Welcome to Energy Bytes with John Kalfayan and Bobby Neelon. Your essential guide to the intersection of data and energy. This podcast dives deep into the world of energy, shedding light on how data, AI, and technology are revolutionizing this sector. Each episode equips listeners with insights into the most efficient tools and resources, paving the way for a data-driven future in energy. From technical nuances to broader industry trends, Energy Bytes offers an unparalleled perspective on the evolution of the energy industry. Join us as we decode the algorithms of energy, one byte at a time.
I've lots of fun uh butchering of French names the first time I started calling on people in Louisiana. Yeah. Well, that's different French. That's Cajun. Right. Yes. I was like, "Oh, Mr. Fontenot." And he was like, "It's Fontenot." [laughter] I was like, "I am an idiot." I I was I was in New Orleans once and I asked a woman like um I like a deli shop uh how do you say that word? And it was cochon which means pig and she said coochon. I was like, "Okay, you're [laughter] not not not as Cajun as you were advertising." You were advertising. So. [laughter] [snorts] What was the snowflake news? Uh I mean, it's just the last couple weeks there was like a a data or like a cybersecurity deal, but How did that originally come out? It was somebody somebody like leaked a bunch of [ __ ] right? That that [snorts] I'm not even sure. It just kind of came up. Like I, they ended up sending them something like, "Make sure you're using MFA if you're not already," and, um, and like, "We have like IP whitelisting that you can put around." But really, like, everything I've read about it just comes down to like, the tools were there for people to protect themselves. Just You mean not using the default passwords on my database is a bad idea? [laughter] How many How many hacks have occurred because of that? Well, it's so easy just to use MFA these days. Especially now. Yeah, and they make it in through something like with Duo or you can use your own yeah authenticator. 100%yeah. No, I mean, I've been using Authy and authenticator, you know, Google Auth whatever for years because of crypto and I remember the first time I did it I'm like, "Man, this is annoying, but it's really cool because it's like all and now everything has 2FA. So, it's like,"Oh." Okay. Yeah. Well, and what's cool now is I've seen some that are going passwordless and you know, you're just looking up your code and maybe it's combined with a PIN or something. And so, you don't have the two steps, but you have the dynamic just on my phone approach. Yeah. No, I like uh the the two most functional ones I've come across HubSpot uses 2FA, but it's like if you have it on your phone just like Gmail or whatever, right? It asks you every login attempt and you approve it. Um or just like the magic link. Like I know it's annoying to go back to your email, but that's way easier than remembering your password. Yeah, it just sends you the link. You know, it just sends you the link. But, no [snorts]. That's uh the never-ending trend on the security side is use the security tools that people offer. Yeah, human I'm humans are always the point of entry. Yeah, 100%of the time. Well, there's one company I heard about that they um never mind, let's skip that one. I probably shouldn't tell [laughter] that story.
So, speaking of Snowflake the how kind of I know I I assume you guys do a lot in Snowflake with different customers and stuff. How kind of how have you use Snowflake over the past couple years? Yeah, mentally, you know, it's one of those um we were talking about this earlier that do you bring a bazooka to a knife fight, right? And so, a lot of things that we do is very much um within the Microsoft stack and we love that because um we just inherit the arc enterprise client's existing security and just give me a login. Um and once we have that, then we can operate within their ecosystem. Usually combination of their power platform at like Power BI is the most popular, but Power Apps and Power Automate, as well as their Azure or AD APS instance. And so, we can inherit all of that. Um so, I actually don't think too much about cybersecurity cuz when they ask that question, it's like mine is yours and um it just kind of sends that question right through. So, pretty easy to get running. Yeah. [snorts] Y'all, are you seeing a lot of people using Snowflake these days? Um, I feel like that was like the option, um, 5 years ago as far as like really adaptive data models for machine learning and AI. But a lot of people are picking up much more native tools, or the [ __ ] DBs of the world, to make it to just take on a little bit different flavor. And there's just so much, uh, ingrained toolkits now within some of the clouds that when you go standalone now, if you're starting today, it you're you're signing yourself up for some heavier maintenance, right, and longevity costs. It's always the pro con, right? It's like,"Oh, well, Microsoft can do it, but now I'm locked into Microsoft [laughter]." Yeah. Yeah. So, it's like,"Ah, change every 5 years."So, or at least threaten to change every 5 years and [laughter] renegotiate. Yeah. I mean, I think the problem like say Microsoft on the data side until maybe recently and we'll see if you know, Fabric is the panacea that they are pitching it to be, but they've really struggled in the data warehousing, you know, the data lake side of things. Like they you know, I think also there was like Teradata and everything was able to do some of this heavy processing like on prem and that was I remember when I started at Conoco in 2014, it was like,"Ooh, we got Teradata." Yeah. You know, which again, not to knock it. It's it's a great it's a great engine, but it's more of an on prem kind of thing. Then the cloud became a much bigger deal and then [clears throat] Redshift, you know, kind of took the world by storm on AWS and able to use that multi-parallel processing columnar kind of stuff in the cloud with a Postgres data flavor. But even that [clears throat] like it's just if you blink, everything passes you by and then all of a sudden like Snowflake came in and it was just like, "Wow, this is I just turn it on and it works."Yeah, or BigQuery is really similar. If you were looking at a more native tool on on GCP, but um Microsoft has really going to struggled. I mean, like, over the last 7 or 8 years, I remember looking at it and it was Azure Data Warehouse, and then they rebranded to Synapse, and now it's Fabric. And it's kind of hard to tell: is this just another re-purpose garbage, or is this actually all cohesive yeah, you know, platform that's actually going to like solve a lot of people's problems. Well, and I think you, you just talked about kind of the evolution we're seeing in the industry, and so Microsoft seems to be kind of going through their own internal evolution; and as they learn and they're, they're adapting and rebranding and repackaging things, but you know, technology is moving so fast these days that, um, the tech companies have to keep up. But I, I think where, you know, the enterprise clients are is, it's where is your home base, right? Establish your home base that allows you a lot of flexibility and agility to keep going. Um but pick up the new tools as they're value adding and I think that's the key thing is what's value adding, right? If If you can use a SQL database and it achieves your objectives and it doesn't sacrifice something for the future, use it, right? Yeah. Um and so I see I feel like blockchain was such a big thing like probably like 2016, 2017 where everybody was like,"Oh, I need blockchain."I was like,"You just need a database. You want to go from paper to a blockchain?"And like there's no like issues with um um you don't need a distributed ledger to help with that that no trust kind of approach. And it's just like pick the right tech, you know? Let's just put put it in Excel. That's better than [laughter] your paper exercise. Well, I remember John and I went to there's a blockchain conference at UT and this guy Jimmy something like Asian dude had a wears cowboy hat. Like he's pretty popular in the Bitcoin world ever, but like he's like,"It's just a [ __ ] database with rules."[laughter] And then the big thing is like, you know, I think you're I see this a lot in different technology now like like databases and everything were kind of everything was all wrapped in a tightly coupled thing and now it's really everything just been turned inside out. Yeah. You know, so like a a blockchain is just a database that's turned inside out. And now like it's a you need a cohort of people that agree on these rules and like that's the hardest part. Again, the people part of the people processes and technology. That's people in the processes part of the blockchain is the hardest thing and it's been turned inside out, but it's at the end of the day it's a database. You can do those same rules and if it's something you're doing internally Mhm. then you you're in control of those rules anyway, so you don't need blockchain. Yeah. Like I said, to your point, you just need a database. 100%, yeah. It's when you have that external no trust kind of a person and you want that distributed network. I think is really where it starts to unlock value. Yeah. Yeah. Yet another beautiful point of Bitcoin. It's like, well, I mean, when you really think about like the psychology of how the code works and like how it incentivizes people, and they knew that they were going to need incentives in order to keep the network going and growing and all this stuff and over, you know, generations, not just a few years. It's pretty crazy to think of all the foresight that kind of went into all that. Yeah. Well, and what I'm excited about, particularly in the blockchain space, is more of the uh distributed computing. Um and so today you're just crunching purely to process the transaction, but I think Polkadot was one of the first quality ones out there right about it like 2 years ago. Um but if we think about the um natural gas uh um Bitcoin rigs today, that's just creating coins, but what about distributed computing power? And so you could literally compute wherever you have uh an existing um energy source instead of having to transport it a long distance and I mined, uh, I think it's called Theta, um, for a long time on a laptop that I just had sitting at the house and. Their whole concept is basically distributed caching for streaming services and just high compute loads, but it's like you can literally have it on your laptop, have it on your phone, and you can imagine this world where it's like,"Oh, [ __ ] Everyone has this Apple buys whoever, and now it's baked in on everybody's phone, and they literally have a the largest distributed compute network and so Bobby's phone is buffering the video that I'm watching and like all this crazy It's going to be crazy sure you want to talk about Silicon Valley? [laughter] The show. I love that show. It's such a great show. One and what it could potentially mean is that all you really need is a screen and a and a Wi-Fi card or cell card. And then all your computing power is done elsewhere. And so we think about, you know, just minimizing consumption on on the masses, you don't need full CPU on every single device that you you have. Just send it into the cloud and it's taken care of, and it could be 10 ft away or you know, 10,000 miles away. And And then with I think with 5G is it's going to make some of that more possible too because like because like, you know, again, it's great that you compute, but then you got to transmit the data inside too. But I mean now what you can do wirelessly with stuff like that or Starlink or you know, the lower orbit satellites, you know. And again, it just everything accelerates so fast. We've seen it with AI, but it's all all across the spectrum. It's the compute power which has made the AI possible, and you know, it's all interconnected. Well, that's one of the things I love about the Industrial Revolution 4.0. I feel like one, two, and three was just around mechanical [clears throat] advantage and electrical advantage and then you know that combination of like Henry Ford and process optimization and then the semiconductor. With 4.0 it's not one specific technology. It's just that there's a whole group of technologies that gotten good enough and now the theories that came out in the 1950s for machine learning we actually go do something with it. And so, it's just like so powerful. And and you know, people are like,"Oh, we're on 5.0 with with um AI now and the large language models," but I just think it's going to be a continuous, cuz if you look at historically, it was 100 years and you know 70 years and then smaller and smaller, and I think you just get into this kind of continuous innovation cycle. Well then the quantum computing is going to blow everything on its tail. I mean quantum computing. Yeah. I mean all the security stuff like oh your 256-bit [laughter] encryption I can solve that in 8 seconds. You know like whatever. I I I read something that China and the US are storing all of the intercepted documents that they can't crack that are encrypted and they're just hoping in like three or four years they can just crack everything. Waiting for the quantum to come out. [laughter] That's crazy. I was like that's actually kind of clever but scary as heck. But then, when that's going to be combined with the AI or the models and stuff like that, just with the way they're going to crunch through the data, is insane. And then again, just data creation it's, you know, it's not a new story, but I mean, it just, we're creating so much data like every Yeah. literally second or millisecond you know. If we really think about it, all the industrial revolutions is really around driving down costs and driving up productivity, and then the cleverness is: how do you apply that in a value adding way? And so, it's the businesses like the Fords of the world that were able to apply a little bit of technology and a little bit of clever process and lock an entirely new industry. So 20 years you went from all horses to all cars and so just night and day difference. Incredible when you really think about it. In in the early 1900s right? Yeah. Well and that's the thing too is it's like if you would have asked the people of the day what they wanted a better horse like a faster horse or more endurance. But the only thing that changed from their perspective is it went from a quarter million dollars to like 30 grand. It's like,"Okay, now I [laughter] can afford this."So. [clears throat] That's true. Yeah, and then it kind of brings up So, I think there's a guy I've shared stuff with you quite a bit, but Kyle Body, he's on the baseball side of of the world, but he does a ton with all different kind of technology and how they use it. It's called Drive Line Baseball or whatever. But, he's been helping out Great follow, by the way. He puts out a lot of really cool stuff. Yeah, and now he's I mean, whether you nerd out on baseball or just tech I mean, he he knows a lot. But he's... I think they had a Wall Street Journal article that came out 'cuz he's consulting with the Red Sox, and, um, I think they had a little blurb in there about him. But his his whole thing was like with, with his AI stuff like you, kind of to your point about people would want to buy a better horse or whatever but he's like,"You A lot of people don't have the creativity to like think how to use this."You have to show them like the value. Like, and if you can just Sometimes you can't sit there and talk about it, you just got to go do it, and you got to show them, and then like then people will understand the value." But, it takes a visionary, you know, visionary too to just like that can really understand that there's value there, but you know, sometimes you just got to like Look, here's something I did, but you got to show And people can see what the value is. Otherwise, they're going to be like, "Oh, well, I don't need that." No, that's a huge just product startup company hack in general that I've really learned working here, honestly, is that you can talk about it all day long, but showing a customer, even if a crappy POC, is way better than just trying to describe it on a PowerPoint slide Yeah. in most cases. is worth a a thousand slide decks, for sure. Absolutely. So, um you're coming I guess we can kind of jump right into it. Yeah, let's let's introduce Josh and And we've got Josh Poliet uh from his Act Solutions and Consulting. and Consulting. Um, and yeah, again, you guys do I know a lot with, you know, Microsoft Power Platform and, and you know, different solutions, but I'm sure across, across the board. But I mean, what, what's been your experience, say, with you know, whether it's AI or different things, but on that kind of like getting customers or people to understand the value of certain solutions that you either have or that you, that you think could be useful to them. Yeah, you're uniquely suited to answer this because you know, the the managers are like, "Ah, well, another data solution. Like, what do I need that for?" But, you guys are actually doing a lot of really good like tangible stuff, right? Like that Well, I think it goes back to what we were talking about. It's all about the value add, right? And so, start with a business problem, and um people and IT really struggle with change management, and it's a huge pain point and struggles to get the value. But, when you change something that somebody wants and like they're complaining about, then change Yeah, there's still some change management. It's going to take a little time and effort to get it over the hump. But, then they're going to use it cuz it solves a problem. Yeah. And so, we do both the strategy side um and then kind of the brainstorming as well as the execution. And so, we start on the business side and and do a brainstorming workshop and really consolidate the the you know, all the pain points. Um and I think one of the things that I found that a lot of people struggle with is they want to go pick a single pain point and just solve that. When you build that portfolio of pain points, you can really take a step back and say, "Okay, you know, over the next 6 12 months, here's the combination of things that we can do that solves this one 100%, that one 30%." Um and you can really build some momentum around a a transformation program pretty So, as far as like say working at you know, working into a company and kind of trying to create that road map or the um or where where do you find like you get the most traction? Cuz, I mean, sometimes you hear like, you know, the best ideas come from the bottom up. But then, sometimes again, like the you really need buy-in from the top to, you know, to kind of really push things out and make. Or, I mean, again, sometimes there's people in the in the middle. You know, but like where have you found the most? Um I I I think that generally speaking um it's someone that's close enough to the work. So, I would say director level or VP mhm will a lot of times engage us initially. They're saying, "Hey, but that they're they've been talking with their people and saying,'What are your pain points?'" And everybody's saying the same thing. They're like, "Okay, we need to do something about it." And so, [clears throat] if I was going to give one advice to a business, it's just listen to your people. What What's keeping them from helping achieve the business outcomes on a day-to-day basis? Yeah, enabling your employees to do what you hired them to do is a Yeah. It's a oddly hard thing for most people to to do. And and a lot of lot of times it's driven by finance, right? Is do you have a framework that allows you to give credit for productivity? Um and some companies really struggle with that. They want cold hard cash um kind of impact or get out of here kind of thing. Yeah. So, how do you mean it mitigate that on the consulting data side of things? Because I I'm sure there's plenty of people listening who are in these companies trying to implement change or want to have all these problems that they want to fix, but they can't figure out either how to sell it to management or you know how to frame it up right even. I think in my opinion that's a lot of the times the problem, right? It's like engineers are not we're not very good at selling things. We've never been in sales most of the time. And so we we're not very good at communicating effectively on how to on a value proposition generally speaking. And so I think that also happens a lot too, right? Where it's just like yeah, there is a problem, but they can't communicate the value or the relative impact that the solution to that problem might have. [snorts] And my experience has been the closer you can get to the bottom line, the easier sell it is. Um but also don't overbake it. Just figure out what your value case is that kind of justifies it, but don't go go overdo it. And so a lot of times in our business we're making pretty heavy risk-based decisions, right? And so if you're a production engineer designing workovers, you inevitably um are making intentional decisions on how best to um manage the risk that go on during an operation. And so if you can be higher consistency, higher quality um and how you develop those procedures, then you can deliver consistently um and that mitigates risk. And ultimately, if you don't drop a tool in the hole, that saves you cost. Um you can get revenue back up. And so I it's just creating that kind of um link of if we do X, which really drives a decision then and that decision has consistent higher quality outcomes, then here's how it's going to impact. And so there's that loose relationship. Um but where I think it early on when I was leading this at a Fortune 500 company, where we struggled is we tried to tell everything, right? Oh, it'll do that, and it'll also do it save time, and then, you know, this will help our culture, and here's all the ways it'll help our culture. And it's just like, just do the least, and Yeah. there's a lot of questions about that. Just risk it. Say like, you know, if if we only achieve 5%of this value, still lucrative. And usually that's how it is with these things. It's it's very, very attractive with compounding impacts. You don't even have to include Yeah. Yeah. I think that's the big thing, too, right? Is it's like, you have a bunch of these little problems, but a lot of them end up having the same root of like, well, if we had all of our data structured in a database, we wouldn't have all these issues. Or if we didn't have data coming in from 20 different vendors in 20 different formats, we wouldn't have all these issues. And it's like, that might end up causing a dozen different problems. And so, you know, if you just present it as, oh, well, this is one little problem, but finding the root, to your point, it's like, oh, well, [ __ ] all of these problems come from one thing. So, it's actually way more impact like, these things compound, and I feel like that's the hardest thing to communicate most of the time, right? Especially to management of like, if we do this, then this can happen, which means this, and this, and this is really all the root of a bunch of different problems, not just one little thing. And so, being able to find those compounding problems and point them out, I think is a normally a really it's a faster path normally to being able to get And it brings me back to kind of building those list of pain points and that portfolio of opportunities, cuz a lot of times if you go solve one, tends to be the sexiest one, but it might be the most complex, and you need things from other thing like, three other data sources. So, just go solve those first, but if you have that road map, you know that you're solving 10 other things partially, and it's an enabling further things along. So, tends to be the less sexy path, [clears throat] but it builds program momentum, and allows people to really see the art of the possible. Yeah. Well, I think that, too, and then also, yeah, when you build that road map, maybe you can shuffle things around and realize that there's a dependency Yeah. That this one needs to come first. And then like if you really want to get here I got to do this first if you want to do it right. Yeah, we can we can skip a step and do that but then it's going to cause other like second order consequences. Yeah. One thing you hit on and someone had a thing of LinkedIn post about this past week and I'd like to get your thoughts is they basically said, like, time is not a value proposition, like saving time is not a value proposition. And whether you agree or disagree, I think it's thought provoking, cuz in the end, like we always talk, time is money, but like... Yeah. And I've fallen in that trap too, but it can be really hard to quantify, like, actually how much time you're saving. And then, you know, now this you know, if I'm saving 5 hours per person times their hourly rate, probably now it's this, but I, maybe you can kind of get to that, but, like, it's almost, I think, you need to find more concrete business, you know, impact than just, like, "it's going to save people time." I think it's it's an easy one to kind of throw out there and then no one can really validate it necessarily or not if they if they haven't really outlined how much time is taking to do these processes like So I heard maybe two or three kind of challenges with that approach. I think I read that LinkedIn post so what was what were they saying you should do instead? Again I I don't remember exactly I I started getting into the comments you know a little bit but Fair enough yeah. So so to me there's a few ways to approach it. First one is: take a baseline, so usually that's a survey how much are you spending on this process? And then when you have enough people doing it, which often is the case, you just take one that's after the fact and. that you know everybody's going to have a distribution of what their answers are, but you at least see that the mean shift. So that's a really effective way to demonstrate that and the other piece is just general job satisfaction. No one likes to hit control C control V so if you [laughter] Fair enough. They they think they want they like it until they realize there's much better ways. Yeah. One and so I think that um a lot of times though where the other value prop is is that consistency or quality and it tends to be the elephant in the room. No one wants to say I'm doing a bad job um which you aren't you're just human and so if you allow machines to help with the low level stuff that you're more likely to make human mistakes on then they it just drives higher consistency. Yeah, computers don't have fat fingers. Yeah. Or or need to sleep or never hung over so. Yeah. [laughter] Yeah, they make I mean they even when they do mess up, they will mess up in the same way. So like I you know I got to like at least you program it and you weed it out, right? Once you find the one area you're like, oh okay, now I can find all the rest of them cuz they're all going to be relatively the same. Yeah. That's that's one of the things, like, I've been using language models for, for like data extraction and stuff. And it's like, yeah, it's not perfect by any means, but I was showing you earlier: like I can feed a bunch of docs in and extract the metadata from those docs without me having to manually open them and go through them and find the title, author, publish date, all of this stuff. And it's like some of them for whatever reason, the first time I did it, it couldn't it's like it didn't read the PDF or OCR the PDF. And [snorts] so it just made up a title and an author, but it made up the same title and author every single time it had [laughter] that issue. So once I noticed that first one I was like, oh okay, sort filter, okay, there they are. Manually do those and you know, it was you know, an 85%solution which Yeah. beat the hell out of doing it by hand. Of course, yeah. Well, and and I think the the interesting thing and where I'm seeing more and more successful application of LLM is a very value-driven approach like I need an output, how can I get this? Let me use the tools at my at my discretion, but Let's say it was 60%. We're having issues, what you possibly could have done was well, but 95%get the the titles right. And so you could extract those and then go find a database to extract the the rest of the metadata. Um, and so it's just having that, you know I'm an engineer by training, mechanical engineer, and so my approach[] everything is: How do I apply this technology, not how do I create new technology? And it's, "Let's just get the most efficient highest, value answer quick to, in place, so we can create value." So let's talk about like how [clears throat] you and or your company is applying AI. Like I mean how are you or LLM's? Like, are you using them? How are you using them personally? I mean, I'm sure there's probably personal use cases that you're using them. But, then also you know, sorry, excuse me. Say through like the Power Platform or Microsoft, like, have you been able to utilize, you know, stuff out of whether Azure Cognitive Services and so on? Like, and like, how's that worked kind of applying those to the business? Uh so, me personally [clears throat], uh I'm I'm ChatGPT. Just go straight there for any of the um I'll call it the creative things that I enjoy doing, but is usually you know, you have that writer's block when you sit down to write either a blog post or, you know, long-form content of any sort. And so, go there to just kickstart it. Mhm. Um but, uh for one prospective client, um they were looking at a few different options to solve a particular problem. Um and I just went to ChatGPT and said, you know, what are the top five or top 10 best-in-class softwares that for this kind of use case? And we were able to find some additional ones. And Nice. I just said, like, "Hey, if don't use us if um if you can find a third-party software." So, I just sent that back to them and let them run with it. And so, and so, um use it for that kind of um Google Plus, if you will, right? And so, I'm going to go to Google to do something, might as well cut out the going to 10 websites, extract all the details. Yeah. Yeah. Yeah. Yeah, and that. And just get the curated answer. Um but, looking for more and more ways um on the functional side. So, business owner, right? So, you spend a lot of time on the HR, um the finance, and marketing, and sales side. And so, um it's really lifting a lot of that creativity and and mundane things from doing that. Yeah. Yeah. As far as application with clients, very value-driven. And and this is probably not popular in tech, but my my entire tech strategy is let's on the simplest technology that achieves the value without sacrificing the long-term thing, um or the long-term value of it, um or long-term vision, sorry. And so, we've used it very little at this point. We help people move from just spreadsheets in a very manually driven process to an automated process and stewardship of people through that process. Um but, what we're we're starting to play with a lot more is bolting on the co-pilots or virtual agents, if you will, um at the end and so you can just kind of interact with your data in more meaningful way. So, instead of opening up a dashboard every day at 8:00 a.m. to say, "Which projects do I need to go focus on today?" You just pop in a quick chat in Teams or whatever um and ask the bot that and it'll cure curate that list for you. Nice. Are y'all using any or have you had any experience using any of the Azure AI tools platform or any of that stuff and how has it been? Um we played around with it. Um we use Azure portal for a lot of um hosting services um as a back end to some of the Microsoft power platform. Specifically on the AI, um I was telling you earlier on one of the use cases um we had was around capturing um French and then translating that into a transcript and then translating that into English and giving people a kind of a history of um of different use cases depending on the language it is in. It took a day. Um so, this stuff is like insanely fast when you adopt that kind of pre-packaged, ready to go technology um and have the tools to connect everything together. You talked about inverting the databases earlier. Same thing's happening with the tools. It's pick your kind of central tool and then just go find the right tool for um a task X and then Yeah. be done with it when you're done with that task and go find the next tool for task Y. Yeah, that's the the good and the bad thing about all this all the AI open source stuff, right? It's like there are so many tools which makes it really nice, but then it's also overwhelming cuz it's like there's always a new one or there's they constantly change and stuff, but it's uh you're right. It's like you have all the building blocks you need. Right. And from multiple sources in most of most of the cases, so it's just assembling it how you want to assemble it. Yeah. And now like APIs or Zapier and stuff like that, you can just All right. Well, if this works for what I need to do, I'll set up here and I can just call it from you know, another platform. Right. Or power automate. Exactly. Uh but yeah, and and you know, like before the enterprise technology approach and I saw this in in the enterprises was we're going to do a study. We're going to do identification and then we're going to look at all the different options for this and then select it. And what I've done and and I I would recommend to anybody is invert that around. It's here's my core toolkit. If if can I get these to do my use case? If the answer is yes, use that. Then trigger. If the answer is no, then trigger that longer ID cells. Yeah. Even just like the Whisper API. I remember first time I used it and it's like two lines of code. And you're like, holy [ __ ] Like yeah. Two lines of code and you can translate or transcribe literally anything. It was nuts. It's Well, I think like you're saying though too like, you know, trying to stay in a kind of a a narrower tighter like toolkit that you really understand is important. I mean um talking with our IT guy, one of our softwares we've been having just issues with them. Like but like they've got our dev version of our, you know, one of the databases is on like Azure, one's on AWS, another one's on this open metal. And then somehow the VPN stuff broke. But it's like yeah, they can't get like heads or tails and they can't figure out how to do it cuz they're trying to balance like five different platforms. Like if they were just doing it through one platform, they would be the expert on it. They would know how to use, you know, how to set up the subnets or VPNs and stuff through that. But because they're spread so thin across and well, now I got to go figure out how open metals or, you know, one services work. And it's like, no, just be good at, you know, a handful of things and you can probably like I said solve probably 85 90%of the problems with those. Yeah, definitely. Well, and and you you touched on something that I wasn't even talking about is really the talent piece, right? And so when you use that consistent toolkit, you create experts and don't want to have to source other talent or train people up. It just slows you down. So thinking about all those things that create waste or additional kind of lag in a project. Well, and then it's like again on the talent side, fact that you're leaning into like Microsoft. I mean, I mean every kid in America gets taught how to use Microsoft. Yeah. In schools and everything basically at this point. I mean, it's you know, but at a higher level even professionally most people are interacting with at the very least Excel and some other, you know, uh you know, the um SharePoint. Microsoft Office. Yeah, you know, all all that side of it, but um Yeah, so you brought up Power Automate and I think that's something I I definitely to talk about like I have this vision that I'd love to do it. I probably need to do it more just internally. I have the capa-well, the the peo-I have the power to probably do it or put it in place. The capacity is probably com- becoming more and more, you know, available as I have more people coming on. But, um, I'm I'm always astonished like how few people like know how to use like Power Automate or like I mean we we pay for this enterprise license of Microsoft Office, and there's so much untapped value I think that people don't even realize. And it wouldn't be that hard to train business users on how to even use basic Power Automate or those tools. But I remember even telling my CEO at University of Maryland a couple of years ago, I was like: "I was like, this makes me my skin crawl, but like, if we just teach half people how to record a macro," Yeah. there'd be so much more They don't even know that's available to them, you know, like But I mean, there's these free in the sense of we're paying for this like tools that we don't need to go, you know, use XYZ RP, you know, Right. RPA tool like you've got one built into your platform and it's federated and, you know, you can use your uh your authentication everything with it. Yeah. What are some Yeah, what are some interesting use cases you guys have seen on the Power Automate side or some really cool things that y'all have done with it cuz I've got a lot of experience on in Zapier, but it's effectively the same thing. And so like I am a big fan of them especially in smaller companies because they are so like instead of writing an API script and then making it a service, I just click a few buttons and it's done. Right? Like Yeah, and [snorts] so you actually talked about two separate softwares [clears throat] there and it gets confusing cuz they're both called Power Automate. But Power Automate Desktop is the RPA version and that's for recording a macro on your desktop and I I haven't played with them much yet, but some of the AI tools they put on top of that is just ungodly and so then I was I was thinking more about the one in in the cloud. I mean, I guess something like that would be but yeah, since of I knew there were desktop flows or whatever you could do. So then um what what you're talking about, John, is the the cloud-based and and it's exactly what that what you're talking about, right? So if you um if you need to move data from A to B, that's the most common and widely used um case. And so a lot of times it's just somebody sent me an Excel extract the data I want to put it in a database. Um but what you can do a lot more with that is um once you're using third-party softwares, then you can apply that in in a meaningful way. And so we actually um have an API into ArcGIS, a really common um mapping software in oil and gas. And I don't know if you've ever used it, but it's it's definitely feels like it was built 20 years ago, which it was. Um super powerful, but just really difficult to use. And and and costly resources to to to hire to to do that. And so um we we we tunnel in through that API, and we actually compute um anywhere in in the world if you spilled a drop of oil and it ran to a river and all the way to an ocean, we can plot that whole path up. Um and then within Power Automate, we're realizing sometimes ArcGIS was sending that back inverted, so the path was starting at the ocean and ending at the other piece. And so we initially were trying to brute force think about, okay, what's the best way to do this? But eventually we just said, what's the first point? What's the last point? How close is it to my origin? Um and then determining that that's the starting point. And so we're able to do all that in Power Automate and and put that into um a database and put Power BI on that. And now we have a visual [snorts] for the users when they drop in a single point of hey, where would my my oil go on this? Now we can tell them exactly where it is and plot it up on a map. And it happens in 30 minutes uh for 100 sites, which would have taken weeks for people to do previously. That's awesome. Yeah. Yeah, no, they're [clears throat] they're incredibly powerful, especially for like SMEs or end users and things like that who understand the problem or like I just need this data set but it's over here and IT's blocking me from it. [laughter] It's like I I can just set up a uh an automation to move that over to me. But I mean that's one of the biggest recurring problems in our industry is just data siloing or you know having to maneuver around IT instead of with IT a lot of the times and so yeah these the crazy part is how many integrations they have especially Power Automate. I mean like there's so many well beyond just you know the Microsoft ecosystem too that it's like but yeah I mean I I was was looking at using that when we were working together for like our marketing stuff. I set up a a form that then went into a Trello board for like marketing post ideas and all this stuff and it's like oh this is awesome you know super easy workflow. Well, and they make it super easy to build custom connectors too. And so, another use case was with the EPA, like a client that needed EPA data. And so before, it was, you know, "Here's exactly all the details of my permit, here's the limits I have." But now, you just pop in your permit number. And like, it wasn't built into Power Automate, but we have the custom connector and we just go pull that out and it's 100%accurate; there's no fat finger risk. Your permit gets updated, it's all reflected. And so, pretty, pretty powerful to do. You know, it takes two or three days to build that custom connector, so really quick to do some stuff now. And and I guess I technically was but wasn't talking about like desktop flow but I'm actually really interested to hear more about that. I mean like say you're saying like that they have some AI stuff you can put on top of it and it those can run on your desktop and run kind of little things that internally Yeah so I think the best analog is like Excel recording a macro but instead of just Excel it can be your entire desktop. It's not traditionally called RPA, but I think some of the new features they're releasing around some of the co-pilot stuff is going to be much more adaptable and. And the user can, like, literally just narrate, I believe, and say, "Here's what I want it to do," and then it'll go draft some of that stuff. And instead of having to say, you know, "Make my Chrome window 800 pixels wide and go to position, you know, XY and, Z, that it the,that" it's using contextual data with AI to know that I need to just go find the title and then, you know, go the next row down or things of that nature. And so it's what much more robust and consistent cuz before RPA wasn't was rule-based and when you pair that with statistics, doing both rule-based and statistics is crazy powerful. Sure. Yeah, no, I think I'm a big I'm super bullish on these multimodal models and their ability to automate a lot of this stuff. Just Or like even I I don't see a world where OCR isn't replaced by multimodal in the next year or two just because it's so damn good. I was playing around with extracting data from drilling reports earlier this week and I was playing around with some of the multimodal models and it's like holy [ __ ] man. This is give it this shitty PDF report and that's all tables and you know, fields and and tabular data and being able to have it find exactly what I want in a repeatable fashion and then output that into a JSON blob. It's like man, and I didn't have to write a single piece of code and I don't have to worry about the OCR messing up or recognizing anything. It's it's pretty wild. And that's just on documents, right? Like then it becomes this whole other like level of thinking when it's like, oh, I can do this on my entire desktop of like Yeah. Man, how much stuff do we do on a daily basis where it's like I could just automate Even now as I'm like just talking you know, with the GPT-4 Omni you can feed like video into it now and everything too. It's just like it's going to be ridiculous. Well, and I think you were talking before if you know, how do we coach people on how to use whether it's AI, large language models or even the lower level technology and um you know, I'm still experimenting with some of it and the use cases are growing, right? Like with the fast evolution of the technology. But one of the things that I actually stood up back in 2019 when I was leading digital transformation and I've seen successfully deployed and we've my company's helped helped do it is a citizen development program and it's really creating the boundaries with an IT. So, instead of working around them and and under the shadow IT, if you will, you're working within those confines, and through that you can help not just educate the how, but also. The what, and and really help people bridge that gap of, you know, "I saw this use case; how is that comparable to my job, and how could I apply a adjacent use case?" Yeah. That's kind of what we were talking about getting along like in some of the enablement like Yeah, empower. Yeah. Well, and it's both, right? And so to me enable is just give me time to go do some of this stuff and power is giving the tools and and the the authority and the support to go do it as well. Yeah, no that's uh I mean that's that's the whole kind of uh value prop of all most of the low-code stuff, right? Is it's like you don't need you can live within IT's boundaries, you know, but you can enable the SMEs who understand the root of the problem the ability to solve that problem or get the data that, it, whatever that may be, um, without them violating any of your IT protocols or things like that, so to speak. But it's also [snorts] just one of those things that, like you know, the IT side doesn't there's no way they can understand all of the problems of every user in a company all of the time. And so, it's like the IT guy should be doing the stuff that IT guys know how to do and are good at that the average person doesn't, and the people, the SMEs in the business units, should be doing the things that they're good at, not IT. And so, this enables that kind of on both sides. And so, like I said, I've been we've been using I've been using Zapier and all these no-code tools for a long time, and they just keep getting better. Most of the time Zapier has gotten gotten really shitty lately just user interface-wise they just keep shoving stuff into it and it keeps making it slower. It's just like there's no reason. Like I come here to make this fast not because I want to sit and wait for a page to load. Um Uh to go back to your point around uh people creating uh like citizen development know their own process best. I think one of the recognitions I had as I was really strategizing in digital transformation was that people been creating applications since the dawn of Excel, right? They just happen to be shoving it all in there and some VBA macro. Yeah, I used to be called Dr. Excel cuz I would create my own little nasty ones of all sorts of stuff going on in there and even without macros. You just You effectively say here's what my process is and you put it in a really logical sequential fashion so you can put data in and get what you need out without a bunch of kind of extra QA every time. And so, all I'm seriously seeing what we're doing a lot not, not all the cases, but a lot of them is to, to help people move from 40 different people having 10 different spreadsheets, which is 400 spreadsheets, um, to creating a single database and a single interface. Um, but what's really makes it super easy for us is they've effectively documented and prototyped the process and the solution, right? So we can now just use a combination of Power Apps for interacting with your data, Power Automate to retrieve it, and then Power BI to report and visualize it, um which is the three capabilities of Excel previously, right? You got the graphs, you got the data storage, you got the calculators. Um and so they've just kind of broken the the core functionality of Excel into those three tools. Yeah. Well, and that's the the other thing too is um giving the I just completely lost what I was going to say, but [snorts] the uh
We'll move on cuz that will come back at the very end of this, but So maybe if I can steer in one direction, we talked data lakes before a little bit. And I have a different philosophy on data lakes. Um there you know, there's that big kind of enterprise for the highly consistently used data, but I think the other piece is that it's the Excel data, right? And and when you combine the SAP or the Oracle data with Excel, when people start doing a lot of the copy and paste. And so people dream of oh, I'd love to you move all of my Excel data into the cloud so that other people can access to it. But I see it from a value proposition perspective. You could move it into the cloud, or you could just take it one step further, put it in something as simple as a SharePoint list. Say, "Here's where our single source of truth is." People go and update there and pull from there. And that achieves 80%of the value. Yeah. Yeah. from a a data lake. Where I think the true data lake, where you contextualize all your different data sources valuable, is that that true ERP or enterprise data. Where everybody in the company is trying to ping that you don't A load up the the database itself. I mean, do you need that relationships between that? And then everybody else just uses those relationships to kind of marry their data to that. Yeah. Uh I was I remembered what I was going to say, but the low-code stuff, in my mind, to your point of everybody's been building software basically from the beginning. It's this like low-code is the perfect way to test out whether it's worth building a a full app or workflow around, right? And it's a very quick and easy way to do that. That's the last point I was going to make on that, but Yeah. I I would maybe challenge that, actually, cuz I think that where Microsoft's gotten with some of these tools is that they can be your production instance. that's absolutely. And it's that same kind of strategy around, you know, what's the lowest the simplest technology. And when it comes to UI, you don't need to go build a a React front end or C#or anything. Just build it in Power Apps. But, where you need to plug in some additional tools, some AI, just interact through the database or directly through Power Automate with those tools. And we've created some very sophisticated applications um and with a lot of microservices within our toolkit. And It's very fast and very powerful. awesome. Yeah. So, you were talking [clears throat] you brought up stuff like SharePoint lists. Um And let's talk about some of that with like the interaction with Power Apps like in like some of the back ends that you have available. I mean, Power Apps can connect to almost kind of any database behind it. But, then there's also that makes once even a SQL Server doesn't make like a premium you know, app or whatever. And then like you have to pay per per user that's using that app. But, um we've done a couple things where we're writing right to like SharePoint lists and then that makes it actually free Mhm. at that point and um you know IT workaround. Well, what's [laughter] I where but where have you seen but where have you seen like the uh is there an inflection point where like it makes more sense to do it in database versus like a SharePoint list or Usually I say the threshold is 100,000 records. Mhm. If you're going to have a million records in a single table, then don't use SharePoint. Yeah. Um but I I think we're um it really depends on the magnitude of the users as well. Um if you're going to have five or 10 users maintaining, you know, 1,000 or 2,000 record, then yeah, you use SharePoint. Yeah. Um but if you're effectively creating an enterprise application, you just want to do it for a 10th of the cost and a 10th of the time, then go use the the dataverse or the sequel or whatever whatever Yeah. is your database of choice. Um and you can build your business rules in there and um we have some clever ways of doing SharePoint relationships just through IDs and whatnot. Um but you can actually put in database relationships into dataverse like you would a a normal sequel table. Mhm. And so that allows you to really drive much better data consistency um and do a lot more efficient development um by by leveraging a lot of those business rules and relationships and whatnot. Um and so we we tend to do more and more now with dataverse. That's fair. The the threshold we have to get through is the the the pay per user, but Yeah. if you're saving 100K 100,000 on the software. [laughter] and on the development cuz we're you know, we're replacing six-figure softwares probably Yeah. once once or twice a quarter across our portfolio of clients and so um and you know, you're using 90%or 10 [snorts]%of the software. You put data in, you got to report out. Yeah. Sure they had some other bells and whistles. You didn't need it and so let's just replace that with a database. Cool. Now that's uh Yeah, these things are so powerful. [laughter] At the end of the day, they really are. And you know, I I came to it early in 2019 we're using this in the enterprise and it's just a few years before that. Um, but it's really about that clever application of technology. And like 99% of Power Apps I see is a single, maybe two screens and like a form, and it's like, okay, we're actually building rich interactive apps. And so, I have to apply that stigma a little bit at times, but as soon as I, you know, give that demo, people get pretty excited. Yeah, it can be as simple or as complicated as you want to make it. That's the beauty of them. It give us guardrails, right? And so, you know, sometimes we'll build something, we try to shove too much into one screen or connect too much to it, and we just, you know, prototype and find other ways to approach it and. And so, we think we had [snorts] like probably 20 layers in one, and it was a combination of like seeing all the data and then also editing the data, and so we just split that into of: see the data and edit the data. Right. Um that drastically improved performance. Can you help me and probably people listening grok what is the data verse cuz it seems it's very like broad like I think when I think about it or I've you know I've interacted with it I know I have but it's like kind of like Probably everybody you ask has a slightly different definition but I just see it as an Azure SQL table and so they've created an abstraction layer between traditional database development and citizen developers. So now it feels as simple as just adding columns in an Excel or SharePoint list but you can start to use the more and more advanced features over time like the relationships. But you are actually using Azure SQL and so the performance of the cloud based and retrieving I think the most we've shoved into a database is 1.5 million records but no issues at all with that and you just keep going. Nice. And I'm not, you know, I think it's all intertwined there too. But how much have you done or interacted with Dynamics? Cuz, again, even that, like, at first I thought it was a CRM. But then like, but then I hear about [laughter], you know but then, but then there's like, no, but then there's massive, like, I mean, like, I think I'm pretty sure Oventus uses it as their ERP. Yeah. And it's like and it's an option like it's like I check in on it every kind of 18 months if you will and you're 100%right like 5 years ago, it's like super clunky, terrible UI, like, why would I ever use this? Microsoft in general, right? Like, Dynamics sucked, and now Teams, you know. Yeah, prototyping [clears throat], right? They've got that agile mindset. Um, but, yeah, I'm seeing the same thing and. And both in the SMB space, where small and medium businesses are using their entire, um, warehousing, inventory, financials, all in Dynamics, and seeing more and more of it used at the enterprise level, and it's funny what people end up picking. Like, there's one major, um, midstream operator pick it up just for CRM. It's like, who are your CRMs that you have an interact with? But, seem like more like a contact list, but not not neither here nor there, but, um, certainly seen that capability build, for sure. The challenge with a lot of these ERPs, though, is you're so ingrained in your day-to-day business that people are are really hesitant to make a jump. Yeah, that's a hard software sell. Yeah. People spend billions of dollars on SAP over, you know, a decade, and it's like, how do you spend this much on a single software application? So, you won't find me, um, giving any sympathy towards SAP. No, it doesn't. Or any votes for it, but, um, it has a time and a place, but people have tried to use it in way too many, uh, ways, and they've tried to build that that vision of a single database. Yeah, yeah. Well, SAP, to their credit, has made it very configurable, and again, then no two implementations are the same, so, yeah, that's They've made a lot of consultants very rich, too. Definitely. That's the the nuance of it, right? Is like, if you're trying to be everything to everybody, you're nothing to nobody until you have the custom development for your thing. But, but once you're there, you're Right. But, when you're in there, good luck getting out of it. It is a sticky, sticky thing. [laughter] Yeah, yeah. No, those are ERPs are tricky, man. It's, I mean, Bobby and I used a low-code ERP for at our prior company, and I they've gotten even better. It was QuickBase. Um, but it's like, having that flexibility of like, oh, I need to add this entirely new process to my company, yeah, is really nice. But yeah, I I'm sure Microsoft has a lot to make it nice and integrated across everything and then power apps on top of that. So Well, the problem is the user of this app I'm calling say, "Why can't I do this here?" I mean, like what Salesforce was a CRM and now it you know does all these things and QuickBase. But again, like once again, it's all might come back it's a database, right? Like I'm just Yeah, they're just relational databases. Yeah, you can create these relations and stuff in the background and then why can't I do whatever the hell I want my business in this? Right. Well, and it it goes back to that conversation where the decision tree now is what can I use in my Microsoft stack that I have and then go look elsewhere and the answer is quite a bit. Yeah. And more and more so, so it's a very powerful. What's a uh What's the technology in the power platform that people are sleeping on? That's a good question. Define sleeping on. Elaborate what you mean on that. Like Not really being used or Not really being used or not really being talked about or talked about powerful. Um
It's probably for me it's just the application of the technologies, right? I think the the technologies are not terribly complicated. It's it's more and people are still using a a power app just to as a form into a database. They're not using it as a rich application. Using power automate for moving data between two databases and and doing a little post processing or that ETL if you will. They're not cleverly combining two or three data sources and helping extract new insights. Power BI Um we haven't really talked about that much, but Yeah. Um that was probably the highly most adopted by citizen developers. It's simple. It feels like Excel. It's drag and drop. Um but people are using it um to just report out on on numbers. They're not using it for true analytics and I don't even mean like put machine learning in there. It's just Like Right. Know what your question is [laughter] and put the only visuals to help you answer that question versus just like a bunch of cards and stats. Yeah. Yeah. Um but then of course, once you know what your question is and you're able to prototype what you're looking for then it's easier to start to layer in more and more of the automated analytics on top Okay, so I'm bringing up Power BI cuz yeah, again, I don't know why I didn't think about that. Yeah, I I think about, you know, at least at least through my conversations, like, you mean, I'm clearly you're using Power BI, but I feel like a lot of it is more like the processing, like getting data into it and base. But, um, how much, where, where do you see him say with Power BI versus Spotfire and adoption? I mean I mean are you seeing it going that way or are people still pretty sticking to Spotfire or is it both? Yeah, so back in 2019 I actually ran a study to figure out for for my digital transformation which one I want to use. And so Spotfire was the native one the reservoir engineers and and production engineers were most comfortable with that. Um but what we found was Spotfire is much better for um more complex larger data sets and Power BI is better for someone that's used to Excel and needs to make that jump. Um and so if like Power BI has more and more of this but at the time it was if you want to go write some Python code to help process your data use Spotfire. Um but you know, as a small business owner I I looked at it and Power BI is free until you want to publish and then it's only $10 a month um and from going from no Power BI instance to having one is 10 minutes if you have a Microsoft account. And so whereas Spotfire you know, originally standing up your own nodes and now it's the cloud-based stuff and so um it's it's just there's a big hurdle to go and do that for both the users and from an IT perspective. And obviously Microsoft's really trying to um ex-expand and extend that capability with you brought up Fabric and so some of the cool things we're doing to do kind of a pseudo data lake if you will is create data flows and and that's just a way to like publish a data set then other people can hit whether that's with Power Automate or with Power BIs. And now they've got some data marts which um we've only played with a little bit but Yeah, those are Azure SQL back end, right? Exactly, yeah. And you can actually hit those data marts from Spotfire as well. Exactly, yeah. And so um they're really broadening out the capabilities within their ecosystem. And so I would generally say like unless you're just an expert Spotfire user, if you're starting from scratch, do Power BI. And then if you hit those ceilings somewhere, particularly on data density or something around that, then maybe look at alternative solutions. Yeah, that's my I've got a very tend to agree with that experience. I think engineers like Spotfire because we're used to looking at frac stages and drilling report. We're look We're used to looking at high-density time series data. And that's honestly my only complaint about Power BI is they don't have a great way to do that. But then there's really an honest conversation of should Power BI have that natively at least [clears throat] capability or should it not because it is just it's a BI tool. It's not necessarily like a or at least that's what they Yeah. Branded as, but it's like instead of like you can do that in Power BI, but I'm not have to. Now I'm have to write my own Plotly dashboard in Python to get it to look how I want it instead of dealing with all the crap that Power BI like all that like, uh, the [snorts] concatenation and categorization and stuff that they kind of force you into with time series data and stuff like that. But, um, I still I completely agree with you, especially if you've used Microsoft products at all. [laughter] Power BI is a very easy uh intro, especially comparatively to like Spotfire, right? Like Spotfire can do a lot, but it's also harder to I think the hardest part is going from Spotfire to Power BI because you expect all the stuff Yeah, with all those tools, yeah. Even compared to Excel, there's less like customization it seems like at first glance in Power BI. And it's gotten a lot better. I mean um I mean like like we said earlier, everything Microsoft they continually improve the products. Yeah, they get to where they're yeah, I'm definitely world class or up there, but um, but I guess it's I mean, that's kind of my understanding of Fabric right now, is it is kind of Fabric is enabling like that more analyst type person to start doing some of the data engineering and everything, right? I mean, like, it's [clears throat] moving back up up the stream. And I mean, is it at the point now? I mean, I guess Power BI datasets or semantic models, they call them now are like...but or is it data flows? Like, so are those able to be authored, say, with like Power Query? Or you know, cuz I mean, I I I I think that's that'd be a huge thing if you were able to kind of author a an end-to-end, you know, ETL essentially, then it landed in a like a data mart or like a database, but you were able to do that ETL with like Power Query, I think, would be super powerful. Yeah, and it is. Um that's exactly what you can do and sometimes [clears throat] you can build it. They've got a lot of pre-configured steps so it's just like, you know, dedupe my data, um sort it by this, and change the data type, and things of that nature. You can pop open the advanced editor, and you can just write all that code. And so, you know, I, I've seen, um, a really strong bias from traditional full-stack developers against anything low-code. But, you know, we're, we're trending towards what I would describe as fusion development. And so, very similar to a lot of times, if, if you're a full-stack developer and you're not learning React, you're becoming obsolete, then I'd say it's the same thing for low-code, cuz the, the guy that is using it is going to create higher-quality products faster, um, with a, without it. Yeah. And that's what I've always I mean Power Query is probably my favorite part of the whole platform. You know, it's just, you know, again it it is the drag drop but, it the fact that it creates the code is is a huge...it's always been a big deal for me. Cuz then I get some of that flexibility, like, cuz even, like, sometimes if you want to have a parameter, it's, like, date, you know? Like, you create a filter; well I, I just create the filter with my old drag drop down, but then I can go plug, go into the code and plop in my parameter date into [clears throat] that code that it generated. Or you copy it from one place into another. A lot of times I I need to like transform a dataset and then like collapse it down so I just like have the dataset in one query, copy that code, and and then paste it over here, and then just like collapse it and and transform it. Exactly. Like crazy fast. It's just like every time we do like try something new it's like, "Oh, that was way easier than I thought it was going to be." Well, it makes it easier to learn, too, right? [clears throat] Cuz I'm I'm very visual. I'm a very visual learner and like that's what I liked. It It took a minute to get used to in Power BI because everything is, you know, different than Spotfire, but, um, the fact that you can do everything either through code or through a button like pretty much on everything is incredible because, like, "Hey, I'll do it through the button and then I'll go look at the code so I understand what it's doing," and now I've just learned how to now code that thing without spending a ton of time having to actually figure out how to code it. Right, yeah. And then [snorts] yeah, then you get more advanced and you can start stringing those together into the query and all of that stuff. I feel like you you've gone down this rabbit hole and and I did some time ago as well of pagination in Power BI. So if you're hitting an API call and you're only getting 100 records at a time and you have to paginate through it. That That is the one thing that is like okay, I have to get in the code and get a little bit clever with some of them especially for the APIs that are not as rich in Yeah. I see the pages. My favorite workaround for I've got I've still got that link stored in my I've got a Power BI OneNote notebook of just like random [ __ ] that I have come across that I'm like oh man, I'll use this again a lot. Um, and for the pagination, it's essentially you create you just create a blank table, you use the functions to create, you know, whatever the offset and limits are for your pagination, and then you just loop the uh loop the call; you put a variable in the call that is going in calling that table, and then it just loops through all of. But [laughter] it's the most like backwards way to possibly do that and you could just write a loop that actually does it like it's supposed to. Yeah. We might have [snorts] invested a little bit more time and done the loop, but it it is it takes definitely it was, you know, a 10 YouTube kind of day. Right. [laughter] But that's the cool part about it, right? Is it's like if you want to do the full custom code, you can. If you want to figure out the gooey workaround way to do it, you can also do that. Totally. Yeah. Yeah, one one thing I did Power Query was particularly proud of and it wasn't even that big of an impact, but like our we used DBT for our transformations on top of Snowflake, but they have it generates documentation and they have a GraphQL API that you can actually you know, the same information you can see on their documentation website it generates so you can go get it from their graph QL API. But I mean there's no graph QL API button in in [laughter] Power BI. But again, because like you have the flexibility through that I was basically able to create the the post request within the Power Query and like post that and then I think it needed a bearer token, too, cuz again you can hit some APIs but it's not super robust. But again, basically because you can create like a curl request or whatever Mhm. Yeah. post request from within there I was able to feed all the headers and everything and boom, send it and then it got the response back and then you can parse it from there. Yeah. But it's yeah, I mean it opens up a lot of possibilities. 100%yeah. Especially once you learn how to call APIs within it and you're like, oh man, this is awesome. And my background is mechanical engineer, right? And I did petroleum engineering for a years, but I consider myself a fairly proficient developer. I will say APIs were like one of those steeper learning curves. I can do a lot of it other pretty natively, but once you crack it, it is so powerful. [laughter] Yeah. No, that's the annoying part is it's like I can do it over here in Postman all [clears throat] day long, but getting it into Power BI how I need it Yeah, cuz yeah, Postman has all the different languages but it doesn't give you the nice little here's how you do it in Power BI, right? So, you have to hack together like three different blog posts and a Indian YouTube video and uh Yeah. But you get it there. And then that's why you hire experts for you know, the the niche things that are more complex and you want like to truly productionalize it. Yeah. Yeah, that's where I was going with like the whole prototype stuff. Even if it's not like a fully blown Yeah. full stack custom coded thing. It's like, hey, here's my hacky shitty way of calling this API, make it better, give it to But you can giving something to a developer that's already there and they can use it is like it's another hack cheat code in my opinion like especially for streamlining development, right? And even Excel does a good job at that. And so, a lot of times when, when we're working with customers and trying to understand their process and their data, it's just: "Data dump your all your Excels on me," and we can come back you know, 30-minute discussion, data dump on me and we've got 80%of the use case framed up after that. And so. Yeah. that kind of ironing out the process and the data and whatnot and prototyping it whether it's Excel or otherwise is makes it so much easier to create some really high-quality apps uh quickly. Yeah, for sure. That's yeah. Until you've gone through the development cycle where you go through all of your tasks and items and don't include mock-ups and then get the results back and it's like, "Oh, well, they did the function that you asked them for, but they didn't do it in an intuitive way or put the button in the right spot," or like... Yeah. I will never I try every single time I have any kind of anything that has to do with a front end to have a mock-up with it just because it it will save you so much time and headache. Yeah. All day long. Well, that's usually the first step. It's like, "Hey, we we mapped out your process, we've optimized it. Here's the data tables we're planning on." And then first step of development is there's not a single button here that works [laughter], but here's what I think it should look like. And usually have that 10%changes, but by just ironing that out your development cycles thereafter are so much faster. Yeah, yeah, big time. Damn, we're already over an hour. That Yeah. [laughter] Always does. Yeah, I know. Well, it's it's also been a minute. Thanks for everybody for dealing with us for travels in summer. We missed an episode last week, but [snorts] it's good to get back in the uh studio, too. Yeah. Yeah, I know you haven't been in person in a minute. Yeah, good to see you, but [snorts] we'll jump into the speed round and really get crazy. Okay? Should I hop out? No, we'll we're we'll just ask you random quick answer questions. Okay. Uh I think I know the answer to this question based off all these conversations, but what's your favorite cloud? [snorts] Yeah, Microsoft. But, it goes back to what we were saying that it's use what you know, right? And there's not a perfect one out there. They have pros and cons. Yeah. Yeah, for sure. Um favorite either like social follow like just someone interesting in the during the tech and energy space that you follow that is worth following or um we can do a favorite book or good book recommendations that you've either Well, obviously I follow Colin McLellan. Don't tell him that. He doesn't need any more of that.
Yeah, admit admittedly I probably don't do that much on the social side. Um so and you know, the way I educate myself is I don't like to read just one book over three or four weeks. I like to read 50 articles and then kind of splice together and so I'm a much more I get on socials, but I'm just looking for educational content a lot of times. lurking lurking or lurking, yeah. Yeah. No. That's passive engage Yeah, absolutely. That's what for now forever and now I'm Yeah. I'm a podcaster and then [laughter] Just use chat GPT to generate all your blog posts. Yeah. Um What's your favorite non-Microsoft technology? Hm. That's a good question. My iPhone? [laughter] Fair enough. What about any kind of software or Python libraries or anything like that? Yeah, admittedly chat GPT's really become a critical arsenal. I've been paying the pro version, or premium version, for about a year, 18 months. And you know, I feel like it was a little bit of learning curve, probably even in college, to learn to go to Google every day just to ask it questions and now that was only like when I didn't know something. Now it's more of like not when I have a question, but when I have a task I have to think about oh, can I get chat GPT to do this and talked about there's a learning curve there, but um that's definitely a a critical tool in my arsenal. Yeah, no, that's I had that SPE workshop last week. Shout out to Kyle in the SPE OKC group. That was one of the things that I talked about was like this is an iterative thing. You have to learn how to use it just like you would learn how to Google. It's the same thing, but but also interrupting your just teaching yourself to be like, hey, I want can GPT help me or do this completely? Like just getting that into your brain and starting to think that way it completely change how much you're able to do with it. And just messing around and finding it like Yeah. FAFO, [ __ ] around and find out, and see. And where I started was the task that I tend to avoid, don't like doing. I started just having it do the heavy [laughter] load, and it was solving my pain points. And then I kind of allowed it to creep into other things. And then it's just can do higher quality things faster. I kid you not. And it just keeps getting better. Yeah. I haven't quite gotten to the habit of just talking to it on a daily basis. It's there. But I think it would drive my wife nuts down the hallway, so I avoid doing that one. Colin Colin swears by it. Like he'll go on a walk or a workout or whatever, and he'll just have his AirPods in and just talk to her it to her, whatever its name is. Um fake Scarlett Johansson voice. Uh [laughter] But it's, you know, it logs everything, right? So it's just in your chat. And I was like, damn, that's a crazy way to think about like being able to use this for productivity all of the time. Like record conversations that you're having with people and just transcribe them, right? Just to have that. We were talking about that with a with an operator the other day. Like why aren't why isn't every operator transcribe recording and transcribing all their AARs? Because I guarantee you what's in the PowerPoints of the AARs is not it's It's much more meaningful conversation. information is in that conversation in that room. And it's like now there's no reason we shouldn't be doing that on all of them because the problem is the answers to the to the problems that you're actually trying to solve are in that conversation, not in the deck per se. Like the details about it. Well, and you know, the Collide product, you're allowed to like search videos and things of that nature, right? And so having that could completely contextualize you have a future questions that, you know, tell me about a previous after action review that you had this scenario. So you can go back and learn from that as you're planning the next job. Exactly. So, yeah, that'd be crazy powerful. Yeah. Well, cuz there's just so much, I mean there's so much information that gets passed around at the field level, right? Like with consultants, with drilling consultants, with whoever of like, "Oh, yeah, at this depth you're going to get this gas kick, so be careful. This is how we mitigate that." Same thing on the frac side, right? But it's like none of that ever gets written down anywhere. where everyone knows where to find it. Where it's useful, at least. Yeah, or yeah, don't even know where to look. [snorts] Um we got one more. Let's do um favorite board or video game. Uh lately I've been mostly playing Destiny 2. It's like um like for the last 5 years. It's the only game I play on my Xbox. It's like, uh got a little bit of time, need to recharge the batteries, and so I'll go play that. But all [snorts] time, easily Halo 3. Um it was like definitely my college one. And it was the one that I I, you know, I could have some fun with some of my buddies in college, but they would all get so mad at me because it'd be like 25 to 4, and it's Josh versus [laughter] three other people. And [snorts] uh they're like, "There's no way you had that many beers." But whatever. [laughter] Pretty good on that. What's your weapon of choice? The battle rifle or the sniper? Or Uh definitely the battle rifle with the um the rechargeable little I can't remember what it's called now. It's been so long since I [laughter] Strip your shield with a little um gun, and then uh a single battle shot rifle right into the head. Whenever I my roommate in college turned me on to the whole like battle rifle head shot, and it's like, "Oh, [ __ ] that's why I was dying so fast." [laughter] Like that was such like a life hack when you're playing that game. It was like same thing. And it had like an auto finding on like for the actual head shot. Like it gravitated towards it, so you could I could just sweep my whole my cursor [laughter] across the screen, and like pull the trigger at approximately the right time, and head shot. Close enough. That's awesome. Cool, man. Well, thanks for coming. It's been a a pleasure. Where can people How do people find you, get in touch with you if they want to Um easiest thing is our website. Um it's ascension-team.com. We've been going through a rebranding, so called Ax Solutions and Consulting now. Um realize that, you know, we're not just consultants, and we're not just selling a PowerPoint deck. We're great value, and so we wanted to put the solutions in there, but uh Nice. or look us up me or the us up on LinkedIn. Got a pretty sizable following and growing there. [snorts] Great. Spell your last name for everybody. Uh it's like polite but with two T's so p o [snorts] l i t t Cool. Appreciate it man. Yeah. Thanks for coming. Yeah. It was good seeing you. Great.