max would be.
00;51;05;04 - 00;51;30;21
Unknown
And so it's like you're also saving the money. You're saving is actually yours instead of like, oh, well, why don't I do this for homeowners and auto and all my other crap? Yes. So it's amazing you would not not have had the time or the inclination to go that in-depth into something that you're comparing right now. Like, yeah, like if we were looking at cars before and we just like you just copy and paste a bunch of links into it and it says, oh, this is the best option.
00;51;30;21 - 00;51;47;16
Unknown
And you would otherwise if you were looking at a car, you would have to look at so many different things, especially if it's a used car and you have that one. It does all the analysis for you and then like two minutes. Yeah, it's the green I got. Like I got the, you know, in the actual physical mail thing from base core, you know, like, oh yeah.
00;51;47;18 - 00;52;05;15
Unknown
So I fed it everything about that and said, here's my current, you know, what energy is doing for me now, which probably won't be so but like, know compare it like, what am I missing? Like, well, it's actually be the same price per month, right? Is it actually a good deal? Yeah. It sort of the whole thing out and I rebuilt energy.
00;52;05;15 - 00;52;18;17
Unknown
Oh girl. Yeah. In an agent as well. Yeah. Because I used to once a year, a couple times a year to do the whole party juice and pull a couple things down and have a little Excel sheet. When's the break even on this or if I use 2000. Yeah. But like, you don't have to do that now. Yeah, we have that spreadsheet too.
00;52;18;17 - 00;52;33;21
Unknown
But we probably need to convert that into an app now. I mean, I did it for my brother in law's job one day and I was like, right, maybe a year or so ago. And cloud was getting really good. And I was like, yes, I mean, the two efl's or whether they're called like just fed them into cloud and like, tell me the difference.
00;52;33;21 - 00;52;49;09
Unknown
And here's the usage per year, average usage and like, oh, this is the best one because actually, you know, the some other thing, a lot of people don't realize that your home meter is a probably a smart meter, and it has an API that you can get access to. So you can go pull your actual actual data.
00;52;49;10 - 00;53;04;00
Unknown
Yeah. Historic data. Well, that's what Energy Hacker does. What you know, you pay them ten bucks a month and that's all they do is manage your stuff. But the first thing you do when you sign up is you connect your smart meter API so they can pull your data so they can run it against the plans. But yeah, I did the same thing.
00;53;04;00 - 00;53;27;06
Unknown
I was just like, well, hey, here's power to choose. Yeah, pull data from that every time we run this. And then yeah, it's crazy. Anyway, fun, fun stuff. Fun fun nerd stuff. It's it's great. So I mean, it could be about well worth or just the industry or the tech. I mean, what do you guys most excited about the next like 2 to 5 years?
00;53;27;08 - 00;53;53;13
Unknown
You have any big features coming up? Yeah, we're as we said, we just deployed these couple of interesting features. And so now we're thinking more about AI and how to have how to make the user experience. Because, you know, obviously people just want to type stuff into stuff now and say, just give me the answers. People think that, wow, I'm still not sure I buy off on just a chat.
00;53;53;13 - 00;54;09;00
Unknown
Interface is the sole interface everything is going to operate out of. I don't know that I'm still I'm fully bought in on that yet, but I mean, everyone thinks they do. I going to the left in that you will still need clicks or are you saying it can be voice operated so you don't even have one voice operated?
00;54;09;00 - 00;54;32;03
Unknown
Yes, I think that becomes a default for a lot of things. But I'm also saying that like you need specific views for different departments based off, you know, a drilling engineer is going to want to look at depth. Yes, up and down. Right, or however they want to view it, versus an accountant who's going to want to look at a completely different plot in a totally different way.
00;54;32;05 - 00;54;54;10
Unknown
And so like the the concept of everything being done through a chat interface, I don't think makes sense across that. Yeah. Only that. Right. Yeah. It needs to have the base as well. Like you're you're trying to stick. Yeah. You're deterministic software becomes the architecture for AI. So if somebody still wants to use, well the traditional way they can.
00;54;54;10 - 00;55;12;22
Unknown
But if you wanted to not look at all of that. Yeah. Still have the deterministic software. So that whenever you give it one plus one it always gives you two. Yeah. But you don't have to put in the one plus one. You just have to tell it to do the one plus one. You take for instance, you just give it the files that you have and you say, okay, build this model of developers.
00;55;12;23 - 00;55;37;05
Unknown
For me, it does it for you. Okay. And if you wanted to go back and see what it did, you can see it. So that's what we're envisioning with developers right now. So we're excited. We're finishing off this, almost done. Right. Yeah. There's something called a PSC, a production sharing contract where, in companies in, countries where the government owns the asset, they bring in foreign operators, and then they take a cut off the top, they take multiple cuts through the entire value chain.
00;55;37;05 - 00;55;55;03
Unknown
They take lots of modeling that lower 48 economics is very different from that. Right. So we had a, international customer reach out as a potential customer back then. And they said, can you guys do this? And so we usually don't build features on an ad hoc basis, but it's okay. This is something what an interesting challenge. So we took us a few weeks to build that out.
00;55;55;03 - 00;56;12;19
Unknown
So now now we got it to a point where it's doing exactly whatever they are. That's awesome. So and for us it was a good, learning experience in that we had never I'd never seen economics one gas economics model this way. Right, right. So for us, it was the challenge was how do we fit that into the way we are running economics right now?
00;56;12;21 - 00;56;32;10
Unknown
That's a good exercise. We managed to do that. So once that's done as in we'll just finish that off. And then we want to the midterm goal for us milestone is get it to a point where customers can describe what they want to do and then have the application do it rather than having to click through. Like I said, build the risk profile, build the financial statements so they can describe it.
00;56;32;10 - 00;56;58;10
Unknown
And then the engine would suggest, okay, these are the things that you can, build variants on. Do you want me to do it? Which one? Which model is harder in your experience? Which model was harder to build? The lower part? It was much easier because it's fairly tabular. There's no, multiple waterfalls or things like that with the PSC model that that had a lot more, variances, bussing.
00;56;58;11 - 00;57;18;12
Unknown
And again, the, maybe I'm wrong on this as well, because if I build it from scratch, maybe both of them are equally easy or difficult to build. But I think the challenge over here was fitting that into what we already have. Right? And end of the day, these companies, they still operate where they have lower 48 sort of assets, and they have these assets.
00;57;18;18 - 00;57;36;14
Unknown
How do you combine all of that into a financial, a GAAP sort of financial statement? That was the tricky part. That's crazy. That's awesome. And this is what I'm talking about, as in building the algorithm for this. Right? Okay. Looking at because you know the application, you know what needs to get built building the application, I could obviously you can spend two hours to write a really good prompt.
00;57;36;15 - 00;57;56;10
Unknown
You'd still not be able to One-Shot it, but you might take certain iterations to get that done. But where I am right now, it's easier for me to just build the algorithm and then break it into smaller pieces. Then I write fill in all the smaller functions. That was easier. Yeah. No, I was just curious, you know, like the at least on the mineral side state side.
00;57;56;10 - 00;58;17;05
Unknown
Right. You have you can break it up by depth. You can. And then you've got all the distribution against all the grandkids and all of the ownership stuff. Whereas internationally there's no individuals involved anymore. Right. It's so it's cleaner on that, that side. But then the contract terms that sounds like have much, much more interesting provisions built into them.
00;58;17;07 - 00;58;33;18
Unknown
One thing I'm thinking about when we're talking about earlier is like, how does this well worth work within, like a larger company was, I mean, just throwing a name out, let's call it ConocoPhillips. Like where you've got, you know, all these different business units and even asset like, because people could use just well worth within as a singular asset.
00;58;33;20 - 00;59;03;17
Unknown
But is it also possible for like, say, a finite person in each asset to work there and then for it all to roll up into a corporate like an overarching corporate model or. Yeah, and it could so that's a really big company. But yes, you could have. So you could have multiple people use the software. And there could be, you know, you could limit some people to see certain pages and you could limit people for various projects.
00;59;03;17 - 00;59;26;08
Unknown
So you could have, business unit projects that are, you know, limited to people in that business unit. And, there could be one project that is basically the whole company. So for FBA purposes only, then maybe the finance team has access to that project. Okay. So depends on how I guess everybody. So well, was it so modular that everybody's workflow we've tried to accommodate for that.
00;59;26;08 - 00;59;45;19
Unknown
So if you wanted to remain within the same, same project and have people just look at a few pages, then, you know, we can allow for that to happen. Right now we're mostly working with small mid sized companies, so we haven't really prioritized, that feature rolled it out yet, but it's like it's trivial for us to just call as an, like getting to that modularity side of it.
00;59;45;19 - 01;00;01;06
Unknown
Right? You can create subsets of Wells Caldwell sets so you can have, say, your entire company might have, say, 50 bucks. I'm using random numbers for 50 wells. You can have a subset of ten wells. You can have a subset of 20 wells. So one team that's interested in only these ten wells, they can run a full fledged analysis, including up to financials on only those ten wells.
01;00;01;08 - 01;00;16;11
Unknown
You can have another team looking at the other 20 wells. At the corporate level, you can look at all 30 wells together. You might have a different capital structure over there. And because these are all plug and play right, you can change the capital structure in real time. You can do that. So you have, like you said, analysis where one team the other team and the overall.
01;00;16;16 - 01;00;31;13
Unknown
Okay. And then we have these ownership profiles where you can have some wells which are overlapping these two. So these ten okay, might have three wells which are also present in these 20. Right. So the ownership of this group and that might be working, this might be say 50. And then this group has a working version. Additional ten okay.
01;00;31;14 - 01;00;47;28
Unknown
So you can keep changing that as well. And that is another piece in the selections that you make. So with this selection now you're looking at all 30 wells and a 60% ownership in those. Okay. So you can do all of that play around. And then each of these can be a scenario then. So you can look at how do they perform against each other.
01;00;47;29 - 01;01;07;27
Unknown
Nice. You can look at how is each unit performing compared to. Yeah you. It can be as complex as you want it to be and it can be as simple as you want. That's one thing that that was one thing that we took pains to make sure that it is modular in that, okay, so so people can use whatever structure they want to use, and it doesn't stop you from, using it doesn't force you to use stuff that right.
01;01;07;28 - 01;01;31;23
Unknown
It's not relevant to you. And also when when we onboard people, we help them bring in their data and set everything up. So we go through the analysis of what's the best way for them to structure their project. And then once it's done, then we know we also have to provide like a help file, which is a PDF, and it walks them through what the best practice for them is like, how will they, maintain their project on an ongoing basis.
01;01;31;24 - 01;01;48;22
Unknown
Got it. And that think that we're still looking to improve, like you said, moving data between us because this is such a pain, right. So we have built the data import wizard. It does a lot. It makes it a lot easier. I feel we can make it even easier. Like you just dump in a random file and then the engine looks at it and says, okay, this is what I've seen, this file.
01;01;48;24 - 01;02;05;24
Unknown
Okay. Is this interpretation correct? Make just like man mode, right. Make changes to whatever interpretation I've made. Okay. Now go then it's going to import all of that data. So we want to get to that. That's the next step. That's sort of going to be awesome near to mid term. Then no it's pretty wild because you I mean they're at the point where you could do that right.
01;02;05;24 - 01;02;27;28
Unknown
Like yeah. Hey this is a financial report or whatever it is. Right. Look for X, Y and Z things. Yeah. If you find them let's match them. And then yeah, you store all that data and you make it smarter the more it does it. Yeah. This basically whatever I do manually right now, I can we can just write, like logic around it and tell it what it needs to do.
01;02;27;28 - 01;02;44;19
Unknown
Like, okay, this is if I let you got check if it has water costs, check. Does it have water production as well? Okay. If it doesn't then you tell the user that it needs water protection. So that's happy with the sort of the path that we have taken in that took a lot of patience to not get in the fancy stuff and just grab some of the engine.
01;02;44;26 - 01;02;57;25
Unknown
But now we are in a position where we have the entire engine built out, and now it's just a matter of, you know, exposing the right tools that they can call. And then make this install, because the reasoning is never going to happen with the numbers, as in, it's not going to run the financial model within the engine, right?
01;02;57;25 - 01;03;19;09
Unknown
It's going to require the tools. So it's just a matter of exposing the right tools for this engine that we have built, and then do the reasoning within it. Now I think that's going to be really nice. It's going to be awesome that time, right? Yeah. This wow. That was go fast. It went so fast I know I haven't yeah I genuinely was shocked that that well reads what it is now.
01;03;19;16 - 01;03;40;02
Unknown
That's crazy. Yeah I thought we wouldn't have much to talk about everything. Okay. And then we started talking and then we could go on for hours about it, but I'll get into, like, a little speed around. Yeah. What is your go to? Are y'all. Y'all are in Houston. Yeah. Okay. What's your go to restaurant when you know someone's in from out of town?
01;03;40;02 - 01;03;59;15
Unknown
Where are you taking them? In Houston, I guess is good. I mean, it's better than good, right? So this is the move. This is my my. It was the most with a few people. And I know, I guess most people who've eaten Indian food might not even agree with me. I think the food at Himalayas is better. The eating experience at Argus is.
01;03;59;15 - 01;04;17;27
Unknown
But. Gotcha. Yes. Okay. So overall atmosphere, this is the choice of getting takeout. We would maybe go for the Himalayas if it's a matter of taking someone to go eat Indian food, we might go to our. So that's interesting. And also in terms of food, I really like the food that I've got. The Afghan village might not be the nicest place to go to, but their food is really, really good.
01;04;17;29 - 01;04;42;17
Unknown
Rumors that Afghan villages are aircraft. Okay. Yeah. No, I saw that August is like the most ordered DoorDash restaurant in the country. Yeah, yeah, yeah, yeah. Well, they just opened one at 99 and ten and it's just to go, wow, it doesn't have an end. Like there's no restaurant. It's just to go only I guess because they had so much DoorDash to me, which is why, yeah, that's why I'll.
01;04;42;17 - 01;05;03;14
Unknown
Yeah. And they did anything about that. Yes. We didn't even get into this like what's your favorite cloud platform or where they already develop and run and run. What's good for being able to spin off, instances as easily as you want quickly on a, on demand basis? I'm experimenting with one right now, so it's good.
01;05;03;14 - 01;05;22;12
Unknown
I haven't played around a lot with the models, but I like when for now. But again, like I said, I haven't had a lot of time to experiment because of the yeah, a couple of features. But now I'm going to get deeper into this. I really I use Quinn three eight. I think on a on a daily basis and my enemies as my like main orchestrator.
01;05;22;12 - 01;05;45;04
Unknown
And then the crazy thing with Deep Sea, the flash model is so fast and it is so dirt cheap that even if it's wrong multiple, you can just iterate. So like for example, I did this classification test and I don't remember the times, but it was like literally one tenth the speed and cost of any of the GPT models.
01;05;45;04 - 01;06;06;24
Unknown
That and I could hit it with 60 concurrent instances instead of just the one on GPT. Yeah. So I was just like, yeah. So it's not just compounding model to model, but then I can also run 60 parallel. So it really can balance on. And with 60 still a fraction of the cost of the one public or no open source model.
01;06;06;24 - 01;06;24;09
Unknown
Just crazy stuff. What's your take on, you know, owning the hardware versus running it in the cloud, especially given that Ram prices? I don't think that you guys think the ramp will get back to normal in the next two years or so. I mean, has a backup right? Like I think the demand is over or is it overstated or.
01;06;24;10 - 01;07;02;07
Unknown
I think it it always is. Like every time it doesn't. I'm not saying just AI, but like any kind of new technology, they always dramatically overestimate the the short term demand is really bad at short term. We're much better at well, I mean, depending on who you ask. But and so I don't know, I think the, the more interesting part is what they're able to do with all the quantization and all these new techniques, like, I've got 96 gigs of Ram in this old laptop with a 4070 on it, so it's only got eight gigs of Vram, but I can put a 60 gig M0 model in the Ram.
01;07;02;09 - 01;07;21;02
Unknown
Yeah, and it only loads eight gigs into the GPU and I can run the whole thing from that one gaming laptop. And so it's like, that's fascinating. And what this could look like, you know, in the future. And so you've got, you know, all the premier labs saying we need more. We need more and more horsepower more, bigger, bigger, bigger.
01;07;21;07 - 01;07;36;15
Unknown
But then all the open source stuff rolls out, and then it's not days before people have a quantized version. You can run, you know, in two gigs like, and that's crazy. And so I struggle with it. Right. Like, let me say this too. Like, I mean, some of these models are as good as you need them to be.
01;07;36;15 - 01;07;50;05
Unknown
Yeah. Now or they were like months ago like, so it's like and they'll be for like you said like one classification thing, like it's as good as it has to be. Right. For that use case. Like, I mean, so things will continue to get better and maybe we'll get to a point where like, oh, you know, fable that was trash.
01;07;50;05 - 01;08;08;02
Unknown
But like, yeah, but it already does a ton of what we need it to do. And that's like that's the other thing is like if I'm doing a small a genetic specific task, I don't want a thinking model. I don't want a big, heavy, slow model. I want something that is dumb and is not going to think and is going to do what I asked it to do, and that's it.
01;08;08;08 - 01;08;21;19
Unknown
That's what you want the coding, right? But I mean, I guess going back to your buying hardware, I mean, it's just it's going to come down. I mean, I obviously there's a race to build more data centers, so I'm sure there'll be plenty of compute. But I think you've always been paying a premium to use cloud compute. Right.
01;08;21;19 - 01;08;40;11
Unknown
Like and so if that's probably the same as it ever was, like if you, if, you know, you have this baseline of compute that you're going to need consistently, it probably makes sense to have some of your own compute, at least hard going up. I think, like I said, the 5090, it was 4500 a month six weeks ago, maybe 7500.
01;08;40;11 - 01;08;58;05
Unknown
Yeah. The only listings for 4500 will say out of stock. Yeah. Now it's, Well, and that's where I like I wonder is it's like, okay, well, you know, is it going to boil down to you have to have your own hardware or you can't get enough access to open like even on the open. I mean, you guys run into limits.
01;08;58;05 - 01;09;24;11
Unknown
I mean, you're like so you're getting throttled by Azure, AWS, whomever, I mean like so I mean, if you have your own local stuff that. Oh yeah, it's a very real risk on our end. Khan has brought that up quite a few times. And yeah, we're actively looking at ways to mitigate that because of it. Right. But even like, I mean, because you're selling something that stands on top of that to your client runs and then it comes back on you, I mean, or, you know, it's hard to even give a quote on what is this going to cost?
01;09;24;11 - 01;09;44;23
Unknown
Because, right, you can't you don't want to give a fixed bid per se, because. Right, 10,000 a year and then all of a sudden it cost me 400,000 a year to serve $200,000 contract. Right. We're hemorrhaging money. Well, and there's there's still the elephant in the room in my opinion, around like, all of the frontier labs are hemorrhaging money.
01;09;44;23 - 01;10;03;24
Unknown
Yeah. At some point the prices will go right. Yeah. And so like, maybe that is the correction on the hardware side is that prices now go up. And so people quit using the cloud stuff as much I don't know. Or but it's so easy. I mean the average thing if someone's Joe who's going to have or the optimistic view is when cloud computing came around there were similar thoughts.
01;10;03;24 - 01;10;20;12
Unknown
Right. Okay. What if the price of compute goes way beyond my unit economics right now? Yeah. What would happen. Right. So you have you've got yeah. You would have plans in place. Well okay I'm going to go out and buy a server and I'm on my application of it. Right. But again they never went out of reach. So now practically every SAS application runs on right now on Moore's Law two.
01;10;20;12 - 01;10;38;02
Unknown
Right. It should get faster cheaper. But the good part is there is just that one bottleneck of memory manufacturers being able to divert from the consumer to SVM. Yeah, I remember what five, six years ago the concern was that, you know, Bitcoin and all these were not efficient enough. And it was right. It was using so much. Now I was using so much more compute.
01;10;38;08 - 01;11;09;22
Unknown
Oh yeah I mean people don't credit the crypto space to all the that was the predecessor of all the GPU acceleration that we've had on the Lem side of things. But yeah, it's a fascinating space. I don't know, I, I think if you can find a hacky way to get, you know, 100 gigs of a model somewhere, you're, you've got a pretty cool little sidekick companion thing that you'd be, you'd be surprised at how much you'd be able to start using it for different things.
01;11;09;24 - 01;11;30;09
Unknown
You know, from your side now, I mean, this was this was great. It's good to see you guys. We appreciate you all coming out. Where can people find you? If they want to reach out or get in touch? Our website is well worth app.com, and my email is Samara at well worth ABC.com. Y'all are on LinkedIn too.
01;11;30;09 - 01;11;47;23
Unknown
Yeah, we're on LinkedIn to. Yes. Yeah. I mean, it's, I know your time up before we got on and but LinkedIn's not. Yeah, but hopefully the algorithm finds you. Maybe we should. Jacob, make this clip and post it and just see how many people see this. Yeah, that was it. I'll link to them and see if it throttles it.
01;11;47;27 - 01;12;10;20
Unknown
Right. Real. I'm talking crap. Yeah. Thanks for having us. Yeah. Thanks for coming out. Thank you. Thank you guys. Thanks, everybody. We'll see you all next time. While some may see them as the crazy ones, we see genius. Because the people who are crazy enough to think they can change the world.
01;12;10;22 - 01;12;14;00
Unknown
Are the ones who do good by.