Welcome to another episode of Energy Bites. Uh I'm Bobby Nean. I've got my trusty co-host, the Rad John Calan here with us. What's up everybody? And uh really excited. Want to introduce our um [clears throat] our guest today. We've got Catalina Herrera from Dikush. She's the field co and uh Nathan Simone is our or their account executive and um but both with a ton of experience in oil oil and gas and energy and uh tech. So I think a lot of good nuggets going to kind of come out of this. So yeah, I'm excited. I know we've had to reschedule on both sides, so it's it's fun to finally get get you guys in the studio. Thanks for coming. Yeah, good to see you guys. Glad to be here. Yeah, thanks for having us. Of course. So yeah, I mean, usually want to just kind of jump in and talk about, you know, what the journey looks like for different people to get into oil and gas or andor tech. And um because the cool thing is everyone's story is different. You know, it's pretty wild and I think we'll see a pretty strong dichotomy even between the two of you. Um, just based on what I know about you, but um, I mean, you start with Catalina, but I mean, how did you get in? I mean, was this was computers and stuff? Was that something even as a kid that you were really into or or kind of how did you make that trek into energy or tech? I didn't have an email until 95. Okay. Or so. So, no. No. and uh I come from Colombia so it was a lot of cultural pressure of being a doctor or a lawyer. Okay. So I had those two options clearly and when I mentioned and I never had a Barbie either and the one Barbie that I had and actually open the back to see the engine that helped the arm going up and down. So I think that was a signed. Yeah. [laughter] I am a lot into music and sound. So I told my mother, I want to be a sound engineer and she said that's not going to pay your bills. What are you going to do with that? So that's not going to work. And she was a single mother at the time. So it was a very [clears throat] important decision. How are we going to spend that money? Right. Yeah. So I started with, okay, I'm going to go for medicine. And then I realized that it was no [ __ ] way. I was not going to sleep my eight hours per night because I actually do like a very balanced life. Sure. Believe it or not. And that's when I um heard about electronics. Okay. Engineer. Uh so I did that in Colombia and it was a very good experience. Saw that they offered me to be a professor at this university that I did my bachelors. So I was teaching for the fourth time and this class on semiconductors and I was teaching about silicon wafers. Yeah. And I realized like I am wasting my time. I mean I do really need to complement my information and my knowledge and we don't even produce these things in Colombia. So I need to get out of here. Sure. And that's how I started my journey. At that time I already had a masters that was uh from Colombia in electronics and computer science. Okay. But then to get out into the United States, I I kept my F1 student visa that allow me to go into a campus in West Texas&M University. And it was a cultural shock. So just FYI, Columbia, 10 million people, cultural life, theater all over, canyon, Texas, dry town. Yeah. Oh boy. You cannot find a beer without getting from Marindo or 20 miles from [clears throat] and that was very shocking. I bet. Uh so I went through that journey and with a 4.0 GPA, Texas Tech gave me um scholarship. Okay. That was tied to Texas Instruments. So Texas Instruments has a deal alliance with um tech and if you have some decent grades and perform they sponsor you your education. I was lucky enough to be part of that group. Amazing. That's awesome. And then I actually was interviewed by Texas Instruments. They offered me an internship. I was in Houston. Yeah. And they gave me a computer and guess uh what was installed on that computer? A Spotfire. Nice. was installed in that computer. I didn't know what it was. I didn't know anything about it. And my job was a Yel engineer. Okay. So my responsibility is how are you going to look at all this data coming from all over the place with this computer and whatever is installed there and give us insights into how are we going to optimize Yelt and reduce this cost of producing scrappy things that are not going to work. [laughter] And that's how I discovered the power of data. Okay. So that was an aha moment. Spotfire became like a session. My third arm. Yeah. My third and fourth eye. Like wow. Right. So that was that was a very important moment I would say in my career. Oh for sure. No I I would say that was my gateway drug as well. So yeah I like wow. So and connecting the dots as well like oh if I produce this visualization that represents how many millions like what? So that was a very important moment and then after that I I started consulting. So Typco Mhm. offered me a position and I started in the analytics consulting business pretty much driving the seat. So now I was driving the seat with the spotfire very into hey what's going on what's new how we can maximize the value that we are providing and how we can maximize the opportunity here because now I understand what it means so it's I'm happy to share it sure and and then life happens so life happens is technology evolves we learn we continue learning and as a consultant I like driving the seat meaning I like state-of-the-art technology therefore uh I am with data core right now that I believe is leading the market into what we can do to really maximize uh what's out of data. So that's a little bit of my story. No, that's that's awesome. A lot to unpack, I think, across that, but definitely want to, you know, see where where Nathan Yeah, my journey as well. You get to decide who's uh gone through more turns, but I know I think this is our our first with two, so this will be fun to We had them in two. Yeah, navigating two guests is definitely different than than just one, but Well, I'll try and keep it interesting. Um, so I, uh, my dad worked in IT and oil and gas his whole career. And so, um, I had no desire to do either of those things. It's funny how many people you talk to, it's like, "Oh, my parents are oil and gas and I said I would never do it." And here I am. I I don't think I was even against it. I just was off doing other things. [clears throat] So, um, I was a radio TV film major. I wanted to be a sports reporter. Um, we were just talking earlier about how my last semester in college, I got to intern for Madison Square Garden, which was awesome. But, um, even then that was sort of like this crest of technology change. I was uploading podcasts, okay, that were like the interviews with the coaches, the, you know, sort of game debriefs. Um, so, you know, spring of 2005 was uploading podcasts into, you know, you know, a website so that nobody could download them. But uh it was [laughter] uh you know RSS feeds and things like that. So um but yeah came from a world where I was going to do a lot more of this um as it related to sports. So um I as I continued that path I realized that almost anybody in a sports career if you're not going to do the journalism side has master degrees. So I went and did a double masters at UCF in Orlando and that included my MBA but also a masters in sports business because I wanted to pursue um you know the uh professor that I studied under there used to be the VP of marketing for the NBA. Okay. And so what I found really interesting about the sports world was uh the sponsorship side of the business because you have to create these winwinw win between the fans, the community, the this owner, the stadium, and then all of the potential sponsors that are forking over a lot of money but need to see the eyeballs and the value and return. That's right. And so I was really passionate about pursuing that and then finished my grad school in the '0809 downturn. And so got uh sort of tied into the technology world. I did grow up with technology. I think I'm in the narrow slice. That's the Oregon Trail generation. And uh so as technology evolved, I was I keep I kept uh jumping on to the next iteration. Would you say you kept foring the river? I did. I fored the river uh and only lost a few people along the way. Uh so um no kalera. Yeah, [laughter] dysentery etc. Um yeah it's uh it's been a a very interesting journey and so then as I got into technology um I started off you know in recruiting and then got into hardware and then got into software and then got into analytics and um the thread there is that I kept finding it more and more compelling to get closer to the business problem. Sure. Um because it seemed more interesting and challenging and complex and more difficult to solve. I mean, some of the stuff that's easier to solve, more commodity oriented, it's it's a race to the bottom. And I just found this space so interesting and compelling. Um, and so I was drawn to it. And so Catalina and I had the pleasure to work together, uh, back at Tipco. Uh, and so I sold Spotfire for several years. Uh, and then joined Data Iiku two and a half years ago. And the first thing I did when I signed my offer letter was hire Catalina. So um we were glad to reunite and um yeah so now I lead uh our energy vertical for uh basically for the US and then um you know have a lot of uh influence on where the product goes in terms of what we do for our energy customers because data IQ is uh horizontal so we um you know cover every industry we're um I think they actually finally uh made this public so we're um up in six figures in revenue and uh 250 million in revenue. So, we're um we're growing quickly and uh we've hired over half the company since I came to work here two and a half years [clears throat] ago. So, it's um a lot of fun. Yeah. Yeah. It's fun hitting getting into a a startup when they're hitting that growth mode because it's it's just so different from when you started and you're like, "How the hell did this [laughter] happen?" Like, yeah, absolutely. And it happens so fast, too. Well, and when it's been a challenging time in tech more broadly. True. Yeah. True. Yeah. And you have the double whammy of tech and oil and gas. [laughter] It's always one or the other, right? That's right. It seems like I can uh and actually that was a big part of it. So I think um you know I had my eye towards what is a compelling technology for oil and gas but also that I can in an oil and gas downturn whether that while I see the rest of our company do well. Sure. Um there's going to be cycles like this that are, you know, challenging for the broader macro that oil and gas is doing relatively well and then that might flip at some point. But um to be with a company that can weather that either direction. Yeah. No, I think that's a big pro of being in the energy tech space. That's more tech than just energy, right? It's like that was at hivesell that was a appeal to me. Yes, I was in similar situation running leading the vertical but as a company at least it wasn't all you know I started at hivesell the year after we hit negative oil price and so I was like listen just a year ago this is what the situation was so like don't expect people to come knocking [clears throat] down the door for this brand new thing and this new concept that you you know we just convinced the industry to move to the cloud and now I'm coming in to tell them to move back on prem right like and so it's a yeah it's The it's always nice to be able to see that like people who really are passionate for the industry get into positions where it's not always tied to oil price, right? That that can be brutal at times. Yeah. And and probably obviously to a degree, Tipco isn't or wasn't, but they're they do a big part of their market share is, you know, energy and oil, especially oil and gas, particularly on the spotfire side for sure. And obviously uh I mean uh Tipco is used across lots of different verticals as well, but they have a massive suite of products and that's a different story than what we're doing where there's one product and so we're [clears throat] serving lots of different industries with with one piece of software. Sure. Yeah. That also makes it a lot easier to communicate with your customers about Absolutely. instead of having to try and tell someone about that is a Spotify user about stream or stream base or whatever it was called. That's right. convincing them that it's worth the money versus it's a more cohesive [clears throat] story and and we're both very happy to be here because it is a clean story. Yeah. So I mean and you can tell me if there's a better way to go about this, but I'm thinking maybe we start from like what is data IQ and like why and whatever and then maybe from there work backwards towards other tools and stuff that you've seen do the job and that you've seen or used throughout your career and like you know what it solves and then but what were some of the pain points that happened before that that it solves? Yeah, I um yeah, I think I can I uh just start with kind of what we saw all the time in Spotfire Land. And so um in that world uh we saw lots of really complex and cool things that people built, but all of that logic got trapped into a dashboard or even an app that somebody built and and that became super powerful for that use case. but then not very accessible to other folks. And um and then when you do that at scale in an organization, it gets really messy to manage. I think every single one of my customers was like, why is there not more around version control? Why is there I mean like I think they finally are did they in the in the library? I think either the last one and definitely this new one, they finally have like a a form of version control in the library. Yeah, it's I mean but that that's the kind of thing that happened for so many years that uh it was just it was a it was a frustration from [clears throat] customers. So like we would see these sort of um complex environments because it was such a beloved tool but that it enabled this self-service paradigm without as many guard rails as it should have. It created a little bit of chaos and so or a lot of chaos. Well, and this is I think this is a good topic because there's a the whole idea of the institutional knowledge, but then also like it's also one of these like um cautionary stories on how restrictive or how loose IT organizations are. Um because I've [clears throat] heard, you know, and I've been there. I was at KICO, which I mean I think they had a pretty good balance. Honestly, I had pretty good accessibility to what I needed, but like I've heard people like, "Well, I all I had access to was Spotfire. So then I'm writing SQL in information links to do this, whereas like probably better off pushing that upstream to the warehouse or some other side of it." But then you have people that are too lzair and like, you know, everyone's got their own copy and it might as well be an Excel sheet with, you know, Jimmy version 25- date, whatever. Um, but again like I'm big on institutionalizing like knowledge and like again like all that business logic gets trapped inside of a BI tool. I mean Spfire is one of a handful that this is an issue whether it's Tableau, PowerBI, whatever. I mean there's really I mean like or Excel I mean like really important business logic that is trapped there and and again sometimes only that person knows how to use it. Again you see it all the time in oil and gas. I mean companies move people between assets. So, you know, again, let's talk Alex Lack, you know, like it was on our thing. I mean, guy built some really cool stuff for us in, you know, Kaneko Bakan. Well, he went to Oxy. I probably died on the vine. There was things that I left behind when I left Kico. I'd like to think that he kept being used, but I don't know. I mean, like that's right. Um, so I think it's a mandatory journey. All right. And like it's been 20 years of learning of digitalizing more data of putting more use cases into a data pipeline understanding what we can do from the descriptive to predictive and what it means and all of that. It's been a journey but I think we proved the point like predictive can be very valuable. Yeah. So now how do we let's go back like okay let's zoom out for a second now we know that that works. How do we optimize this thing and how do we really make it better? And what it means is that not only the logic is trapped there, there is also a lot of change into the infrastructure itself. Yeah. There is also a lot of change into the teams theirelves because they are leaving and changing jobs and this guy downturns and layoffs and didn't document anything and let's start over and yada yada yada. You know, it's the the same old same old. So I think it's like okay, so what's different? Sure. And that's the thing. Okay, what's different now? Well, technology will continue to change. That infrastructure will continue to change. People will continue to change. Yeah, predictive is valuable. So, what's different? Let's orchestrate it in a way that we can leverage the team that we have and that we can keep scaling. Sure. And by the way, keep learning, right? So, that's that's what's new. That's what's different that we learned the lesson. Yeah. And that's I mean that's what drew me here was the you know when when seeing those challenges at customers particularly the largest ones I mean because as it scales it becomes so problematic that um abstracting all of that logic and orchestrating and automating and making all of that something that's sustainable that the next user can come in from a different team or from another organization and launch off quickly without having to go rebuild stuff themselves. And so that just it becomes a very virtuous paradigm to work in and it enables so much more for people who don't have a coding background or don't have those sort of skill sets and it upskills you as you do it so that you have access to those things even if you don't come from the coding world. Yeah. So like let's talk about that because obviously I come from the coding world. John's kind of definitely a hybrid, but I mean definitely not a coder but but I mean but you I mean you do write some [clears throat] code. I mean but um I let the language models write my code these days for the most part. So I mean like say someone who's used to writing custom expressions or calculated columns in Spotfire like how does like some data IU meet them in the middle to I mean is it a is there a kind of graphical interface to it or you mean is there a query language around it that you know is easy to pick up or just curious. Yeah, both. And the beauty of it is that you take it out of the box. So you don't only solve the problem, however you know how to solve the problem. So use your skills and solve it. But that creates a piece that tech communicates with his piece, my piece, his piece. Regardless of you code, you don't, you code, you don't. Yeah. And that is where the difference is. Okay. That's huge. It is huge. And then you can automate those pieces, right? Yeah. That that's so think about accelerating time to value. Yeah. 100%. So it is a guey. It's a canvas that you read like a book, but at any one of those steps along the way, each any of those nodes in the workflow, you can just drop in your own code. And so it's Python or SQL or R, whatever you have that spring your own. Um, but it also enables somebody who doesn't want to write a SQL code. I mean that might be a coder that doesn't want to write a SQL, right? But allows you to do those sort of transmit transformations visually so that you don't necessarily need to code them. Got it. Okay. So I mean I'm just trying to think about there and I think I talked to you about this too but like I think there's any number of companies that you either compete with but also don't compete with. I mean like and we were talking about that even just Microsoft in general you know competes with data bricks but they're also a huge partner. you know, whatever. But because I mean I'm what I'm hearing I'm hearing it sounds a little bit like a Matilian or SSIS, but then like almost like I almost think of you guys almost competing with data bricks. I mean certain things. So I'm just trying to you know think where you fit in like kind of that puzzle. I mean like I guess it could be any any number of pieces of the process whether it's data engineering, data analysts, data science. So yeah, I can take a stab at that and I'm sure Catalina has some opinions as well. Um, so we were the partner of the year for Snowflake for the third year in a row. Uh, from an A IML perspective. Um, this was our first year as partner of the year for data bricks. And so um, that might set off a whole bunch of other conversations, but the, uh, the place where we sit as an orchestrator on top of those things means we don't really care where the data is. And then uh, we make it accessible wherever you want to. So via API or into your dashboarding tool of choice. Um, so, uh, we aren't going to move data. We're just going to push down compute into whatever engine you guys bring to the table, any, uh, organization. So, um, we run on all three clouds or on prim. Okay. Um, so it's it really is, uh, take advantage of your existing tech investments and lots of people like that because that means they don't need to move. Um, but think of us more as like the last mile ETL, right? I mean, so often it's just T. It's not the E and the L. It's just take what you've got today and transform it into something that drives analytic and insight driven decisions. Okay. So if I've just uh for clarity, yeah, so I've got my Snowflake or SQL or combination of whatever data stack on the back end. Y'all kind of plug in on top of that. So I can then do whether it be ETL or you know AI etc and then all the way through to production deployment of that and management I guess. Mhm. Yeah. And I will add a couple of things. First the cycle when we say end to end right so is connect to a bunch of stuff put the data in a way that it can be consumed and then feed different data outcomes. and when it's dashboards or uh UX that is giving you a machine learning model or whatever else is is happening. So the fact that you can actually orchestrate this keeping the agnostic right whatever infrastructure that's number one number two adding better compute when you have it that's a huge one especially with generative AI coming right so how much are you going to pay for those tokens in the LLM versus do you really need it or you know those decisions where are you going to run that compute do you need a GPU do you need vector database do you actually need it? It's like those things need to be part of it but then at the same time it's like all the ecosystem when you mention we have a lot of software out there that does pieces of it. Sure. So we connect to it in this way and we produce this same spot was very good at that but the what happens after was a huge missing component there. How do I deploy this project into production? Can I do this with three clicks of a mouse? Wow. Yeah. Is this being deploy to all my teams with three clicks? That that is different, right? Sure. Uh how is this being consumed by somebody else and UX? And now we talk about generative AI and the LLMs and the CHPs. Yeah. Of the company. No. So how do we put this in a private tenant so nobody else sees our documents? Yeah. We can feed ourselves can use the technology that is out there. So it is a huge framework and it has many pieces of it and that's where you fit the ecosystem. Oh yeah, we have the spotfire, stabos, quicks, all of that. What is that? But yeah, analysis, visualization, ETL, dashboarding. Yes, you can play with Python here, play with R here, but at the end that goes nowhere. Yeah, because you are not deploying that into any production anything. No, 100%. Right. That's the main component right there. Is the visualization in the platform or to whatever visualization tool you Okay, beautiful. That's exactly what I was going to say. So, I mean, that's a big component of it too, right? That's a huge one. The UX. big problem with Spotfire for a lot of people is it's like I as the engineer know how to use Spotfire because I use it every day, but the finance accounting person whose dashboard I'm trying to build has no clue what that is or how to use it. And because it's built on this stack, it's now stuck in this box, right? And like getting it out of that box into something else that the end user is used to or comfortable with becomes a lot harder. So I love the fact that you guys are nice neutral third party just moving and and transforming and I will say an important part of it is is the UX and the how user friendly is because we do think in different ways. I mean I'm have my two brains pretty balanced and I like to [clears throat] think and I DJ and I paint and all of that and I am very creative as well but you think differently, you have different skills. Uh but then then the so what and the different skills UX super easy baby steps literally they're embedded all those engines are already embedded like hey I need to do geocoding or extract this from latitude longitude all of this is one click away so it's like okay this is a no-brainer even for a coder like oh I already have my p my jupyter notebook I have my machine learning perfect very precise model and now I have the opportunity to compare that and add 15 others. Sure. And compare all those metrics in a minute like wow okay that why not so it's not only connecting dots it's also empoweringmemes to go beyond. Yeah, I think that's a important part. We were talking about it just a minute ago and you know internally here we're looking at deploying some some different language models and stuff to be able to use internally and I was talking to Nathan I was like you know my problem is I don't know which one I don't know how each of the open source ones are going what output they're going to generate for my use case and the idea of standing up you know six of those on my own server and then running through them all through the same test is very daunting to someone like myself who is familiar with code and generally understands it. But from an architecture and backend side, it's like but being able I mean I'm probably going to be talking to you all about [laughter] about trying to see how we can can do some of that on your platform. But I think that's a big thing especially kind of looking forward is everyone we all know everyone in the industry is excited about AI but none of them are going to use any of the paid services because it's all proprietary information and they don't want [snorts] someone you know using their data. So that means everyone's going to stand up their own language models internally somehow some way and then it's like okay well which one do you use? Okay well how do we test that? And it's like being able to quickly deploy and test all those things is not a menial task by any means. And so having software out there that's designed to kind of do that sounds pretty awesome. Yeah. Think about it as a funnel. So the base of the funnel you may have 10 15 of those LLMs. You have the llamas the GPTs the whatever doll is but then at the end you want to democratize that and you want to just provide a chat GPUx to somebody to ask questions about their PDFs and enhance that. Done. That's that's a very simple use case, but right now key component is hey how much is that going to cost me with this LM versus how much is going to cost me with that LLM and do I need to have a private tenant for that and do a rack project instead of a simple query. Yeah, but I mean you say simple and it kind of brings up one of my favorite sayings is simple but not easy, right? Right. I mean like and I think this is a good segue into like talking about different skill sets and like just whether it's DevOps but now specifically MLOps like the skill set of a data scientist or machine learning engineer is totally different than someone needs to actually operationalize that. But sounds like your platform maybe enables now a data scientist or machine learning machine learning engineer to actually build a model and then operationalize it. So where they don't need someone with that specialized skill set of how do I write an API around this model and then I have to host that API in a container and have to send inputs and get it out. Whereas it sounds like on your platform like they can create it. You create the API endpoints for that model. Yeah. Um and with a super friendly click here and it's that glue that allows these projects to go faster. Yeah. Yeah. Yeah. I I think you know the what is a great customer quote because it's uh a little bit confusing and it makes you think about it is that we're both the lubricant and the glue. So we make everything go faster but we hold it all together [clears throat] and that becomes like you think about just the process that you described which is I we built a really awesome model but now we've got to go refactor it and code it and deploy it into production and did we miss our business opportunity because it took us six to 12 months to actually get that thing deployed and like that that happens at a lot of organizations. Well, I mean, I think if you were working on a model back in, you know, November, October of last year and then all of a sudden chat comes out two months later, even like um yeah, that's I mean, I think that's a big piece of this is the rate at which things are moving is just so especially on the AI side is so crazy that like one of the risks of like from my perspective as a user of trying to pick which one is what if I pick and build on the one that doesn't succeed. Well, then I'm have to go redo all of this stuff. So having platforms where it's like, hey, I can just remove that one and plug this one in or try all of them and see which one comes up with the best results without having to go stand every single one up by myself and then write all the APIs for them because that's how I want to use the models is is uh it's pretty slick. That's right. compare something that you've already coded to like okay here's a whole bunch of AutoML things and then you know let me hop into each of those explanability and what if analyses so that I I can see okay well actually the one we coded is better or this one's interesting and that surprises me. Yeah. And I think we can kind of break into this. I was talking this morning with um well I'll say his name he he won't mind Jeff Hughes and he was asking me you know basically why and this isn't necessarily y'all but I think your stuff enables similar capabilities like why a data lakehouse which sounds just really you know buzzwordy and it is and just the name of it you know there some shock value to it but I mean I was trying to explain to him like say in a traditional data science workflow the way it used to be like if you had just SQL server like all your data was there great it's organized Well, the data scientist still needs to go pull that in to and then load it into memory in R or Python and then do some pandas and this and that or they have to download a CSV of it and then do all these massaging steps to get it where they want it and then they got to train the model and then like all right now the model's good. Well, now we actually now we need to actually apply these steps that I applied in my R script, you know, recreate those as a data engineering workflow. uh you know then it runs through the model but well how do I even feed that data into the model I have to host a you know now it would be containers even before that you had to set up a you know some kind of VM or whatever that was hosting it and you'd feed them in there and then how do how do you get out of there does it go back into SQL server or where does it end up and but now you've got you know platforms like yours where it's like it's working on top of your data the pre-pro the pre-processing work you did is now you know already done like you worked on that to get it into there and you're working I'm assuming you have like notebooks within there. So you don't you don't have to pull it out into your IDE like it's all one cohesive thing sync with your and then the outputs of that go through and those are still accessible. So now you have this whole data governance um side of it where again your data engineers, your data scientists, your data analysts are all using the same data that you're not moving data around. It's again more secure in that way because the data stays in that data platform um and it's just easier to share those results you know with everyone. And now you're your SQL analysts are able to write SQL against the data that you you know was the output of the model. You know there's so much less cognitive I guess overhead even that you have to deal with. So um just kind of get you all's thoughts on that. That's that's part of it. So the beauty of it is I'm back to my framework because it's my favorite word for the like representation because it allows that loop [clears throat] and if you want to push it back to your new better compute initiative because that goes on all the time. It's like oh yeah we're going to invest on this new one. Yeah. Uh but this one is going to move to that one and then that other one comes from a third party that is this one. So that's always going to be part of it. So the fact that you can orchestrate it but within a dev sandbox environment design experiment confirm with your theme check out that this thing is actually working tie that up with the initiative from the use case perspective stakeholders who cares if this model drifts who do I notify why like that holistic view so zoom out and is like okay who's going to be using how and then and that framework is what really accelerates time to value. So agnosticism is key, but also being able to connect these personas and allow a feedback to where you need the data today or tomorrow. Yeah, that is a huge component of it. So that triggered triggered a few more things for me then too. I mean so does in data you can send alerts to people based off certain so I mean like so again an alerting system is not trivial to set up you know you know to do in house Even the out of the box ones are still but even then too I mean like so you know and we just got you know airflow set up in house and like nice thing about there is like running your Python scripts like the print statements become part of the log well but like you know say before if you you can you can certainly we're doing it run Python scripts as a scheduled task you know but like then it takes more effort to write a log out of that I mean there's a lot of overhead so now when you can abstract that and just like all right my print statements become part of the log and now that's there. So I mean I'm assuming again it sounds like you have the the alerting capability which is not a trivial thing to do for most people and then I'm assuming obviously you have some auditability and logging with everything happens a huge component of it uh that somebody can automate what's going on after 20 years of doing ETL in spotfire is running out of guys like load all those information links that we have at Konico. [clears throat] I I'm sure you I don't know if you ever saw my downhole dashboard over there. it was it was famous. Uh but what what's going on behind the scenes information links? So what happens when you have 50 of those and then it starts like uh more starts crawling and right? So how are you how you how you going to change that? How are you going to optimize that? How you going to put gas into it? So remove that friction and allow the normal Excel users to hey automate. Sure. So every day what do I have to do? Connect to these 15 sources create this one table and plot this one graph. Yeah. do it for me every time 3 in the morning or 4 in the morning when I woke up is the the table is populated is clean. The analysis is there. If something drifted, somebody was notified. [clears throat] Yeah. If that somebody was notified with a production very important use case, a data scientist was notified and is already looking at the drift and is already retraining a model and is already redeploying a model. Yeah. And somebody can sign off. Yeah. This is aligned with our policies and this is aligned with what we think internally because that's the other thing that we haven't talked about. But this is also something that we have been learning which is hey not only we're generating data outcomes but we have to generate data outcomes that are responsible. Sure and that we understand what the heck is going on there and if something is wrong we can track it back or fix it. Yeah. Yeah. No, I think that's a I mean that's initially five t years ago when people started talking about AI in the oil field that was the big problem, right? Is you've got you're presenting to a room full of engineers who have like I as a mechanic I'm a mechanical engineer because I have to know how things work. I take [ __ ] apart and I put it back together. I might break it three times until I finally figure it out, but I will do it. Um, and so you're sitting in a room trying to pitch this black box that you can't explain to them and they have no idea and they're already not trusting of you because you're tech and you're big scary thing. But it's like, yeah, being able to have even just auditability of like not even just with an AI model, but over time within a company, right? Like Bobby left and so and so took over and that's where this piece of code changed, which is when all the issues started. And so now we know where to start debugging instead of here's the file. Exactly. And you're like, well [ __ ] I have, you know, you just start going through it. Yeah. So that kind of triggered another thing for me. Like so what what about like data like a a data dictionary, but like maybe even lineage like I mean discoverability. I mean like people like I want to find this, you know, like I mean does your platform platform help enable that too? Like you can find where that data is or where it started, you know, how it got to where it is. Yes. and and think about it as a as a Lego of building blocks. Yeah. So sometimes to find that data you that data can be on a structure or semructure or whatever you may need to still do some mo some processing and some basic NLP but at the end you are going to be tagging or highlighting or grouping worldwide. I mean yes you can you can use all kinds of NLP to surface what you need to surface but the beauty of it is that you already found a use case you can copy paste and change on a structured data and serve it to somebody else and then copy paste and serve it to this other team that cares about this marketing insights sure and these ones care about human resources so they're looking at different PDFs but at the end they're asking the question that they want to ask and that's how you scale are you just going to jump off that point there's there's cataloging so like everything that you've linked to in data [clears throat] IUS like then it's surfaced to people who have permissions to see all those things and so they can see what's being used and how often it's being used and then there's you know not just the what have you connected to catalog but there's the feature store so these are the things that we've worked on and blessed and that's a data set that we should use more broadly in the organization and it's used in all these different subset so you can have a almost like you know maybe gets used too like a medallion architecture or something like where like this is like the gold layer or the platinum layer and like you know this other stuff is more kind of in flux. That's right. Yeah. And then you use that to feed what ultimately data engineers want to work on things that become used. And so what ends up happening is there's [clears throat] you know lots of analysts and subject matter experts who who do things and we want to encourage that work. And as they do things then data engineers say well I need to build that as a robust pipeline and yeah you do but like let's let more people drive that and have it democratize in terms of what the key things are. Yeah. Yeah, because I mean that analyst at the edge knows what the business needs, you know, and you can enable them to do some of that, but then again, it's still done in a more robust or exactly scale. No, I mean just having I think, you know, historically speaking, one of the biggest issues the industry has always had is this silos, right? And it's like they're on the same floor, but they don't even talk to each other, right? like the the completion engineers and the reservoir engineers are literally on the same floor in the same building in the same city and one is saying one thing and the other is saying something different and I as a service provider have no idea who to listen to or what to trust or whatever but thinking about it from the flip side is internally right the problem historically speaking has been we've got all this data but it's in different places or it's locked down by it and this one little thing and as an engineer I'm going to find a way to get the data regardless of whether it locks it down. So now I'm, you know, I found an AI bot that's going in and just downloading CSVs for me once a day so that I don't have to do that when in reality if it just had a platform that had permissions, I could have access directly to the data sets that I need to have access to without interrupting any of that security piece and it can all be built off of the same handfuls of ma mother data so to speak that we know is good and then anybody can use it, right? And so that's that's I want to add a piece there because you just touching a very important point. I worked around it. Oh yeah. Yeah. That's happening a lot. Yeah. Because you have all of these frustrated people that is trying to actually make a simple decision and all they need to do is have visibility. So it's as simple as that. But what I am also seeing is how culturally it is starting to change as in okay I'm not going to be the bottleneck. How I'm going to govern this? how I'm going to give you your data catalog or your future store or your group. Go and have fun with those dashboards. I don't care but I'm going to orchestrate it for you. So I think that's where the real value comes today as in what skills need to change with this technology. is like we have to be the humans empowered by the technology and keep bringing the unique humanness that we can bring to this table that is not replaceable by just doing it on top of the data too right it's not from or in the warehouse or directly from it and that's I feel like that's the thing we we a lot of us me personally uh talk a lot of [ __ ] about IT departments a lot of the time but you know they're a lot of them their main goal goal or function is to keep the company secure because their data is one of their biggest assets, right? And so it's not because I don't like the IT folks. It's just there's conflicting incentives, right? Like engineer solve the problem. IT guy lock everything down so that it's designed friction, right? Exactly. And so being able to have something like this where it's like, hey, you can have access to this and even if you do screw it up, it's not going to affect the actual database. Right. Well, there's that too, but I mean, and I was wanting to maybe segue into some of like some of the cyber security kind of side of it because I mean, you know, I'm lucky right now, I'm at a private company. We don't have to worry about sock compliance and all that stuff as much, but I mean, I'm sure I have to imagine your platform is sock compliant. Um, but also, I mean, just again, by keeping everything in the platform, be able to send things within like I mean, I think a huge data security risk is like people just emailing around Excel files and CSVs. I mean I mean like that in of itself just like sending around crucial information like that way it's a is not um [clears throat] it's just not efficient at all but like there's a chance that it gets sent to the wrong person or someone outside the organization like seen that happen many times you hit reply all instead of reply and then the client gets the data that they're not supposed to or somebody else's. Yeah. Or I mean god forbid like editor you know like I mean I mean because like say you had a data sharing agreement but you sent out you know. Yeah. Um well even then like the easiest I mean in my in my very limited knowledge on the cyber security side the one of the easiest ways people get hacked is through fishing which means they get access to your email first and so oh well if you got all this proprietary data in emails because that's how you're sending them around. Not the most secure way to do it right. Yeah. Yeah. And that is going to be even more important with generative than LLMs. [clears throat] And how are you gonna handle the toxic like what kind of guard rails do you have to put there in terms of private data and all of that? What kind of searches are allowed? I give me the list of security social securityities of like can you do that? What is the what is the guardwell on those queries and all of that and that's part of the experimentation that needs to happen before thinking about deploying to production and LLM use case as well. Yeah. No, definitely. I think when you think about several of the things that we've kind of touched on the last little bit here, it's the challenges that exist with traditional desktop oriented tools where you end up sharing information inefficiently and maybe insecurely. Yeah. the challenges with um handoff points in data processes that ultimately create delays and potentially missed opportunities or um you know just delayed value. And then when you think about how many groups are used to working alone in their discipline or maybe in their asset team um but not as aware or keen to what another group within their organization is doing, there's missed opportunities from that standpoint. And so really why we're so passionate about being here is because the browser-based interface to collaborate across teams across functions and in a paradigm that is both self-service to like enable anybody of any sort of background background persona discipline to do their data driven work and then give it the comfort to know all the governance, security, policy, everything is all tied in with the back end. So, um I know just in my decade of selling analytics and software to oil and gas like security and architecture is just like the biggest way to just get stuck with the technology department because they'll just they'll they have extreme concern. Just to your point, I have never seen a software cleared as fast as data IU from IT security and architecture. They they do deep dives and they're like, "You guys already do everything the way that we would and we work with the largest companies in the world." And so something I've struggled with in prior lives is when you're working up the stack, working towards larger and larger companies, it's hard for companies to scale your technology from a security and architecture point to really serve the the big boys and data luckily has half of the Fortune 10. So we're working down market and bringing that to bear across more operators, more service companies. Yeah. And that's what's really exciting is that um a scaling software like this is most compelling to the largest companies and more and more companies are starting to take advantage of it. Beautiful. Yeah. No, being able to uh being able to come into an office and say, "Well, we're working for whoever, right? top Forbes 100 company is a very easy uh way to break the ice with a lot of IT and cyber security folks because we've been through the pacist with the biggest ones. Yeah. Yeah. Well, I mean those are obviously all public companies and they have to go through the rigor that you know requires and um uh you know another thing I want to kind of hit on and I again I just want to see if if it's possible what you do with it. What about like you say streaming data andor like IoT? Do you all have any kind of your fingers in those spaces at all or? Yeah, I mean we consume Kafka. Okay. And we can do whatever we want [laughter] CFKA. Sure. So, but a majority of streaming use cases in real life end up being scoring realistically. Yeah. Yeah. We always talk about is like what does real time mean to you? Exactly. [snorts] Everyone gives a different answer is yes. And yeah different it means different things for different people but again we orchestrate what needs to happen on top of it. Yeah. So we don't have people usually putting data IQ out on the edge but we what you do is once you've got that data pulled back to something then you'll build a model off of all that historical data and then you'll deploy it maybe at the edge if it makes sense. Okay. Yeah. So you can obviously develop that model, you know, in data IQ and then I mean how easy is it then say to like generate say a container of that out of I mean like is that something that's baked in relatively or I mean what does that process look like? Yeah, that's the one of the huge advantages is that it makes life easier and it becomes a template or a configuration here and there and then it empowers people to get faster and get things done. So, it's a simple UX. Okay. Taking a couple of buttons. So, that leads me to the question I've been waiting to ask the whole time, which is federated learning. How it sounds like y'all have I mean just where you sit and the way that you can do things is like it's a perfect kind of platform for that. And for those who don't know what federated learning is, it's this idea that you build one model, then you deploy that model to many different locations, and then those are their own models running on those locations. They take their learnings, push them back to the mother model, it gets retrained on that, and then pushes that back down to its children models, and it's just this continuous loop. But it's it's literally what OpenAI has been doing for the last year. M you go and you sign into GPT and you do it and it takes the learnings or whatever that you taught it essentially there and it pushes it back up to the main model and now we have GPT4 turbo right like how does how does data IQ handle that and is it sounds like it's kind of set up very nicely for for better learning as well is that we actually presented that was the use case that we presented at the hopes on envelopes in Houston when when was that like a year ago yeah so exactly that and how it branches out. Do your thing and come back. Why? Because we can literally copy and paste blocks. Yeah. So, yes, it's a lot of flexibility. That's very cool. So, what what are some kind of just let's call them wins. What are some wins that you guys have seen, you know, different clients have whether it's in oil and gas or outside of it just using the platform? some of the big kind of hurdles that they've overcome or ROIs that they've been able to achieve. Go for it. And then you complement everything. I'm just going to highlight a couple of things because we mentioned snowflake, we mentioned data bricks. One of my customers right now the aha moment happened when the accounting department was able to use the data bricks like like that simple thing. Yeah. allow them to use that data and automate a report. Like it can be simpler than that. Yeah. But that is an ah moment. So what's going on is that we are actually gluing the opportunity to maximize the investments that are happening in terms of that technology, right? And connecting the dots for people with zero code background to be able to leverage that. So use cases as simple as that. But now more advanced enterprise companies that had been doing this for years, let's say in oil and gas. Yeah. Uh what's going on right now, the aha moment is MLOps and really have a governance framework where you can keep an eye on that AI portfolio because AI portfolio is going to be pieces of things that can be a data outcome. Again, it doesn't need to be machine [clears throat] learning for AI for be part of that portfolio that you need to have an eye on. So that is opening doors for teams to collaborate and work together and use cases now classic use cases like the predictive maintenance that we have been talking for the last few years or so. Now what is going on with that predict maintenance result is deployed in the field. It's in an automation routine where drift happens and the proper people gets notified and all of that but at the same time is connected to a business unit that is also working three more projects and now everything is part of this framework. So I know teams, I know who is doing that, what and I know I have visibility on it. And then the cherry on top is the acceleration that LLMs are bringing to the table when you grab the classic predictive maintenance. And now you add generative AI on it and you just ask the thing, hey, show me the equipment that are in high priority of failure and give me the locations. Boom. Yeah. Right. So those are aha moments and and the fact that you can do that in a matter of weeks instead of months. Yeah. And it still reuse what your teams have been working for the last 10 years years or so. Those are the aha moments. It's like okay finally these pieces are actually being orchestrated. Sure. Yeah. So that made me think of something that I thought of earlier and then immediately forgot because add but uh so could I also theoretically use your platform to distribute information to people external of my company as well? I'm thinking of like my background coming from the frack side. The problem with frack analytics and AI ML is that half a dozen to a dozen different companies generate all of the data in different timestamps in their on their own systems essentially and of course it's all the operator's data but it all lives on their side until the operator gets it and there's no really good way central repository so to speak for all of that to kind of come together you know coore is trying to do that corva is kind of pseudo doing that But being able to have all of that data like all the service companies push it to the operators uh instance and data IQ and the data IQ says hey kind of like a pseudo CFKA you get this piece of the stream because that's the data that you need for your job and you get this one over here but you don't get you know the other stuff from the other vendors and stuff like that is that a feasible conceivably you could do that yeah for sure and so that like because of the data cleaning capabilities in data IU then you could set up those, you know, as you receive those into files or databases or however you want that to then you you run those transformations and you stitch that together and then you let the people listen to those automated outputs that come. And so because some of the alerting and things that we were talking about is data IU can trigger actions on new data or schedule or whatever you want it to be and then as a result of that trigger do the alert and action and whatever. So that again that orchestration means that um you have there's no easy button for it. You do have to set it up but then it makes it accessible to the people that need it in the organization. Yeah. Oh that's interesting. Yeah. Uh, one thing I was curious about, so like you know people have their snowflake or the data bricks but also maybe running data IU on AWS or Azure or internally like do you does it have like its own Python engine within it or does it have its own like R engine because I mean you saying I could plop in a Python script or an R script or whatever like does it come with one or do you have to you install like how does it handle like where the where that runs because I didn't know say with Python you know say snowflake has their snow park and then you Obviously you can run Python and data bricks you know and that compiles down to like some spark code like so does data Iiku compile those things down to those engines or if it's just a generic Python script where does that run just kind of curious on that the engines so we leverage snow park in a snowflake and that's a big aha moment as well. Yeah. Uh because you are connecting a lot of dots there. So yes, we leverage the infrastructure. Yeah. And it goes down. Okay. But yeah, I mean if you need to bring in your Python libraries and all that can be managed within data. Exactly. Yes. And and as a data scientist, I mean I like independence. Yeah. and I like control and just let me experiment with my three libraries and my packages and let me have this kernel here and let me have my code environment here where I have fully control into what's supporting that use case. Sure. And and and then I am ready to operationalize and let's share this with the world. Yeah. But yeah, that independence is is key. And you mentioned ids and there is a lot of people that still loves the ID ID UX and if you like that go for it. So we sync okay that and whatever you are doing that that is going to be represented in the flow itself. Okay. And so then I was that's kind of my next part of that. I mean is that through like a git integration you have um because or do you have say a VS code extension? Yeah you I mean is it either both? But and depending because a lot of people is already doing some CI/CD at at some level and and git is a big part of it. So we we're not going to be disrupting the way that they're doing things. We're going to optimize [clears throat] where they are doing things. So we're going to pretty much understand that CI/CD and put data coup on top as that orchestration of those pieces but for these people as well. And now we all have the visibility on the CI/CD plus the alert team piece, the drift uh governance and the whole thing. So it's a it's a more user friendly CI/CD that smooths that framework for these teams to really maximize what they can do together. But then say can someone write say a Jupyter notebook with you know within like data data IQ's notebook but then like push that down to get from the platform too. I mean or you can you can write your you can bring a Jupyter notebook that you already have or you can start and write your Jupyter notebook and it's going to become an orange bubble literally. So it becomes an artifact. Yeah. And and the traceability and the interdependencies and all of that is embedded in the flow. Okay. So you don't have to sit down and think about documenting anything because it's doing it for you. And that's another aha moment because we don't like to document things. No 100%. I mean like and again yours platforms sound like this on steroids but you know say we're using like dbt you know for our SQL transformations but like again you have that traceability between models and like reusability and then it does some auto documentation and again you there does need to be some work to enrich that um but at least out of the box you can see how these things play together and see like where this came from and what code generated what I'm looking at and everything. So, and I get just like onboarding people too has been huge. Like when someone and that's kind of what I will sell if we're interviewing like an analyst from another team like when you walk in day one I'm going to give you access to this documentation and then when you're using you know this SQL you know view at the end like you can come to the documentation see what transformations are happening there what's happened downstream and how it you know it's being calculated. Um, so it's just, you know, as an onboarding thing too, people can come in and be dangerous almost immediately. Exactly. Innovation, that's where innovation happens. Let people be dangerous. Let them experiment. Let them try LM versus another one. Let them create that interface. Let's ask questions. Let empower them. Yeah. Well, and I mean, it sounds like you're using it as a recruiting tool, but it's also a retention tool. If you give people better tools, they're not looking around to go elsewhere. Yeah. Oh, I mean you I've seen it on the other side where uh I won't name names or our ERP is not great and like you lose accountants because they're like I can't do this every month. Yeah. Like you know bad tooling will ruin someone's life. And I think, you know, we had Mark Ear from Dakota Analytics and he had a great story about how some of the tools they built, like I mean, actually helped an accountant that was spending, you know, nights closing the books till 1 2 a.m. and then she was actually able to get home 600, 7 at night and see her kids, you know, like she's like, "You literally changed my life." I just by having better tools or better, you know. So, I mean, it's it's bigger than people realize. Absolutely. Yeah. Well, and people don't unless you've had to do it, you don't understand how hard it is to go into somebody else's code and figure out what the hell they were thinking or why they did this specific thing there and when and why or any of that stuff. And so just that traceability and auditability of things is so nice to have when you have it because when you don't it's a huge pain. Yes. I personally started from zero a couple of times because of that. Yeah. Yeah, it's like that's easier mostly. Yeah, I'm taking control over it. Yeah. No, it's very true. Okay. Well, [ __ ] I think that was quick. [laughter] Yeah, it was quick. But as I expected, you know, but um yeah, I want to kind of jump in wrap it up to the lightning round or the the speed round. Speed round. So, we'll just pepper y'all with questions. Just give us short fun answers back. But, uh favorite cloud GCP, Azure in trouble. Yeah, getting me in trouble. She [laughter] has to That was That was a loaded question. So, I'll I'll pivot. Um let's go with favorite. Um open source anything. It can be favorite or most useful or something that you use often or Well, right now I am loving Mid Journey. So, if you haven't tried it, go for it. Yeah. I love it and it's empowering me as the human. So yeah, generating I mean generative AI and generating these images from text and for image and prompting and seeing the power of the prompt and yeah seeing that generating an image out of it. I I am fascinated by it. It's it's crazy this stuff that they're doing and they're already doing stuff. I was telling uh Ding earlier today, you know, they've already got stuff where it's combining, you know, language processing. So, it will translate the language, but then it will also generative fill the mouth so that the mouth matches the words that they're saying, not like all the movies that we grew up with where it's like they're talking and it doesn't work. Yeah. Yeah. Right. Really bad dubbing. It's crazy the things that are going to come from that, both good and bad, unfortunately. But like the deep fake stuff is terrifying. But for the good, right? Like for us, we're looking at it as in like how can we generate this podcast in five different languages so that more people can hear all the stuff that we're talking about, right? Like that's very valuable and very powerful and fascinating all at the same time. Yeah. [snorts] Yeah. Yeah. I don't know that I have any uh strong opinions on this. I think you know what's interesting from my standpoint I don't I'm sure I use open source things but I've never coded anything in my life and so um what I think is interesting is the technologies that are open source because they pull so much community adoption um and so there gets to be really powerful user bases um that lead to roadmap of product and stuff like that so data IU doesn't happen to be that but I think some really interesting um products have been launched over the last you 10 15 years that led with you know open source open source and community first crowd sourcing helps you know improve the product faster on it [ __ ] I mean there's so many that yeah if only our industry would learn that well they had an attempt with the ocu I mean I think that it's still going on come on [laughter] there have been a number of consortiums that I've watched spin up and slowly uh die over time. And I'm not saying that specifically about anyone, but the my problem with consortiums in our industry generally is that they end up getting led by uh or a lot of the power comes from the big service companies or the other people that are involved that have obvious incentives for it to go one way or the other. And the whole point of open source is that you put those incentives aside and think about what is best for the mass and for the greater good and the community. And I don't [clears throat] know the solution to that, but I hope we figure it out sometime soon because I mean Wi-Fi, USB, Bluetooth, like there's so many everyday technologies that everyone uses that came from just that concept. [snorts] Um, who's your favorite social media follow? Well, I I actually have a few. Okay. And uh I will say that I like the AI and advanced analytics female community. Okay. So, of course, uh Kate, you know, Kate, I follow her a lot. And Casey as from from Google. Okay. Well, ex Google because it Yeah, that's Casey Koser. Yeah. Uh I have been in the market for a car for a long time until just recently. So, uh I was watching car dealership guys podcast was [clears throat] loving it. I mean, um, so I did purchase a car, so I'm not following it as close as I was, but was a fascinating run as he kind of like appeared from nowhere on Twitter. Yeah, some really awesome information and then great interviews on his podcast. I actually have not listened to the podcast. Yeah, I follow his Twitter quite a bit and you did it at an interesting time too with the car market and stuff all going well and it it changed so much month to month that it was uh I mean it was a fascinating to follow but it was really there was something too where I feel like he kind of moved the market on Buicks like he mentioned some how Buicks were like you know they were selling really low and then then it was like the next month like Buick had Yeah, exactly. [laughter] The power of It is the power of social media. Power of content, man. Yeah. All right, we'll uh we'll wrap it up. What's what's one piece of advice you'd give to someone coming into kind of the energy tech space? Uh be very aware. Technology has come a long way and this is the time to be as empowered as you can be and bring the differentiator as a human that you bring. So education is key. Understand what's there and and maximize that opportunity. And I will say if you haven't tried any of the generative AI bots and so on, you're late. So get [clears throat] in there and start questioning these things and start understanding what it brings for you and how you bring that special layer which is the humanity on it. Absolutely. So what it means to hallucinate, how we're going to control that, how we going to really connect the dots here towards this one goal that we are sharing in this very diverse group. Yeah, that's% advice. Well, I mean, it's going to enable people that weren't didn't that didn't have access to that power before. I mean, exactly. It's already happened. But they have to be willing to not not be afraid of it and to Yes. Exactly. embrace it. Yes. Practice it, too, because the the degree of output significantly changes based on your prompts with most of these. And I'm not saying you need to go write a paragraph every time you prompt the thing, but it really does how you ask the questions. Absolutely. because we were talked about in the last one too with Ding, but like Andrew Ang or whatever or I think but um he had a really good it's a one-hour course on prompt engineering and you just learn like this is how you have to ask the questions so that you know because again you can and they do a decent job if you ask it a decent question in your your native language but if you break it into these steps or tell it that you're this or you know really outline it it takes it to a whole another level. Yeah. act as and your goal is to are very very important things to put in most prompts because they really help. Yeah. Like well think how you're going to be 10x more productive with the technology that you have in your hands today. That's where your mind has to be. Where where is your side of it on how you're going to have your own wings? Absolutely. Yeah. So um to that point maybe a slightly different uh attack is the um the world will continue to consolidate for operators. I mean that's uh been happening it'll continue to happen and um there has to be an opportunity to differentiate yourself in those organizations as they combine particularly if you're part of the group that's being acquired. And so, um, what I've seen over and over again is the change agents who can demonstrate that they brought the business value, um, are leaps and bounds ahead of their peers for promotions, for, um, even making themselves, uh, compelling to other companies outside of that organization. And so if you want to, it's not even necessarily about advancing your career, but if you want to demonstrate the value that you bring to the organization, the easiest way is to demonstrate your ability to be a change agent because decisions at these organizations are becoming harder and harder. And if you can navigate these organizations to make decisions and make impact, um it is it is something that uh is clearly visible through your management line. Yeah. No, that's good advice. Again, I would go back to but I mean Ding had very similar, you know, talking about how he he navigated different companies and oil and gas. You're going to have the ups and downs, but like are you solving people's problems and like are making things more efficient and better? Like if you are like it gets noticed and you're going to stick around like absolutely or or you're going to make it through the layoff, right? Like that's another big [laughter] one. And now how are you going to measure success somehow, right? Let's align on that and stuck in that. Yeah. What does success look like? Yeah. Yeah. Cool. Where can uh where can people find y'all? And yeah, I'm Nathan Simone on LinkedIn. Um I'm relatively active. I don't do any uh Twitter or anything else, but um yeah, here in Houston and um make my way around the oil and gas towns pretty frequently. So, same. Google my name, you'll find me. [laughter] Google me everywhere. I love that. Google me. Thank you'all for coming on. We really appreciate it. Thanks, guys. So much behind us.