00;00;00;00 - 00;00;31;08 Unknown welcome. Welcome, everybody to, another another episode of Energy Bites. John Callahan here as, as usual. Missing today. We're missing Bobby. He's still, unpacking his his new house and moving, so I don't envy him, and I, I know, he wishes he could be here instead of having to unpack and move, so. But, today we're here with with Wes Hammer, the CEO and founder of, of task. 00;00;31;09 - 00;00;47;19 Unknown Appreciate you coming on me. Yeah. Thanks for having me. happy to be here. Happy to talk technology, man. Yeah. No, you're one of my, I think I've been following you for a long time. Just because you've been very outspoken, but you've been open about what you guys are doing and what you're working on and and all that. 00;00;47;19 - 00;01;06;11 Unknown And so I love, like I mentioned earlier, I haven't kept up with it in the last year or so just because we've been so busy. But it's a good thing. it's, I'm excited to, to chat more about it, because I think what you guys are doing is incredibly needed for sure. Yeah. it's exciting space. You know, as we talked about before, there's so much happening. 00;01;06;13 - 00;01;22;10 Unknown it's really exciting to be building over the past five years and really see the transformation that's happened from, you know, five years ago to two years ago to right now. Congrats. And so, by the way, yeah. Thank you. Thank you. Yeah. I don't know what the average life of a startup is. Yeah. What you hit that five year mark. 00;01;22;10 - 00;01;40;13 Unknown That's a pretty good sign. Especially then you caveat that with the chaos that has happened in the industry over the last five years, and that's a whole like through 2020, through negative oil, you can make it to the negative oil price as a startup today. You got something. Yeah. There you get something. That's right. That's right 100%. You might just be you know, capital discipline. 00;01;40;13 - 00;02;02;24 Unknown But like, yeah, there's lots of lots that can go wrong. And when, when no one is buying anything. So I think it's, just on the startup side. Right. there's so many facets of it, and you have to get so many things right? Right. You can't you can't get, you know, eight out of things. Right? To to move forward, you have to get, you know, ten out of ten things right. 00;02;02;27 - 00;02;26;21 Unknown It's really hard to do from a growth perspective, from a capital perspective, from a personnel perspective. What product are you in the market? How are you going to iterate? What technology stack are you building around? Right. You start off building V1. you have a few client requests and that continues to grow. Right. And then you have to completely cut that down so not to go too far into it. 00;02;26;24 - 00;02;49;06 Unknown John, by any means right away. But I think that's the that's the struggle. Of course, that's the, the pain in it. But that's also the, the great thing about it is you have to line everything up. Yep. And that's fine. Yeah. No, that's I think I remember reading, Crossing the Chasm for the first time and being like, man, this is the most, like, enlightening thing ever, because it's just gives you a, a framework of how this works. 00;02;49;06 - 00;03;05;25 Unknown Right? How you introducing technology products into the market works, the psychology behind all of that and all that stuff. And it's like there's so many people out there that have a good feature that they're trying to roll out as a product, but they don't have any idea about, like how to make it a product or why it's not getting traction or X, Y and Z. 00;03;05;25 - 00;03;23;11 Unknown Right. For sure. That's I agree, it's so tricky and people really don't understand. And the the older I get y'all, we're talking about, you know, startup mascots. That's, I have that in my LinkedIn title because I love it. My wife asked me one day she was like, why do you like the startups? Because she's very risk averse. 00;03;23;11 - 00;03;40;14 Unknown She's very conservative with everything. Right. Hey, talk to me, brother. I was like, and she was like, are you a masochist or like, whatever. And I was like, no, it's just like, that's what I like doing. I like building, I like impacting things and moving fast and all this stuff. But that's where that kind of came from. I was like, no, it's not like I can start a masochist, right? 00;03;40;14 - 00;03;59;00 Unknown Because it is. It is legit. You just like taking a beating. every day, every day. A punch in the nose every day. Yeah. And it's it takes a certain type of person, a mindset to be able to deal with that. Right. Like. And so it's, Yeah. And I think how I describe it or how I relate to it as well, is, yeah, you can definitely have the masochism in you to start this company. 00;03;59;00 - 00;04;14;29 Unknown If, no big capital round all these things. Right. Bootstrap it all these things, at the same time it's passion. Right? Going back to your point on, I really love this and I really enjoy and, you know, to be building right now, as we talked about, right when right when I came in, it was just so exciting. 00;04;15;02 - 00;04;38;15 Unknown And so if I were not building right now, I would I would not enjoy that. Right. That would be too much. That would be too much for me. Right? Absolutely. Yeah. I mean, that's, Jessica you're right. Uriarte I won't screw that up. sorry. Jessica. She was on a couple of months ago and she had transitioned, I think, around Covid to, not to a tech company away from the industry. 00;04;38;15 - 00;04;55;17 Unknown And now she's back. But she just recently came back to CORBA, and that's what she was. One of the things she was talking about was like, you know, I was working day data, working in data as working in data, generally speaking, right there, solving and solving problems with data. Yeah. But she was like but the big difference is I had no passion for the tech product that I was working on. 00;04;55;17 - 00;05;11;20 Unknown My passion is energy and oil and gas. And so it's a lot easier to suck it up and take one on the chin if you care about it or you see the potential of it and you see the future and you care about what you're actually doing, instead of just going in and checking the developer items and feature bugs and stuff like that. 00;05;11;20 - 00;05;33;11 Unknown But yeah, 100%, it's no, I think like it's where I was originally getting the older I get, the more startups I work out, the more I truly realize that, like, the CEO's primary role is to keep the company alive. And like, that's such a hard thing for people to understand. And totally it's hard to accept, even as a person who has been laid off from companies, that that's all they were trying to do, right? 00;05;33;11 - 00;05;53;05 Unknown It wasn't a personal thing. It was just like, we got to keep the lights on. Yeah, for sure. Like that changes the way you interpret everything about the company and how the decisions are made and all that stuff. And I think that's one of those really like especially in the startup early phase and growth phase, like, yeah, as you mentioned, you only have so many resources, people time and time. 00;05;53;07 - 00;06;10;22 Unknown And there's you got to balance it. But yeah, it's a hard, balance. And I think it's, it's, you know, you have different phases of the company. You have the, the initial MVP build from 0 to 6 months. Right. again, what are you going to build? What are you going to show? How deep do you build to show. 00;06;10;22 - 00;06;33;20 Unknown Right. And that's such a balance when you're talking to clients and when you're trying to get something is very capital efficient off the ground. to get your first chance to then, you know, get your data set, have your models working, providing value, tweaking that, talking about user experience, as they already have ten other applications to enter in. 00;06;33;20 - 00;06;51;20 Unknown Right? Which I think is such is is the hardest part. you talk about crossing the chasm, but I mean, the user experience, right? Is is one of the biggest barriers of entry of course you can talk about it. Of course you can talk about data, and all these other aspects that, you know, can become barriers as well. 00;06;51;23 - 00;07;14;03 Unknown But man, how the user enters, you know, the platform, what are they going to do? How is that in the overall ecosystem of the clients software systems? That's really hard to figure out. And so if you can't figure out that, you can't figure that out, you can't figure out utilization. Yeah, right. And if you can't figure out how the platform is going to be utilized because you have all of these. 00;07;14;03 - 00;07;32;00 Unknown Yeah. And if that's not defined. And so yeah, it's it's absolutely a journey. and I think a lot of it is working it backwards from that user. Right. And, and that's, you know, you kind of go into it with, hey, I've got this vision. Hey, this is a better way to do it. but you're not really working backwards from that. 00;07;32;00 - 00;07;52;21 Unknown So there's kind of a coming to a head. and I think you have to change your, your direction and how you're going to build your technology. You know, I think you're spot on there. I, I was very fortunate to kind of learn that early in my, when I was interning, actually, as an industrial intern in industrial engineering internship, we were doing all kinds of 5 or 6 Sigma stuff. 00;07;52;23 - 00;08;10;19 Unknown And, you know, I don't remember what the process we were trying to optimize was, but the first thing I did was go talk to the guys on the line that were actually manufacturing it and doing it. And guess what? They have a long list of things that they've already tried, right? But don't work. That'll probably overlap quite a bit with the things that you're coming to the table with. 00;08;10;19 - 00;08;28;14 Unknown And it's like, yeah, you got to start with the user and work back. So we've had a number of conversations, about like, is Excel a good place for an end user to be able to use data? And the answer absolutely, in my mind depends on what the user is doing, like comfortable and understand Excel. And that's where they everyone works. 00;08;28;16 - 00;08;49;04 Unknown It's probably the best place to serve the data, as long as you can structure it and put boundaries on it and all that fun stuff. Yeah, other times it's not right. It's like especially in our industry too, right? People do not like change. Generally speaking, people don't like change is hard. And then once you're in the enterprise stuff, it gets so much harder to introduce change into that hundred percent. 00;08;49;07 - 00;09;04;19 Unknown Yeah. And I think that change, you know, predominantly has been, hey, we're going to have this new platform, this new tool or this new software. but it's not really defined. You know what? We're going to get out of it. Right? What the stages are, what the plan is for one year, for two year, how it's going to be used. 00;09;04;19 - 00;09;28;13 Unknown Right. And this is pretty basic, right. but there's there's rarely that top down communication when you do have this software partnership about, you know, let's talk about the how right. Let's talk about how it's going to be used. and so I think that these simple things in just communicating and having a clear path, I think should stem also from, you know, what's your strategy towards technology. 00;09;28;13 - 00;09;46;17 Unknown That's that's a broad freaking thing to say. But at the same time, I don't see much strategy ahead. You kind of see hey, look, we want, to to get, you know, digitization here. And this is level one and level two and level three. But beyond that, there's not really a working strategy. Right. What software are you bringing in? 00;09;46;17 - 00;10;06;00 Unknown What software are you bringing out? What's your thesis here. What are you building internal versus external. Somebody might have a snippet of things. Well, if there's something extra on the market that's better than what we have, that's not really the case, right? Because you have all these other things that you write. So, yeah, it's very real. I think at the same time, it's a really challenge to the software companies out there. 00;10;06;02 - 00;10;25;26 Unknown You've got to be you've got to be fucking good. Yeah, right. No it's true. You can't go in there selling a same flavor of X. You've got to really integrate within a day. You've got to have data at that first person's fingertips and make a decision that's better. And automating all their workflows and all their tedious things. Right. 00;10;25;28 - 00;10;54;26 Unknown All of that has to happen within the user experience upon onboarding. Right? Right. Yeah. It's not a we can't just you can't just accept the narrative, like, okay, this integration is going to be it's going to be a year integration. It's going to be a six month integration. That's crazy. That's that's absolutely nonsensical. Right. So I think that it has to it's on the software companies to be able to have a ten x product, a true ten X product across all these user, things that they're doing on a day to day basis. 00;10;54;26 - 00;11;20;17 Unknown Yeah. No, I, I'm right there with you. And it's, it's only getting harder because software and UI UX is getting better. Right. And so it's like that's the thing and you're spot on with it. Right. Like you had you look at the kind of energy oil and gas software space historically. You've got these big giant incumbents that are still operating on the same architecture that they were operating on 20 years ago, and they never changed for a variety of reasons. 00;11;20;17 - 00;11;35;01 Unknown A lot of it, in my opinion, was they didn't have enough competition. or yet then you got all this. Of course, the oil and gas cycles play into that and all of that. But it's one of those things that yeah, you look at like really come a curve comes along, it's like, hey, we're going to do this in the cloud. 00;11;35;03 - 00;12;10;01 Unknown It's like, not I mean not to water down what they're doing. Yeah, yeah. To make it simple or whatever. But like that jump was a step change for most users because hey, now I can throw as much compute as I want to instead of having to be limited by whatever resources I've got available. I worked for a basin modeling company for like four months, and, one of their big like, value ads was I think petrol was limited at that time to, I don't remember, six a some number of clusters or CPUs and then there's didn't have a limit. 00;12;10;07 - 00;12;37;09 Unknown And so like the petrol models would literally take two weeks to run. These huge gas based model three was like, Holy shit. You know, I didn't know that existed at all. Right? Like they're literally mapping the, the, the evolution of the ground over centuries or decades or centuries and, it's, yeah. I was like, Holy shit. You mean like they just click a button and then go away for two weeks and then come back and then hope that the model ran properly, like, that's a crazy workflow. 00;12;37;09 - 00;12;58;14 Unknown And there's so many of those and still like, yeah, it's it's just the crazy, the architecture problem about how you run the models. Right. And, and how you're iterating. I mean, it's still it's still painful. Right? It's still how do you also productionize that, you know, are you training, you know, this data set, and you're trying to look for, you know, what is my scorecard here for this model? 00;12;58;17 - 00;13;13;18 Unknown What am I benchmarking that off? how is that continuous? I mean, there's so much just there's so much in that. And to your point of, you know, you run the model and, you know, it runs overnight. It runs for two days, right? Or runs for a certain amount of time. And then you're going to review the results. 00;13;13;18 - 00;13;33;28 Unknown Right? Like even that is is challenging to do. Right. And so even just the the basics of getting it set up, you can come in with a lot of whether it's configuration issues, whether it's, yeah, data issues, whether it's complicated, some multivariate less theoretical and there's so many knobs. And so that's like the that's like the building blocks, you know, building the model. 00;13;33;28 - 00;13;57;24 Unknown Right. Putting it out into a production environment and putting out a production environment where you're, you're, you know, you're making predictions of, of a suite of models every day. whether it's, you know, generative models, whether it's, you know, time series models, whether it's, you know, image models, whatever the case might be, right? That continuous cycle, there's a CI, CD to it, there's a continuous improvement of it. 00;13;57;24 - 00;14;14;19 Unknown There's how are you going to, you know, update that model. What what are what is your scorecard? and how do you do that on a day to day basis? And how is the model improving? So wow, there's just a lot that is I would say not too much talked about. Right. Because everybody says, hey, here's a model. 00;14;14;19 - 00;14;28;21 Unknown Okay. Well, I'm going to plug your model in and plug this model out. Okay. Well how are you thinking about doing that? Right. What's your architecture? How are you guys set up right now. And then it's quickly. Then three months down the line they're like, oh yeah, well we've got this new thing that we want in it too. 00;14;28;22 - 00;14;44;25 Unknown Can you do that? Yeah, yeah, yeah. Oh, sure. or yeah. You, you build a POC and then it starts gaining traction. You're like, well shit, how do I productionize? And that is a whole nother level. And then once it's in production, like you said, taking it and it, it's a continuous improvement, continuous development thing like it never ends. 00;14;44;25 - 00;15;04;23 Unknown Yes. I think that's what people don't realize about software. They never stop. So like yes, there is always a bug. There's always a new feature. There's always an improvement. Something. Yes. Right. And so that's now that you and how is that, you know, very, you know, operating very smooth for the client across of course, you know, many, many, many clients here. 00;15;04;26 - 00;15;26;08 Unknown Yeah. But I say as a startup trying to streamline that so you're not spending. Yeah. And, you know, and I was just chatting with Colin about this. Right. But what you put out as your V1, right, your MVP, what you're putting out for users to use. And this is something that we've I would say been pretty, sensitive to how we do this in the market. 00;15;26;08 - 00;15;42;21 Unknown Because your first version where you're going to get client request one. Well, can we, can we change this and what can we add. This will really I want to know this. Right. And so you kind of you're iterating around that product. Right. You're trying to answer all these questions for your users because they have all these complex workflows. 00;15;42;21 - 00;16;03;25 Unknown And this is a workflow game in AMP, especially in oil and gas, as you have so much underlying workflows that you know, the product is not even touching, right? Yeah, it might be touching 10% of that right now. Right. And it's so important to put a scalable MVP. What is the future going to be out there first and iterate off of that. 00;16;03;25 - 00;16;22;22 Unknown Right. And so it's it's I would for anybody trying to or think about, you know, building their own software, trying to put it out in the marketplace, really go to the market with I think the common advice is, you know, you should be embarrassed about your MVP that you put out. And of course, there's some truth to you can't have everything refined perfectly and you just have to go. 00;16;22;29 - 00;16;43;28 Unknown But at the same sense, man, it can bog you down. Right? And we've kind of been in that cycle where we had to tear everything down and rebuild everything up because we were continuing to, to patch or to to fix the things. But we didn't want to fix those things. Right. So you stuck in this cycle of, you know, fixing things that you don't really want to fix. 00;16;43;28 - 00;17;01;21 Unknown And if you if you thought about this problem in a different way, you're beginning. Wow. If we could only have iterated around that to start. So I think that's also an important part on the software side on how you're developing and how you're pushing to production. How you're iterating is is huge. Yeah. Back to the masochist. Yeah, yeah. 00;17;01;21 - 00;17;22;16 Unknown Back to how many startups have, torn down and completely rebuilt. Yes. Some portion or if not all of their stack multiple times. Yeah, 100%. No, that's, Yeah, it's spot on, man. And so actually, I've got a question for you around that. Yeah. how do you, how have you what strategies, techniques have you seen that work and don't work around managing that? 00;17;22;16 - 00;17;41;29 Unknown Because it is such a fine line in the startup space, especially on the energy tech and the tech world. If P.l.c. just, you know, move fast, break stuff and iterate quickly. But there's ways to move fast that don't scale, and there's ways to move. And then normally the scale things take longer so you can't move as fast, right. 00;17;41;29 - 00;18;02;14 Unknown And you've got the whole project management. You got three things right. You can only have two. And I love this question. Yeah, it's a great, how we've done it. in, you know, in the beginning, of course, getting a PSC or getting something out there in the market is the the biggest thing that you need to do to engage the market on what the future can look like and how does it look like? 00;18;02;20 - 00;18;22;07 Unknown We're tasked with whatever the company is seeing is believing to like 100% under that 100. Like you have to have an algorithm that does something in the background and then spits out something is not nearly as good as letting the people actually see it, letting them have the hands on with it. And so and so I think you have to, of course, you know, there's other ways to approach the market. 00;18;22;07 - 00;18;39;16 Unknown You can say, hey, if we're going to build X, y, Z, would you be interested? That's it's harder to do with enterprise. I've heard these stories. I've seen some of it happen. And so I think it's a way to, to create demand. And that's completely fine. As another way to start off, and then you start really trying to build this thing. 00;18;39;16 - 00;18;55;21 Unknown Right. And so we, we get our, you know, first few contracts and we're trying to build this in the background to make sure that, you know, we're connecting the pieces together. And so where, you know, what database are you going to pick. How are you going to deploy models. How how is the feedback going to feed back into the models again. 00;18;55;21 - 00;19;11;21 Unknown What's the scorecard of accuracy. Well can we even measure accuracy right now or are we worried about our database or as you say is it on prem or in the cloud. And then yeah that changes everything. Yeah yeah yeah yeah. We early on we're like look we're not going to do anything on prem. Right. Was like you know this is this is so much easier to do in the cloud. 00;19;11;27 - 00;19;29;05 Unknown This is the path we're going to take. you know, but to your point, even on that small thing. Right. That was almost a it could have been a concern. I don't even know, you know, ten years ago or something like that. Oh, you're in the cloud and there's still companies thinking about that. Well, you know, I'm the data sensitive, and I'm kind of scared about this. 00;19;29;05 - 00;19;52;00 Unknown We haven't really done that before. And I'd say that's just that's just technology in general. Right. And how people are viewing what is acceptable risk or not. Right. And I think that's the some of the problem. Right. When you talk to. Oh absolutely. You know everybody your organization's all about risk. Yeah. Once I realized that yes. And the industry like first and foremost think about the risk with any decision or any product that you're trying to introduce. 00;19;52;00 - 00;20;16;06 Unknown It changes everything. Yes. Because I was so jaded for the longest time about like, why are we so slow with adopting technology? Why is this like we have all these people working on this really cool shit, and there's really low hanging fruit problems that we can solve today with the combination of SMEs and tech and everything. Yep. But it somebody broke it down to me and they're like, you know Facebook having 12 different models. 00;20;16;06 - 00;20;36;03 Unknown And then you not being able to get notifications is a very different risk than, you know, a drilling rig blowing out or, you know, Macondo or things like that. Right. Like that's how they think about these things. And once you can approach it from that mindset, it definitely helps. Yeah. And I think to that point, it means you have to have a high degree of accuracy, right? 00;20;36;03 - 00;20;54;28 Unknown You can't. And in the day and age of, you know, losing generations of AI or, you know, referencing wrong data or coming up with wrong predictions, it's, you know, it's not just accuracy, it's explainability. It's explainability too. Yeah. And so you kind of have to get all these things right, for the users to trust the system. 00;20;54;28 - 00;21;11;10 Unknown Also for the system to continue to be valuable on a day to day basis. And so I think that's that's always going to be a challenge. Right. And the biggest companies in the world are still facing this as a challenge. Yeah. How do you get it right for specifically the problem that you're trying to solve is is really the key. 00;21;11;10 - 00;21;30;21 Unknown But I'll go back to it. You asked me, you know, what did we you know, how do we build and scale? Because I think that's a really important question that could be, you know, I think you you could you could create a whole nother podcast on this, right? From, from everybody, which is super interesting to me. But yeah, I think you have to take some type of MVP approach and maybe break this down to four things, right? 00;21;30;23 - 00;21;49;18 Unknown You have to take some MVP approach to to get it out there, engage the market and get it stood up right. And that's like version. You know that's the first step right. And then you're trying to scale that technology out making sure that you're building around the right things. Right. And the third is really what is your differentiated technology. 00;21;49;18 - 00;22;11;12 Unknown And then how do you really scale the business from there. Yeah. And you know, I, I would suggest really trying to understand that before you start building, of course, sometimes it doesn't exactly happen that way. And sometimes it happens, of course, through trial and error. Right. Yeah. And you, that's why you have to rebuild things. Because what you know, what we know now absolutely would not have architected it this way. 00;22;11;12 - 00;22;26;28 Unknown You start right or built this way from the start or even the tools available. Right. Like yeah. Oh this is available. Oh my gosh. The the tools, the toolset within I mean it's like yeah it's like what's your stack? And it's like, you know, 12 lines. It's like 30 line of advice. Like it's not just like three, right? 00;22;26;28 - 00;22;51;04 Unknown Or you're not like, hey, I'm in SQL. Yeah, yeah yeah yeah, yeah. It's not that anymore. Right? It's like, what are you using for blogging? What are you using for, you know, identification. Where. Yes, yes yes yes. Right. Like what do you use for deployment, you know, all these things. So, I think that the key thing that I would say out of all of that is really focus on what your differentiator is going to be. 00;22;51;04 - 00;23;08;20 Unknown And this is, you know, sometimes you have a pretty good idea of that, at an overview, when you're starting up. But, you know, I think as you start getting more into it, right, and you start understanding the data more and you start understanding what the models are doing and understanding, you know, how you can make these different connections. 00;23;08;27 - 00;23;34;03 Unknown Then you have to go really deep, really quick. And by going really deep really quick, that means that you can't spend time on managing a database, right? You can't spend time on, you know, these other peripheral things that you shouldn't be controlling. And, and there's a, a relatedness to the MPs out there because it's the same problem that they're dealing with that you deal with in software about. 00;23;34;03 - 00;23;53;00 Unknown Again, it goes back to your strategy of how we're going to do this. What am I going to spend time on? What am I going to spend not spend time on. And his, you know, so me and he's trying to create their own database. Like what what do you why why would you do that. You know that's that's like there's a billion databases out there that do a much better job. 00;23;53;00 - 00;24;11;19 Unknown Yeah. And so I would really target that right. As find what your IP find what your core thing at doing. Well that you can scale up that you know really works, that you've seen work and is differentiated. Yeah. And spend the most time that you can there. Yeah. And sometimes that goes through cycles to figure that out. Yep. 00;24;11;22 - 00;24;30;28 Unknown if I, you know could do it again I wish we would have figured this out at the beginning and just gone deep there. Right. And so I think we we tried to in the beginning, we tried to paint this more holistic version of operations and all those things. And by no means was that wrong. But the entry point where then how do we scale up from there? 00;24;31;00 - 00;24;49;12 Unknown what is really important? Yeah. Now that's the trickiest part about software, in my opinion, is that it's it's iterative and there's building blocks. Right. And so you're selling the future state, but you're trying to get there through the guidance of your users as you're as you're building towards what you think is the future state. And then the users help kind of direct that or put bumpers on it. 00;24;49;17 - 00;25;09;12 Unknown Yeah, on either sides of it as to where you go. But it's, yeah, it's, it's it's such a hard thing. I think you nailed it though, because there are one I don't I'm curious to know your thoughts on this, but I don't know why so many people in our industry, like, are so adamant about doing everything themselves or internally. 00;25;09;12 - 00;25;28;04 Unknown It makes no sense. So it's not how most. Yeah, that's why tech exists, right. and so like yeah. Even internally. Right. Like the reason this whole podcast came about was because Bobby and I were talking about data and he told me about five Tran like, oh, I'd never hear five tran before. Let me look at this. I was like, oh shit, this does everything that I need it to do. 00;25;28;04 - 00;25;46;12 Unknown And it's a few hundred bucks a month and I don't have to do it. And if the APIs change or the auth changes or whatever, I don't have to mess with it. They just do it right. Like it's like taking time away from my place. Yeah. And it's like, okay, I could have hired an intern or a junior dev or however much. 00;25;46;12 - 00;26;08;05 Unknown And this is, maybe a few thousand dollars a year. Yeah, that's a very easy ROI. Like, why would we custom do this stuff if something else can do it quicker, faster and cheaper than me? That's literally the definition of technology, right? Yeah. Yeah. And so it's yeah, I don't, I don't I think, I think, I think with that I would say, I think the reason why, you know, the industry. 00;26;08;07 - 00;26;25;00 Unknown Back to your point on why would you build X? I think again, it goes back to, to not having to find strategy about what you're going to build, what you're not going to build, I think within MPs is, you know, and you can't discount MPs because they're there every day. They're dealing with this problem. So they have to solve it. 00;26;25;00 - 00;26;41;26 Unknown Right. They can't just say like, hey, I'm just going to wait for, you know, USB to find something, right? And maybe something comes along in our lap and then that solves X problem. And so you kind of it you're stuck in that pattern. And usually that's the way that you're operating. You're operating every day. And there's always false positives or reports not showing up. 00;26;41;26 - 00;27;01;13 Unknown Right. Or you know, the database complete loaded or you know, all these things that happen on your day to day basis. And so that is naturally getting worse right now. I think it's been hard for MPs to make the determination about what do they really want to put the efforts in to fix that? and, you know, I think that's some of the challenge. 00;27;01;13 - 00;27;23;01 Unknown And you've gone going back to, to your question on, you know, why why is it the way that it is you have these built in by cycles. Right. And you and I am a firm believer that, you know, of course, technology is everywhere. And you can you know, that you can go to this marketplace and, you know, sync up the, the technology tools of your choice. 00;27;23;01 - 00;27;49;28 Unknown Right. and get started. But at the same time, there's, there's been this building by where everybody thinks that they can build. Right. And so you're going through that and being a part of that number one. And then also seeing building software. Yeah. You know, on the other side, as a technology company, you're very messy. The effort required going back to the first point that we're talking about to manage to deploy, to do all these aspects of building software in the right way. 00;27;49;28 - 00;28;07;03 Unknown Right. that man, it's it's it's, it's a challenge, right. And it's a challenge to continue to continually improve that. Then you throw commodity cycles on top of it. Absolutely. Like, oh yeah. Yeah. Bill wrote all that and now it all got laid off and there's no one to manage or maintain it. No, I mean we should have bought it from the inside. 00;28;07;09 - 00;28;25;05 Unknown Like, hey, that makes love. And so we've there's many stories like this. Right. It was very true. and we've come across many companies that have had this exact story and like, you know, this guy left and now we're left with nothing. And so we need something holistically right now. And it's like, okay, perfect. Let's have a conversation about what's best for you guys. 00;28;25;05 - 00;28;42;24 Unknown But, you know, I think you have to go through that build cycle, right? And you have to go through the failure. You have to go through you embarking on something that you're not an expert on. right. And, and the Bay of technology's coming lower and lower. And so there is that intersection. What is that right intersection. Right. 00;28;42;25 - 00;28;58;05 Unknown I I'm not sure. Right. Yeah. But I think you have to go through the failure of that. And then you are better poised to make that decision. completely. You don't. I mean, it's so true, right? Like. And I remind myself with my kids of this, right? Like, you don't know what you don't know, right? And like, their worst experiences. 00;28;58;05 - 00;29;16;16 Unknown Their worst experience. Right? But it doesn't mean it's worse than yours, but it's just their worst personal experience. Their perspective. Exactly. Everybody's perspective is real. It is. yeah. Until you go through that, I think that is really where you can have a much better understanding of, like, okay, maybe we should look at, yes, outsourcing this or whatever. 00;29;16;16 - 00;29;34;29 Unknown You know, I agree, to that point, let's, let's talk about about task J tell tell the world what if they don't know which most of them probably do, but share what you guys kind of what your thesis on how it came about, what you guys do now, kind of, you know, I'll pepper you with questions. Yeah. Talk about I'm sure. 00;29;35;05 - 00;30;01;05 Unknown So where task we're focused on production operations. we are the I os of operations. What we're doing is we are plugging in of data, plugging in production allocation data, and then assigning that work out to, control rooms, to, field technicians, to engineers on a day to day basis or a, you know, mini issues that happen within operations. 00;30;01;07 - 00;30;18;13 Unknown we're doing work documentation and then we're doing production reporting off of that as well. And so if you think about your kind of holistically, how an operation is run, I would say if I could, you know, summarize or generalize this for the whole industry, nobody really knows what people are doing on a day to day basis. 00;30;18;13 - 00;30;39;25 Unknown Right? Like at a very high level for sure. Very. Yeah. Very high level. Right. Nobody really knows. And it's because you don't have a, a intelligent you don't have models that have a high degree of accuracy to be able to understand these issues. You mean my paper forms aren't? Yeah. You're right here in yeah. Operational in a lot of companies. 00;30;39;25 - 00;31;01;19 Unknown So. Right. and if you don't have that high degree because everything is, alarm base, everything is threshold rates, everything is a, potentially a SQL query can have models. Right. But those models are usually, you know, somewhat, pointed on a specific issue and not, I would say holistic to, work management. What everybody's doing to day to day. 00;31;01;21 - 00;31;19;00 Unknown So it's at our core, we're really trying to pair time series intelligence with work management. Yeah. And I think that's a different, if you think about those two disciplines, there's usually one versus the one or the other where traditional work management is, hey, Jim, I'm going to assign you a job. Jim, can you do that job? Okay. 00;31;19;00 - 00;31;41;05 Unknown Jim checked it off, but there's a performance aspect to it, right? There's all this production happening at the wellhead. What's normal? What's abnormal if you can't benchmark what the well should produce and reservoir curves have problems and 3030 day average is. And different models have different issues here. And so really paying this all together about what should happen on a day to day basis. 00;31;41;05 - 00;32;04;02 Unknown Yeah. Because if you have a list of let's just say, you know, ten things for weather, this field technician, engineer, control room operator, you know, find out is about 65% of that list is noise, right. It's it's flush production. It's variance. It's about, you know, comms. It's, you know, all these things that don't really need action. Maybe it just needs a set point change so you don't have to go visit on site. 00;32;04;02 - 00;32;31;08 Unknown And so really trying to take all of that together to intelligently dispatch work on a day to day basis. I mean, the simplest, the simplest example of that, right, is like, well, the production curve looks weird. Well, why? Well, I don't know. Yeah, well, it's because Jim was out there messing with like, yeah, point tubing. Right. And and no one knows that because data sets are completely isolated and, and you know, you compound that by the amount of thousands that people are doing every day. 00;32;31;08 - 00;32;51;23 Unknown Right? And nobody knows that you just change the set point from X to Y. Right. And that's the reason why I off. Well, you know, how do you know all these things. Right. And so yeah, it's about connecting these different data sources and, and models to the better explain, operations but to, to also predict what should happen, right, for tomorrow, for the next day and so on. 00;32;51;23 - 00;33;16;18 Unknown And so you know, I think we've, we've been on a journey for about five years. We've developed, different versions of our technology stack in different ways with different strategies. Right. which, you know, happy to go into that as well. and I think really, you know, for us, we're continuing to, to be that voice in that, you know, one system where you can tie all these things together that have been fragmented. 00;33;16;20 - 00;33;37;21 Unknown And right now users are using ten different systems, five different systems. Right. Even if you're using three, right? Three should try to understand what's going on on a day to day basis. is too much in my opinion. And they're probably all different departments and different images and talk to each other. Yeah, yeah. And so I think the it's how you connect it all. 00;33;37;23 - 00;34;07;13 Unknown and then it's what can you do adjacent when you connect it up to really leverage up. And this is the other interesting spot that we're in right now. Yeah. No you're spot on I think that's there's so much opportunity in this industry because of the silos. Right. And like because like coming from that I worked at a in industrial internship for two summers and coming from that five X like lean manufacturing, we literally had stopwatches like it took anywhere from 12 to 24 hours to know you're building wind turbine blades. 00;34;07;13 - 00;34;22;13 Unknown All right. Well, you know, and like, I literally was out there with stopwatches and like, we focused on. Okay, where do the tools need to be so that it takes the worker the least amount of steps to get to them when they need it? Yeah. And like all these optimization things. Then I got from that to working as a track engineer. 00;34;22;13 - 00;34;39;15 Unknown It's like I look outside and there's like 20 sand trucks sitting there charging to merge because no one told them that. We we screened out and like, but at that point, you know, natural gas was whatever 5 or 6 bucks an oil was over 100 oh, it's a ship as a matter. Right. It's that's another like crazy part of our industry. 00;34;39;15 - 00;34;56;06 Unknown Right. Is like once oil exceeds a certain price, no one is worried about optimizing anything. They're always worried about just getting it off the ground as fast as they can before it drops back below. Yep. And then prices go negative or they get super low and then everybody's worried about optimization again. And it's just the same cycle over and over. 00;34;56;06 - 00;35;16;29 Unknown Right. It's it's painful but it opens this door for so much opportunity, especially just around that like, hey, this is a workflow that happens across multiple silos in the industry, multiple data sets, multiple databases, whatever. But look at the power that can be had by unifying these things and leveraging up like you're talking about. So let's get into that. 00;35;16;29 - 00;35;31;25 Unknown How what's what's kind of in y'all's tech stack. How do you guys, kind of do that? What's the what are some of those big, you know, value out of things that come out of that? I can only imagine my first my first job in the oil field was as a production intern, running around with the pumper and doing kinds of random stuff. 00;35;31;25 - 00;35;50;29 Unknown And so it it always blew my mind that the traditional way for like the, the field side of the oil field to work at least around production was the pumper has his scheduled route and he just goes out there and checks on that. Right? Right. And it's like, what? You know, yeah, I can watch my front door from my phone. 00;35;50;29 - 00;36;11;24 Unknown Yeah, but this guy is just driving aimlessly on the same route, regardless of what is actually happening to all the equipment and stuff every week or every month or whatever it may be. You know, it's like it's insane. Yes, it is insane. yeah. So right now, I mean, where, obviously taking those workflows and, and optimizing that for what should be done on a daily basis. 00;36;11;24 - 00;36;32;06 Unknown Right. And taking out all the noise, focusing on what is actionable and focusing on the type of action, type action can be remote, type of action can be on site, type of action could be, troubleshooting type of action can be, you know, many various things. It could be different workflows. And I think, you know, where, the workflows are, are challenging, right. 00;36;32;06 - 00;36;57;13 Unknown Because when you really think about, like you're talking about, all these systems being very siloed and this is the opportunity 100%. and so how do you really connect that those data pools together? The models, but also people who need to do different things because the the measurement tech or the measurement analyst could very much be need to be, you know, it could meet a calibration or it could be at scale. 00;36;57;13 - 00;37;16;02 Unknown Right. If we're picking up meter calibration as an issue. Right. Well that's not you know, going to the field technician. Right. Is going to this other user that's a different workflow. And then if he has to order parts, well that's you know, that's a different complexity. Right. And so that's that's the hairiness that we're dealing with. but at the same time there's massive opportunity. 00;37;16;02 - 00;37;35;18 Unknown But but your question sorry I was on the tech stack side. Yeah. What what's in your stack? How are you. How do you kind of architect this? Yes. Any fun lessons learned or fun lessons learned? I'm sure there's lots of those. Yeah. Cool things that you would recommend or not recommend doing. Yeah. I mean, I think again, we're dealing with a lot of time series data. 00;37;35;21 - 00;38;01;22 Unknown and we're doing. Yeah. Okay. Good quality. you know, tubing pressure is is labeled, you know, at least 20 plus different things. yeah. TB seven and you know, T 19. Well let's t 19 excess tubing pressure. Yeah. so yeah a lot of it is in the driver and aggregating that data and cleaning the data and prepping that for a model. 00;38;01;24 - 00;38;20;08 Unknown I mean, I think a lot of our, our issues have been on the managed management of data. Right. And are we trying to manage that? Are we paying for, different components to manage that as well or different databases to, to manage that. So again, it goes back to you know, what schema. How are we generating IDs? 00;38;20;08 - 00;38;45;24 Unknown Harry? Generating unique IDs. Whoa, man, if we could just push that off to somebody else. That's a managed where we can focus on, you know, really, for us, the the things that we're really focus on is how do you label data at scale and produce models at scale. And I think that's the unlock that we continue to, edge closer to and is really exciting. 00;38;45;27 - 00;39;08;15 Unknown there's different ways to do that, right? There's, there's a we have, you know, millions and millions of labeled data. Right. So what do you even do with the labeled data? Right. how can you turn that into models at scale when, you know, the next client uploads their whole data list? We have 15 models that we have generated for the client based off of their data. 00;39;08;16 - 00;39;28;25 Unknown Yeah. we're getting to that point now because we're we've stopped fucking around with our database. Yeah. And so that's freed up time for us to focus and iterate on the right things. Yeah. which is all the management of, you know, how are we going to do IDs, how, you know, what's your what's your equipment store, what's your model store, what's your data database? 00;39;28;25 - 00;39;49;27 Unknown What's your, you know, your edit edited database. You know, how are you changing your workflows, how you updating that data. And so all of that flow, we were trying to I say managed in the beginning. Right. And we've found some better solutions to manage that for us. And we can really focus on the next pieces of it. 00;39;49;27 - 00;40;06;28 Unknown Yeah. No, that's like a perfect example of what we were talking about earlier. Right. Like doing that for a posse or it's like, okay, well, someone can just go manually relabel or like map these headers or come up with a, a, a key that maps things over and stuff. And it's like, but that doesn't scale once we get to a certain point. 00;40;07;00 - 00;40;27;05 Unknown So it's like, okay, now we're going to use models because we have enough data to actually train a model to do these things. And those are the things that, like, the average person just has no clue about. Right? Like, yeah, scaling to one customer versus being able to quickly scale to dozens. Yeah. Especially with I'm sure the variety of data sets that you get toy is is a whole nother ballgame. 00;40;27;06 - 00;40;48;12 Unknown And you have data that's organized in many different ways. Right. you've got data, you've got obviously historical data, you know, like customer and you've got, you know, classification, models that you're currently running on. how do you backfill some of those if you have a new classification model that you want to to test against? Backfilling is like the yeah. 00;40;48;15 - 00;41;15;20 Unknown Every data engineers exist. That's it. Yeah. That's and that's the one example out of you know. Oh yeah. Like like 50 right. but how can you do all that stuff, from your data management and how can that all be seamless where you're not screwing around with that? and I think the, the interesting things that we're focusing on right now is, yeah, I think there's different techniques for labeling data that I think can be really, really impactful to build these future models. 00;41;15;20 - 00;41;41;10 Unknown And so that's what we're we're currently, focusing on for a release over the next, months. Nice to, to do that. So you could search for any type of issue that can automatically pop up and create your own workflow from that, within operations. Let's unpack that, because I was having this conversation on earlier with or I think earlier this week with somebody else, but the, yeah, I think a lot of people, even myself included. 00;41;41;10 - 00;42;07;28 Unknown Right. Like the way that you approach you have to approach data at scale changes, especially with all the tech tools that have come out in the last few years. And stuff. But like, yeah, traditionally, if you're building a relational database or whatever, you have to come up with a schema and you have to come up with all the semantic things and variables and naming conventions and all this stuff, and it's like, but the best way to do that is actually what the data tell you. 00;42;07;28 - 00;42;27;15 Unknown Yes. Right. So that's a completely it's literally the opposite of right. And so but when you do that it's way better. It's way more flexible. It's way more accurate. All of these things. Yeah. And so can you kind of just talk about that in general. Yeah. How that has to go on. Any improvements and stuff you guys have seen because of that. 00;42;27;15 - 00;42;46;08 Unknown Yeah man I think you're spot on. Right. the, the engineer in the beginning is providing that structure. And then you find a billion cases. Right. And it's and it's always a fringe case and it's like, well, fringe case. Well it didn't fit, you know, the schema. So we have to think about how we, you know, well should we create another table. 00;42;46;08 - 00;43;05;02 Unknown Okay. Well how does you know that link to. Well we have all these workflows and well we have to run a different process. Right. And so the you know, you've just taken one fringe case that has blown up one customer for one customer, right. That has completely just like blown up your whole, time on, on a Friday because you have to re architect re engineer. 00;43;05;03 - 00;43;25;15 Unknown It's always on a Friday. Yeah, yeah, yeah, yeah. And so, you know, I think you, you're spot on in that where it was human driven before in creating a somewhat of a, limitation there where as now you can look at the data and say, hey, can you provide me a schema or what are the best outputs based off of this data, based off of all this data that you have at scale? 00;43;25;15 - 00;43;48;04 Unknown Yeah. For recommending the even data labeling. Right. This is where we're at right now in our technology lives. And it's so exciting because you can let the machine, you can let the model, you can let you know any different form of AI do it for you. And this is really about how do you leverage AI in different models and different ways to optimize the whole workflow. 00;43;48;06 - 00;44;08;15 Unknown It's not just that AI is a model and how do we use it in this app to tell me all the answers today? It is embedded into how you develop, right and how you run. You know, all your different facets of operation of structures and data structures, of labeling all of these things. What makes you so much more dynamic too? 00;44;08;20 - 00;44;24;22 Unknown Yes, when you get those outlier cases where you get a new customer or a new data set that you're not accustomed to or like it wouldn't have fit in your old schema because it is new or different. Yeah, well, I don't have to worry about that. Yeah, yeah. Like I'll have to rebuild this or. Yes, figure out all these relationships, this other shit. 00;44;24;25 - 00;44;41;09 Unknown Now I think that's one of them. And it's always like, where do you, where do you put the injection point there. Right. If if you're going to generalize what your, what data types you need. Right. And how to you know, it's somewhat not not, you know so generalize. Right. But it's not so specific where you're finding the schema right. 00;44;41;09 - 00;45;02;05 Unknown There's a middle ground there that I think, is a better sweet spot to operate in. Yeah. No, I think that's, I think Diana asked me about what my thoughts are on on AI and how people are going to use that and all of that stuff. And that's one of the biggest areas that I think is being so slept on is, yes, obviously it's, you know, worker augmentation, acceleration. 00;45;02;05 - 00;45;21;11 Unknown Sure. Whatever. Yeah. But there is so much I mean, even like on a weekly basis, there is new stuff that comes out on like, oh, hey, now we have this way to map the into this entire PDF document and all the shit that's in it and the images and labeled them and like, yes. And like that would have taken a human so long to do at scale. 00;45;21;13 - 00;45;37;21 Unknown And that's what they've been trying to do with OCR, for example, ever. And it's like, but now you can see, I, I got a bunch of metadata on a lot of our sources for our language model by OCR in the first few pages of the content. So eating it into GPT and saying, give me the author title and all this stuff. 00;45;37;21 - 00;45;55;02 Unknown Absolutely. And it works really well, really, really well. And it was pretty damn instant compared to me doing it this time, doing 100%. No. Yeah. Or like we would have when we had like big data problems, we would distribute that across the company and be like, hey, we just need everybody to go pick. Yeah, here's your 20. You go, yeah, we'll knock these out. 00;45;55;02 - 00;46;20;13 Unknown Right. Like like, no, I just feed it into a model now and it's done. And it's crazy. Yes. And I think that is a fantastic point 100% agree. And what it tells me is that it's just about pace of iteration. Right. And everything is pace of iteration. Right. If you if you can take the, you know, tomorrow's problem and solve it within 30 minutes, where four months ago it took you a week. 00;46;20;16 - 00;46;42;11 Unknown That's what we're after, right. Because pace of iteration for all of these problems is how quick can you solve a problem? How quick can you scale it up to market? How quick can you take these problems with the market in a more scalable fashion? Right. With limited resources, with limited resources. So, you know, I think there's a, you know, there's there's a huge amount of value that being can be created by doing this. 00;46;42;11 - 00;47;02;18 Unknown Right? Yeah. And just to your point, the inner workings are how do you put that all into your development cycles. How do you how do you take all this time that you're spending doing this manual effort, doing all the, you know, it's it's manual effort that humans hate doing there they get I mean, humans are flawed. We make mistakes. 00;47;02;24 - 00;47;17;15 Unknown And we also even in that example that I gave where it's like, if I would have given it out to the company and had them labeled departments on jobs or something, right? Like, yes, each person would pick different things, different ways where it's like if a model does it, it's going to do it the same way every time. 00;47;17;15 - 00;47;32;05 Unknown Yeah, at least I have that consistency. Yeah. Like as an engineer. That's one. Yeah. I would much rather have a model doing it than a. Yeah, you're not consistent right. Like the human performance is going to degrade over time just like everything else. And that's guess what. That's what humans or technology is for. That's what computers are good at. 00;47;32;12 - 00;47;55;28 Unknown It's all the mundane shit that humans hate doing. Yes. And aren't get lazy about or you know, make errors with. Right. And so it's like, well, I think a 100% spot on. And in the normal day to day work, right, there's just so much B.S. that everybody has to deal with at the enterprise and within AMP, and you've got your reports and then well, but you need to update your slides for Tuesday and then but you've got to fill out, you've got to check your downtime because everybody filled it out wrong. 00;47;55;28 - 00;48;21;24 Unknown Right. So the whole construct of what everybody's doing on a day to day basis is packed with meetings and packed with work based off of those meetings and packed off of reporting to feed on up, because everything's messy. Right? And so the value, you know, everybody thinks that, well, AI is going to tell me everything to do is less, you know, I would say, and this is probably something that I've definitely changed on in the beginning, I was like, hey, we can procedure like this and say, hey, this is step one. 00;48;21;24 - 00;48;35;27 Unknown This is step two, and so on and so on. But the value is just and taking out all this B.S. work, right, where people can actually think, have the time to iterate on the next thing and go from there and that's where the value is, right? Let the human do the things that the humans are good at. Yes. 00;48;35;27 - 00;48;55;08 Unknown Of that. You can't teach a computer. Yes, and let the computer do the boring, mundane shit that we don't like doing any better. That shit is like 80% right? Absolutely right. Like it's it's just way too much. Yeah. No, that's, in the digitization age, right where you stand up your first forms and you stand up your first environments and all this stuff. 00;48;55;08 - 00;49;21;01 Unknown Right? We're now trying to redo that. Yep. Almost in a sense. Right. Okay. This data gave us great some good clarity on to optimizing certain things. That's good. Right. That's progress. But it's like now you're spending your whole time managing this and you're kind of at a plateau. Right. And how do you break through to that plateau where you're not managing this initial digitization effort that you stood up because it's not intelligent? 00;49;21;01 - 00;49;43;05 Unknown Yeah. So I think that's that's where we're at right now. Right. That's awesome. No, that's I love I love hearing that I love, I love I know I just I love this period of time. Yeah. Industry because there's so much one there's so much cool shit happening in the industry as a whole. But then to overlay that with all the tech stuff that's happening, it's I'm very bullish on on our industry moving forward because of all these things. 00;49;43;05 - 00;50;06;10 Unknown But yeah. And now that's, well, so since we're talking about AI, everybody wants to talk about AI. Yeah. Yeah. Sorry. It naturally goes. No, I mean absolutely. It's it's well, I mean, like we said, it's really useful in a lot of different ways. And I think that's where there's so many creative ways that we're not even talking about or thinking of that people are going to be using these things to accomplish all kinds of stuff. 00;50;06;10 - 00;50;35;29 Unknown But absolutely. How do you, how do you see I kind of or let me let me refine that. How do you see kind of the language model trend boom, etc., impacting the industry or people in general? Yeah, absolutely. It's impacting, you know, as we were talking before, right. The amount of, progress with LMS is just I mean, the pace is just absolutely crazy, right? 00;50;36;01 - 00;50;59;16 Unknown I think if you look on, you know, the models that have had the best scale over time has, you know, where we're in the stage where LMS have absolutely taken over. and I kind of come at that from a perspective of even just think about time series data. Right. So if I could, if I could, you know, paint a picture black and white of time series data and the progress of LMS, right. 00;50;59;16 - 00;51;18;29 Unknown Like LMS with, you know, post Bert. breakthrough. Right. And, and the initial version of GPT, you know, you were kind of taking those steps and then you take this just drastic step, right step change. And the step change is massive. And you haven't seen, you know, I think envision there's there I would say you're you're pretty far. 00;51;18;29 - 00;51;39;09 Unknown You know you're further progressed as well but not as scalable as as a LMS have taken off. I think the question and the interesting thing is, how do you use the best progression of the models that are right now? and what are you going to bet on progressing. Right. And this is a technology issue about again what are you going to pick for your database. 00;51;39;09 - 00;51;59;01 Unknown Right. What are you going to pick. That is we're not going to to have a breakthrough here. Right. We're going to bet on XYZ company having the breakthrough. and I think it's absolutely going to be used in many different ways. And I think we, you know, when in the past, I don't know, two years. Right. And two years ago to a year ago, of course, everything was chat bot right. 00;51;59;01 - 00;52;17;11 Unknown And everything is, hey, chat with this and chat with that. Right. And you can chat and all these different ways. And I think from our perspective, we saw that as you know, that's that's it's definitely value add. You can upload your documents, you can upload your data and you can have a conversation. And the value there is that the thing, you know, data is at your fingertips more readily available. 00;52;17;11 - 00;52;43;15 Unknown There is value. Absolutely. for us, I think we always try to think about how are we going to have a differentiated solution with us. And so, we really focus on data entry and recommendation and, and auto filling forms and live search new data, all from a singular experience. Or it's working back from the end user and all the workflows at the end user has to do, and making that just one right instead of ten. 00;52;43;17 - 00;53;03;18 Unknown You mean manual duplicate data entry isn't on the top of the list for all of you guys favorite things? Yeah yeah yeah yeah. Exactly. Right. Yeah, exactly. Yeah. I already told you I'm on this. Well, actually, you know, I'm on this email because my phone is here and you know exactly what pad I'm on. And, you know, when I got here and, you know, like, yeah, no, I so so it's we're using all that again going back to the 80% BBS. 00;53;03;18 - 00;53;23;03 Unknown It's happening in operation from the initial digitization one. Those are all opportunities for the human to screw it up too. Right. Like anytime you can have a database auto populate something that you know is right versus having a human input it, do it. So I get to I hate got the the amount of, you know, forms and manual data entry stuff with all the ERP, CRM. 00;53;23;03 - 00;53;38;10 Unknown Something I've set up is just like it's it's bad. What? You just don't know until you start getting into it and you're like, okay, well name? Why? Why would the well name be anything other than the well name? Well, you call, you know, it's like you, you call the superintendent and he's like, yeah, we're on the up 22. 00;53;38;12 - 00;53;56;24 Unknown Okay, great. Then you go look up, you all 22 and you can't find you all 22. And it's like, oh well it's the university lands giant, you know, string and like, okay, well there's a bunch of university lands number 22 with different preferences before that. And so it's like, well which one is it. Right. And you're just so yeah, even I laughed earlier when you were talking about, you know well identifiers. 00;53;56;24 - 00;54;14;08 Unknown Right. Because we've had Bobby and and Jeff and the, the whiskey data night stuff that we've done. That's always a topic of, of debate. So what is a unique identifier? A, well, unique identifier. How do you do that? What does that based off of? Right. And like I had never truly thought through that from especially from an operators perspective. 00;54;14;08 - 00;54;33;19 Unknown Right. And like that's a huge thing that basically arms right. Like you know like how again how you doing it that right. Like companies are still spending projects and months and still trying to figure that out. Right. No effort. But I think I think alums yeah. I mean it's we were just you're just talking about document extraction, right. 00;54;33;19 - 00;55;06;18 Unknown And then schema, you know, construction based off of those documents. Right. And so this is a way where, yeah, it's not a chat bot. Right. But it's it's picking your underlying structure of how you're going to develop your technology. and I think the just more areas that that's going to be applied to, right, whether it is models, whether it is retraining, whether it is data labels, whether it is summarizing, whether it is engaging with somebody, whether it is, you know, an agent, you know, generic workflows and, and multi-model, is there, all of this is coming together. 00;55;06;18 - 00;55;29;11 Unknown Right. And so I think that because obviously there's there's very, you know, there's massive breakthroughs that have happened here. And so you cannot discount that that that will proliferate in different ways because that has had a massive breakthrough. And we all need to recognize that. And and then I think it's it's a race to be to say, what am I going to be really good at. 00;55;29;11 - 00;55;53;24 Unknown Right, right. And what can I scale here that is differentiated? and that can scale across, you know, the landscape. Yeah. Integration. Yeah. Integration. Right. Talk about integration. Right. There's a better means of doing integration. And LMS are nicely suited for that. Yeah. The right really nicely suited for that. And so how can that be done in a day in an hour versus yeah. 00;55;53;24 - 00;56;09;24 Unknown Can you send me the tag list okay. Let me look at it. Well let me tell you. And then you can resend me this and, and all of that workflow right is going to be condensed. And so you have this pain in the ass thing that you have to do now with other LMS. It's condensed. Yeah. Yeah. No. 00;56;09;24 - 00;56;30;05 Unknown And that's the thing like people sleep on the value of just this new augmented search retrieval mechanism that are largely image models. And yes, you know, that's, I, I tell people all the time, like, they get, you know, everybody's worried about AGI and them being sentient and stuff. that's not what they are right now. Right? Like, who the fuck knows what's going on? 00;56;30;07 - 00;57;01;08 Unknown Yeah, what that looks like. But right now, they're just a better way to discover information that is already in there, right? Yeah. It's not it's not thinking on its own. Yeah, yeah. And I wouldn't even put it like I mean, sure, I think there's aspects of that and everything else. You still have to have people, especially in on gas people doing work on assets that's, you know, and probably say there's, you know, some people who say, well, I want the autonomous well, site where, you know, nobody has to touch anything. 00;57;01;08 - 00;57;21;21 Unknown And it's like, you know, look, I understand where you want to get to. And yes, I think everybody's in line with that. but there's so much other low hanging fruit, right? There's so much other, fluff. And I would say most, you know, a lot of organizations because you're dealing with this shitty data management and so let's try to attack that. 00;57;21;24 - 00;57;42;02 Unknown at the same time, these things can be in parallel process. Right? But I think the worker is still going to be doing work on these assets. Right. And I think that's definitely, going to be the case now, all the, you know, talking to regulatory. Well, did you follow the report, what can you assign Jim. Well, the wall was down, but it needs to be filed for all the things. 00;57;42;02 - 00;58;00;23 Unknown Right. That that's all the fluff that I think can be taken out of the system. And so I think what the future holds is that, you know, the office reduces over time. Now is anybody's job. I mean, yes, same conversations with machine learning and with models and everything else that happen. And, you know, if an organization wants to cut, they'll find a way to cut. 00;58;00;24 - 00;58;21;16 Unknown Yeah, right. They can use data as a crutch to make that argument. Or they can do it right. Like it's true. You know, it's very it's it's in both both happen. Yeah. And so, everybody's fear of AGI taking over and making decisions and all that. Yes. I mean that, that will it will come maybe over time. 00;58;21;16 - 00;58;40;18 Unknown But I think there's, there's so much to happen right? Yeah. That is not that. That's the thing. I think one of the interesting things about the machine learning AI into the language model stuff has really forced people to understand that, like data data quality. Yeah. We have a shitload of data. Most of it is shit. Yeah. Like at the end of the day. 00;58;40;19 - 00;58;59;20 Unknown Sure. From a data data engineering. Sure. Right. Yeah. And so it's like they tried all these things and then they realized oh well the data is bad. So we've got to take that step back. And like I got to be baby step our way into this. We can't just. Yeah. And it's the technology hype curve right. Same thing I think we're peaking with LMS right now. 00;58;59;20 - 00;59;17;27 Unknown Sure. I say not to say that they're bad or not use share. But yeah, again, when people are talking about oh they've got AGI or like, yeah, hey how much. Yeah. How I feel about hype. Right. Yeah. Like yeah. That's a great. Okay. You're selling the future state, right. Which but all good startups do. But yeah. 00;59;17;29 - 00;59;40;04 Unknown you know, I think it's, it's going to be wild to see how all this stuff changes. Yeah. Because it's just moving so fast. I was running Lambda three on my laptop earlier. Yeah, yeah. Oh, yeah. This stuff is so lightweight. Local model is super cool. That's where. That's where I think the, the I'm seeing a trend, at least on the newer models, is everything is pushing towards speed and efficiency. 00;59;40;12 - 01;00;04;02 Unknown Right. And I think all of that ultimately you know Apple just put out there as a Microsoft came out A53 and all the same. But I think they're all really the goal of that is they're going to run on the on your they're going to run on. Yeah. On the edge. Yeah. Yeah. And that's another thing people don't think about is it's like I tried to when llama two came out on GCP in vertex, I was like, oh, I'll go sign this up. 01;00;04;05 - 01;00;21;01 Unknown My estimated cost per month for that one llama instance was going to be 30 grand. I was like, well, there's a reason why they're building these models to be smaller, cheaper, more efficient, because they don't want to run on your hardware and not on their their servers just because it's expensive as shit. Yeah. And then it's like, oh yeah, if you're trying to deploy them, you're like, okay, now I get it. 01;00;21;07 - 01;00;40;18 Unknown Yeah. There's like I can't, I can't. Yeah, keep a server running for 30 grand when I could just go buy one for 30 grand. And that's a very tricky kind of balance there. Totally. It's, Yeah, it's going to I'm very curious to see how that kind of evolves because, I mean, having it embedded on all your devices is what is going to happen at some point time wild. 01;00;40;23 - 01;01;10;04 Unknown Yeah. Yeah, I think it goes, you know, it's it's another avenue where, it will infiltrate that median. Right. and again, you know, how much value having something on the edge and how much value is, you know, application or having it this way. Right. Or having it pointed towards Z. Right. I think that's, you know, again, who can who can do this at scale, right? 01;01;10;04 - 01;01;32;06 Unknown In the right way? I think that's the the interesting question. And there's many different ways to cut that. Yeah. And I think just observing that right is, is really interesting. Yeah. That's kind of a cool point. Cool part now is the industry. Of course, we're always yeah. the first ones to, to copy. Right. Or the second ones to adopt essentially. 01;01;32;06 - 01;01;52;03 Unknown But everybody's been kind of like watching, waiting to see. Yeah. Like we're not going to commit to one. We're going to test out a bunch of stuff. I think it's really like you've got to have I think the answer is the architecture has to be set up in a way that you can swap models, because that's the there's going to be a better one at some point in the future. 01;01;52;03 - 01;02;09;10 Unknown That may not be the one that you chose. And then what do you do if like, that's what I don't get. There's so many companies out there building custom models and stuff. It's like it's going back to that very this conversation, like why are you building instead of just buying, right, and being able to architect it in a way that you can hot swap with an open source thing, or just building with open source? 01;02;09;12 - 01;02;31;09 Unknown there's plenty of those out there too, right? Right. Yeah. I mean, it's tricky. So many aspects of this, like, like even just, you know, there's a trigger where where you said, you know, we're just going to swap model A versus model B, right? I haven't really seen that been done. well at all or done at all. 01;02;31;10 - 01;02;58;11 Unknown Yeah. You know, and so, you know, I think from from a generalize perspective, it's like, well, yeah, we want an architecture where you, you know, but the reality of the situation is, you know, they don't they they've got their messy data or you know, they have all these other compounding issues that they're trying to figure out. And there's so far away from swapping models because they haven't, you know, your foundation layer there is not set. 01;02;58;13 - 01;03;18;01 Unknown And you got to crawl before you can run. Yeah, totally. And and again, it's it's like, how is this all going to work together with what data do you have out, what models you're going to spin up off of that. Well what are you going to try to do. What's the end workflow of that. Right. Like thinking about this, this from, you know, systematic level I think is, is really important. 01;03;18;01 - 01;03;33;20 Unknown Yeah. And I think that people are just looking at it from one approach or everybody is talking about a model is that, well, you know, I try to say, you know, hey, what is my production trend for the month? But it didn't work. Right. Well, of course, that's, you know, a data issue, a model, a corpus issue, like all these things. 01;03;33;20 - 01;03;49;14 Unknown Right. How did you, what embeddings did you. Yeah, there's some things, actually. There's so much in there. At the same time, it's like, well, why are you trying to to show what you, what you produced for a week. What's the value there. Yeah. Right. So I think it's is both, it's in the Spotfire dashboard just. Yeah. 01;03;49;14 - 01;04;08;11 Unknown Yeah. For me. Right. Like it's a dashboard. Yeah. That you know but there's a thousand of them right. But it's like but it's two fold right. It's what is the value of what you're trying to achieve. And it's also like there's a lot of intricacies in just getting that set up from, you know, and not even talking about swapping models. 01;04;08;15 - 01;04;32;24 Unknown You know, it's like, can you get that to a point where like, like what are you trying to achieve with just your data pipeline for that? That's why that's why we're pretty heavily betting on the rags rags side of things. Because especially for what we're talking about, which is just search. Yep. Essentially. Yep. Retrieval. That at least allows you to swap the models out because your vector database is your vector database. 01;04;32;24 - 01;04;55;23 Unknown It doesn't matter which model is calling it, as long as you can get that part of it as good as you possibly can. Yes. Then you can find which off the shelf model is best at. Yes. Retrieving toy. But yeah, that's, that's a very unique instance. No. If most people are not it's not a yeah, they're not building on a right architecture or they're we're talking about machine learning and different things where it's, it's not nearly as easy to do. 01;04;55;23 - 01;05;16;15 Unknown Yes, yes. Right. No. Absolutely. And even with retrieval. Right. There's many different ways to approach it. and to have success with it. So yeah. Yeah, the vectorization is super, super interesting. and I think loads so much power. Yeah. And even just choosing the right architecture about, you know, and there's a million databases that do vectorization. Right. 01;05;16;15 - 01;05;36;16 Unknown So pick your flavor. start with pinecone. Right. Go from there. Yeah. but but then that is a question of now that's, that's a very first thing. Right. Vectorization. And then well, your other data is going to sit where in it. Well it's not going to be in pinecone. It's going to be, you know, over here and here and well should we what's the, the system that can manage this all. 01;05;36;16 - 01;05;58;10 Unknown And so then you're going back to architecture. And what systems are you picking there. Yep. So yeah all these things are very important but you know it. You you're you're approaching this from an investor standpoint. What tool or what software system is going to to continue to grow in this way, where it's going to you know, encapsulate these future issues that we're thinking about doing? 01;05;58;12 - 01;06;15;07 Unknown And so you kind of you have to approach it from investor perspective as well. How much money have they raised? Are they capital efficient? Are they doing the right things? Do they have an amazing product? You have an eye for product and eye for detail team. the team. Right. And so you as you're picking software, you have to approach it. 01;06;15;08 - 01;06;32;21 Unknown You know, even if you're building as a software company or within an app, you have to approach it from, you know, hey, is this is this team going to be doing amazing things or where are they going? And am I going to bet on them going in this direction the right way? And if that bet doesn't come true, then you're reframing everything, right? 01;06;32;21 - 01;06;47;11 Unknown Then you're pulling everything out again and you're trying to restructure everything. Right? Or was there, was there get repo updated in the last six months? Yeah, right. Like that's a good sign. People don't check that you're enough. Yeah, yeah. It's like, oh I found this cool git library. I was like, well it's four years old. Yeah. And there and there's no commits. 01;06;47;11 - 01;07;11;12 Unknown There's a bunch of outstanding issues, no new commits. so I'm not going to build on. Yeah, yeah, yeah. Exactly, exactly. Yeah. I found it via Twitter thread. Right. Yeah, I got fed. I get so much shit on my, my Instagram algorithm feeds me all this wonderful stuff. Yeah. And so it's, Yeah, it's always fun though. Then to go be like, okay, well, this guy clearly got paid to make a a real about this or has an affiliate or whatever, right. 01;07;11;12 - 01;07;27;25 Unknown Like I need to actually go try this and see if it actually is what it says. It is. Totally. But man, we're already already past time. So that was quick. Our ten minutes man goes by quick. Well, I'm going to talk about this stuff all day as as I know you can as well. So I enjoy the conversation. 01;07;27;25 - 01;07;57;17 Unknown Yeah I appreciate you coming. We, we wrap it up at the at the end, but I wanted, one just be give you some give some props to another to you alum. Go Golden hurricane. Oh yeah. Bring it back. Yeah. Bring it back. And, but so at the end, we kind of just do like a speed around and I'll pepper you with some random stuff, but, favorite either. 01;07;57;19 - 01;08;20;27 Unknown Yeah. Favorite kind of social follow. Pick your platform. Yeah. you know, we're on LinkedIn or on Twitter. and, have not been posting, have just been, you know, had head down building. I would say that's my default. Yep. but where we're kind of coming up with the marketing campaign and everything else, post this for new releases. 01;08;21;04 - 01;08;43;10 Unknown but yeah, Twitter, Twitter and LinkedIn. Main source. Any specific people on there that you think others should follow in the tech or energy space? oh. I know that's always a hard one because it's, off the top. Any specific people to follow? I would say for me, I don't have a person in mind specifically. 01;08;43;12 - 01;09;10;19 Unknown I'm definitely, following a lot of the companies that are doing different aspects, of AI right now, and using it in different ways. And so really kind of seeing and it's really interesting. Right? You have the foundation companies, of course, and you have the retrieval, you know, specific, packages and different startups there. You have the vector databases and, and, you know, how is this being used across the board? 01;09;10;22 - 01;09;29;29 Unknown and who's going for that? The, the vision, the la, the, the PDFs, the enterprise data, who's going after, you know, different segments of the market. And so I think that's something that I pray, or stay, pretty close to just on seeing where it's going. But at the same time, as we talked about before, you can't get caught up in it. 01;09;29;29 - 01;09;53;14 Unknown You just have to go head down and build and build off of your end goal for what's going to drive value. specific to the user. And so that's really the focus. What about books, either startup or tech books? I wrote I read a lot of books, before starting task. And since every time since starting task is more so been, been less about that. 01;09;53;15 - 01;10;12;02 Unknown Yeah. but yeah, I think, you know, learning mindset by Carol Dweck is, growth mindset is, is a key, a key one. Right. That's really kind of reframing just how you think about life and how you think about success in your company, how you think about and everything is literally everything it is. And so I'm going to make my kids read that like high school. 01;10;12;05 - 01;10;33;13 Unknown Yeah. Because it it's so genuinely think that can literally change your life the way that you perceive and approach things. Yeah. And it's, you know, are you gonna think in a box and are you going to think that success is in that box? And it's super interesting, right. And how you reframe your mind and reframe your mind about how you're thinking about failure outside of the box and, and that it's not even failure, right? 01;10;33;18 - 01;11;00;28 Unknown Is such a shitty word to use. It's iteration. Yep. And when are you going to iterate or are you going to be confined? Yeah. It also to me helped really like explain a lot of the business side of things to. Right. Or it's like, you know, they talk about okay, well you know certain companies chase the quarterly earnings calls because they want their stock to perform well on chasing quarterly earnings calls typically is not a great long term strategy for a company to continue to exist. 01;11;00;28 - 01;11;18;15 Unknown So then you look at the growth mindset, companies like the apples of the world. And you know those folks and it's like, oh, that's a very different company, right? Like the long term game. Yeah. The infinite game of this is very different than yeah, chasing the quarterly time EPs. Right. Yeah. I think the same thing happens in startups. 01;11;18;17 - 01;11;34;10 Unknown They go into our initial point because you have to build this initial product. Right. And so you're furiously building this and somewhat of that is short term. Yeah. Right. But at the same time, there's so many aspects of, you know, you want a good client referral and you want it to grow in the right way. And so you someone have to do this. 01;11;34;10 - 01;11;57;13 Unknown And so there's always this balance. But the point having a long term vision that you're building around is the key most important thing for startups at the early stage public companies. and it's the most healthy way to continue to grow. Right? Because you're growing with an aligned state. Yeah. And that's very clear to everybody. Right? So yeah, getting everybody rowing in the same direction is very understated. 01;11;57;13 - 01;12;17;25 Unknown Yes, yes. I mean, yeah, no hundred percent I mean just in how you build right up like, you know, 10% misalignment drastically slows things down. And so focus on 99% alignment. Right. And and you're not nobody's going to be totally aligned. But that's how you move quick right. That's how you iterate. Well minimize minimize the places where they can't be right. 01;12;17;25 - 01;12;35;16 Unknown Like I've I mean, one of the things I do now, I almost require it, but if it's a if there's a feature request or a bug that's visual, it needs a mockup or a screenshot or whatever. It needs the visual. Yep. Because it's going to go to a developer. Yeah. And the developer may or may not have any idea what we're talking about for that context. 01;12;35;18 - 01;12;53;05 Unknown And so guess what. If it's a mockup they can't screw that up. Yep. Right. I can't tell you how many times I've put in feature requests or new features, bugs, whatever, just with text. And then the developer checks it off because it the feature exists and it works. But yeah, is not intuitive or useful or make any sense in any way. 01;12;53;05 - 01;13;24;12 Unknown But they check the box, right? Yep, yep. the last one. What's, what's all the tumor? favorite video or board game? favorite video? video game or board game? my brother is an expert, board game player. And so he, he knows everything about board games. And so he really, leads me in that capacity. 01;13;24;14 - 01;13;44;10 Unknown but I haven't played, board games in a while, too. Like, my daughters are just growing up, and so they're seven and five, and so I would definitely want to get them into board games. but we're just reaching the point where, yeah, you start playing some, some more interesting ones than, Candyland. Yeah. we, we actually I'm currently stuck on Candyland. 01;13;44;10 - 01;14;03;18 Unknown I'm hoping to move to monopoly as a secondary and then go into strategy game. We got, we got some kind of I think it's Bluey Appley. Yeah it's Bluey. Yeah. And it, it's pretty good but it's just I mean it's still monopoly and still takes forever best their children. Yeah. And then you're like oh man I didn't realize you could be like your income. 01;14;03;18 - 01;14;20;16 Unknown My kids are the same age five and seven. And it's like, oh, I didn't know you have the capacity to to, come in and just dominate an entire section of the, of the board. Yeah. My son, my younger one, he'll just buy them. And every time he lands on the spot. Right. And my daughter will wait until she gets all the way around to the last two edges. 01;14;20;20 - 01;14;47;25 Unknown Right. Oh, it's like picking up on this. Yeah. That's good, I like it. Oh, yeah. Then you play Chutes and Ladders and it's just pure randomness and you always end up losing to the person leading. Always ends up. Yeah. I hate this game. Yeah, yeah. Last one. What? what's some advice you would give to somebody that's, you know, looking at starting their own software energy tech company that you wish you knew or had when you, man, great question. 01;14;47;27 - 01;15;03;16 Unknown there's definitely few things I would say. Go for it, number one. You know, go for it. Life is too life is too short, man. Life is too short to hold back on doing that. I think going back to growth mindset, you learned so much. I learned so much about the business. I learned so much about engaging with clients. 01;15;03;16 - 01;15;22;09 Unknown I learned so much about how to build the right tech. You know, all these different facets that I didn't know about before. Yeah, right. So taking that plunge, I think is if you're curious about it, it's an such an exciting time to be building. and it's okay not to know this. Like, yeah, everyone doesn't know those things when they don't know everything. 01;15;22;09 - 01;15;42;07 Unknown When you start, that's a part of the journey. Just do it right. Reach out to, you know, 100 people via code message. Try to say, you know, say the other thing, right? When I was starting out, everybody's supportive of somebody taking a risk to go do something. And so you have that with you and go engage people. Go send DMs. 01;15;42;07 - 01;16;00;03 Unknown Everybody's going to help you out. Everybody's going to give you the time. I mean, there's so many people in the beginning that gave me time when when we were like four months and had like nothing, right. And, you know, Chris. Right. And and all these people. Right. They were really kind, to, to give their time and just to listen to us. 01;16;00;03 - 01;16;25;06 Unknown Right. And so know that you have this inertia when you start out that can be contagious. Yeah. And it's and it's very much in a positive way. Life is too short. So just go for it. the things that I wish I would have done to, you know, I think for in the beginning, we, we were somewhat, too confident that we could run this as a completely bootstrapped company with no capital. 01;16;25;09 - 01;16;50;19 Unknown And so we underestimated what it took to really get into the market, what it took to start scaling, what it took to build this, you know, differentiated AI and, yes, this technology stack that, that, you know, we believe is, is completely going to have an unlock. Yeah. and so, you know, you kind of you don't know what you don't know in the beginning and you kind of have to reframe and consistently reframe. 01;16;50;19 - 01;17;12;02 Unknown And that's the name of the game is you have to iteration. You know, you find some shitty news, you know, Friday morning. Well, you have to come up with a plan in the next two hours to address it. And then you execute the plan in the next two hours. Right. And then things are better. And so, you know, it's a definitely, level of masochism that you have to go through and you have to enjoy kind of getting Sucker punch. 01;17;12;02 - 01;17;28;29 Unknown But, you know, don't do that all on your a your budget. You know for sure. You know, go try to talk to the right capital providers. And I think we were to, I would say confident that we could do it on our own. and, and that I think, you know, definitely, definitely capped us a little bit. Yeah. 01;17;28;29 - 01;17;49;25 Unknown Check. Yeah. Check. Tales. Colin. Every time he talks about fundraising, he's like, well, I've never I've never met a startup founder that was disappointed that they that was upset that they raised too much. Right. Right. Yeah. You yeah. You can always talk about valuations in the future and well you have to down round and all these things. But you know lucky you've got capital at your disposal like go run make something bad. 01;17;49;25 - 01;18;05;22 Unknown Yeah. Again you see his job is to keep the company alive. And you can't do that if you don't have enough capital to keep the lights on and keep keep moving forward. And you need to do that through just, you know, just just going through it and selling and getting up in front of people and not taking no. 01;18;05;22 - 01;18;27;28 Unknown And you're really spending your time there. But, I think the, the last thing just to leave it on is, yeah, just really think about, you know, do do your research on what's going to make you different. You have a perspective out there. number one, go for it. But number number two, really reframe how that is different than what other people believe and go deep on that, you know, and hold that perspective and be confident about that. 01;18;27;28 - 01;18;45;02 Unknown Yeah, right. Everyone's going to tell you it's not everybody thinks it's a bad idea. Oh my gosh. In the beginning we we had people we would present. We presented at conference. Yeah I, I won't name names but we had some, some were some CVC. Is it just came back to us and like you know we don't even engage with them. 01;18;45;03 - 01;19;03;23 Unknown Yeah. And they completely just, you know Taurus a new one in like three minutes, you know, ranting back to us on how what we're doing is never going to work. and you just have to be able to take that. Right. And I think any type of emotion invoked in what you're doing is a good thing, right? 01;19;03;23 - 01;19;19;16 Unknown And you have to be confident to stay the course that your perspective is different and continue to go that direction. And you know what? If you if you're three years, four years, five years, whatever down the line and that's not the case okay. All right. You learn, you do and you continue to grow. Right. It's the biggest thing. 01;19;19;17 - 01;19;39;15 Unknown It's all just about learning and pursuing your passion. And so, for anybody listening, if they want to reach out to me and they're interested, I'll be gladly to, to share all the knowledge that I have, with anybody who's interested in starting their own thing, because it's worth it. Absolutely. It only takes one to absolutely all the famous people that everybody hears about and thinks of it. 01;19;39;21 - 01;20;00;27 Unknown It's because of the one thing that doesn't mean that they didn't have 20 failures before that. Right? And again, I agree with you wholeheartedly. Failures are lessons like hard stop. Yes, right. Like, yeah, they should be and I don't want I really I would rather, you know, I want to categorize failure in a different way. Right. Because people always say, yeah, growth mindset and this and this and failure is not is is is not failure. 01;20;01;02 - 01;20;18;17 Unknown It's iteration. Yeah. The failure should be scraped out. It's just the wrong way of doing it. Yeah. Just you're talking about iteration here, right? You're not talking about failure talking about iteration. And so if I could change one thing, let's take that out of the the dictionary as a perfect. Yeah. Where can people find you in task? 01;20;18;20 - 01;20;49;00 Unknown find me on Twitter at at West underscore hammer. and then on LinkedIn, West Hammer as well. Jimmy DM happy to talk is a task that tasks out of you. Yep. Go west. Thanks so much for coming on, man. John, thanks for having me on and pleasure. Absolutely. 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;20;49;02 - 01;20;55;17 Unknown Are the ones who do. Goodbye.