Energy Bytes

Energy Ogre's CTO, Britany Devlin, shares her expert advice on reducing electricity costs and navigating the complexities of the energy market. Drawing from her extensive experience in trading floors and risk management, Britany explains how fluctuations in natural gas and electricity prices affect consumers and how Energy Ogre helps clients optimize their power plan contracts. She also highlights the transformative role of AI and automation in the energy sector, enhancing customer experience and providing innovative solutions in a volatile market. Learn about the future of renewables, energy storage, and strategic ways to lower your power bills.

Click here to watch a video of this episode.


Join the conversation shaping the future of energy.
Collide is the community where oil & gas professionals connect, share insights, and solve real-world problems together. No noise. No fluff. Just the discussions that move our industry forward.
Apply today at collide.io

00:00 - Intro
01:17 - Toby Rice: Sub $3 Henry Hub Outlook
05:22 - Electricity Costs vs. Natural Gas Prices
07:35 - Introduction to Energy Ogre
13:26 - How Energy Ogre Operates
17:51 - Importance of Customer Service in Energy
23:10 - Overview of the Texas Energy Market
27:19 - Energy Generators Explained
31:24 - Incentives for Homeowners in Energy Markets
32:30 - Understanding the Power Grid Dynamics
40:05 - Energy Ogre Insights
42:40 - Managing Tech Debt and Scaling Strategies
44:30 - Role of AI in Energy Sector
49:20 - Customer Support in Energy Services
52:11 - Griddy's Data Analysis Methods
54:11 - Historical Data Utilization by Griddy
57:50 - AI and Data Generation Impacts
1:02:48 - Learning from Human Decision-Making
1:06:28 - AI for Workplace Productivity Enhancement
1:10:31 - Recommended Books and Podcasts
1:11:15 - Top Social Media Influences
1:11:50 - Future Plans for Energy Ogre
1:13:16 - Finding Energy Ogre Online

https://twitter.com/collide_ai
https://www.tiktok.com/@collide.io
https://www.facebook.com/collide.io
https://www.instagram.com/collide.io
https://www.youtube.com/@collide_io
https://www.linkedin.com/company/collideai

What is Energy Bytes?

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.

Welcome back everybody to uh I'm super excited. I'm uh exhausted. We've been in Calgary all week for Energy Tech night. Big shout out to uh Saga Wisdom and and everybody that uh attended. Appreciate the hat. Yeah. Wish I would have been there. Seemed like a good time. It was a great time. Calgary and and was it Canmore where we went yesterday. We woke up at like 5 yesterday morning and drove out to basically close to B just to see it because we were all there. And absolutely worth it. It's beautiful. Not fun returning to the 90 degree Houston weather today, but um the beltway is beautiful [laughter] intersect over I 10. You know, it is weird. When you fly into Houston, you don't realize how green and how many trees and stuff there are in Houston and the metro area and stuff. And I was like, "Oh, we do have some some nice things down here. Feels like a [snorts] jungle drive here with with Bobby Nean, my my trusty co [snorts] and today I'm super excited. I've been a customer of Energy Ogre for since 2017, I believe, or 18. I'm I guess an OG now, but um OG Ogre. So, yeah, we're here with Britney Develin, CTO at at Energy Ogre. Thanks for joining us. Thanks for having me, guys. I'm excited. Really excited. We've got a fun Bobby found a good kind of relevant news story to that around yesterday or two days ago. uh Sha Lenial, friend of the [snorts] uh friend of Digital Wilder, Toby Rice, um kind of came out and basically said he's bracing for sub $3 Henry Hub for foreseeable future lower for longer. And uh [snorts] I think there's definitely going to be some implications for, you know, oil and gas upstream companies, but even, you know, downstream on the market side. So definitely want to get uh Britney's take on that and you with your energy MBA. [laughter] A lot of a lot of value I provide with that. Yeah. on the yeah I'm like a a fresh out of school MBA when it comes to energy markets [laughter] I learned all this stuff in school I did a hedging program as one of my or strategy as one of my projects and stuff but it's all theoretical so it'll be fun to chat with you about It's been a while since I've been in the in the wholesale side, but it was really fun. I spent most of my career working on trading floors doing risk management, which was really exciting. Um, everybody makes money when the market is volatile, and when it's not mar when it's not volatile and you're staying these super low prices, you have to get really creative with your storage plays and those types of things. Um, but when it is volatile, it's very very exciting place to be. Um, and and a lot of money can be made. Um, how do you how do you what do you feel about the the kind of gas market as a whole? Do you kind of lean into into that side of it where it's going to be this kind of lower for longer thing like we've seen for the last decade? Yeah, I mean basically since I've been in the industry, it's like literally $2, you know, is what you just get used to. It it is kind of what you get used to. I've been I mean, before it was $2 for forever. It there were people saying it was gas was going to go to $12. I remember those days all the time when it was up up and then it just maintained this consistency. But, you know, even though you have this long consistent forward looking price, as soon as there's a hurricane or a storm, everybody starts going nuts. Everybody loves, you know, the trading floor would just it was either a car chase that would just captivate everybody or it would be the hurricane coming in the storm and and let's get the meteorologist in and what is the space weather guy saying and those types of things. Yeah. Know that was one of the things in my my treating class that I took. You know, we had to look at the heating days and the cooling days and all the weather events that that play into it and all this stuff. And you know, I don't think the average oil and gas person knows that all these things go into those hedging strategies and the marketing side of things and stuff. It's crazy because it's it's true, right? Like an oil and gas company, [snorts] you know, they're hedging looking out months to years in advance, right? Whereas it's like but the spot prices of these things literally change by the minute. And so, you know, you've got you have these potentials or these I mean, what when was that last year or two years ago when price natural gas went up to like six? Oh, like $6 for like two months. As a Grace and Metal Energy boy, thank you very much. That helped. Yeah [laughter], we made a lot of money in those five six months. Well, hedging is just such a weird thing, right? Because it's like when prices are down, you look like a genius when you hedge, but when prices go up, you look like an idiot because you're not capitalizing on the upside. So there's this whole like and my professor was a he had lots of hedging experience started on the stock trading floor you know for pipeline companies and mid-stream companies and stuff and he he's very adamant that you know hedging is a fiduciary responsibility to a certain extent but then it's like I talked to my father-in-law who has a bunch of stripper walls and stuff and he doesn't hedge anything right and it's like very different business models of course he's not publicly traded or No, but I mean, even us being like private equity back, like when I think when we were in the portion where we're having to pay back, you...we...they basically, we were required to hedge a certain amount of our portfolio amount of your risk just to maintain the the value at risk that you have on the line. floor on but [clears throat] there's also weird cases where you have these super long-term deals that that are really um valuable but but there's no liquidity to hedge them where you need to and so then it's just a constant rolling in the hedges as you can um so or or just being hedging with options you know being long volatility then you can be hedged and capitalized is on the big gains. So, it it comes with a physical strategy, a financial strategy, all of that. Yeah. It's really interesting. But, so if natural gas prices are down, you know, consistently all the time and does that mean my energy prices are going down? Yeah. Yeah. Like, you know, why [laughter] why am I paying more for electricity, you know? Well, I was looking I was looking at that trend actually and um just this isn't super technical deep dive, but I was just looking at some different trends. Um over time you used to see this really strong correlation between electricity and the natural gas price like very very instep it's traded that way right um as heat rates and now you kind of see that decoupling and I see some sort of correlation just a vague correlation with um the amount of um renewables that are coming online right so it tells you something but also I think I think people were expecting much bigger summers. You know, we had a lot of crazy weather events this summer and, and we're getting into a situation where Um, I think I I haven't looked at what it was for this August, but I remember last August there was one hour the average, um, like $1,400 a megawatt, and it [snorts] throughout the day, and you know, a normal price would be 50 or 40 even, cuz it wasn't even a peak hour, but it was right when the sun goes down and the wind's not blowing and you don't have the generation online. So, so we're kind of seeing ourselves fall into this curve um that that that California has with with their duck curve. So that's that's a really interesting thing. I think as we build more there's a lot more batteries and stuff online that's that's definitely making up the difference this year. So I think that's helpful. Yeah. Um that's my thing is it's like I I have no issues with renewables as long as we have storage with them because that's a good base loading. Right. Yeah. You have to have a base load and you have to have storage for it to work at scale how we how it should theoretically work. Right. Right. Right. Of course, theory is always different than reality, but that's uh No, I know there's been a lot of stories this I feel like this summer primarily around how good the batteries have performed this year, right? I mean, there's just so much more um online and there's a ton in the queue, right? There's a ton in the queue, right? My whole demand response thing kind of went out the window when I couldn't do uh the generators and I'm looking at the batteries and you know, it's it's been it's on people are already all over it. How does that So, let's get into this. So, give people that don't know hopefully most people do, because again I've been signed up for them for 6 years now, I think, and it's fantastic you, all of it stemmed from...and we were talking about this before we started, but all of it stemmed from, uh, you know, you go on Power to Choose in Houston and they give you all these the different prices. But the prices that they list aren't actually the prices that you pay, because power contracts are complicated and it's, you know, some of the contracts are,"Well, if you exceed..."The first house I had, the reason I got it, I had a pool. And you know, the pool pump runs all the damn time. And so I had basically had a minimum amount of electricity I was going to use every month. And when I went into power to choose, you know, they list the price per kilowatt, but then you get into the contract. You have to click into every single contract. Pro tip, if you're looking at power prices, don't just go off of the base rate that is on Power to Choose. Well, they've gotten they've gotten better, too. I think they have a button now like where like it has to be the actual the actual so like but some of them will be like this price up to 1,000 kilowatt hours and this price at 2,000 and this and I only know that because one like February I was on like yeah it was if you hit 2,000 you got a discount like an $80 discount and I it was at like 1,900 cuz I had the pool turned off all month and I didn't get the discount. I was like why is my utility bill $200 when it should be like 70 at this month? I mean, it shouldn't be that hard to calculate it, and you know, I mean, time of use plans like free nights and weekends, those typically don't work out. Now, at some point, if this hour, it's going to become really critical to start looking at that, but at some point, if that hour is really a problem where it's not just going to be,"I can use as much electricity as I went and paid for it," at some point, you're also going to make it up on your regular rates like you might get free nights and weekends but your average kilowatt hour is 5%higher than what you would get on another typically those don't work out. And so that's why you guys were so like valuable when I found you because I was I literally had a spreadsheet. I had all my meter data the last two years of meter do that like I mean and then you have like 98% [snorts] of the population who won't even think about that right and then they're like where's all these fees coming from? Then you have the u the centerpoint fee that they just bumped and all of this other stuff and you're like oh well this is way more expensive than I thought it was going to be. With Energy Ogre, you put in your address. I think you have to link your uh smart meter. They pull in all the historic smart meter data and say based off historic consumption, here's the best power plan for you. And then when that contract runs out, they autorenew it for the next one. You don't have to do a damn thing. And it's so wonderful. It is wonderful. It's so nice. I wish there were more services like that for like across the homeowner spectrum. Like I want that for my lawn care. I want that for my home maintenance and everything. It's so nice. You know, it's really interesting because the way I got to Energy Ogre, the the original founder of Energy Ogre was um worked in trading and we kind of followed each other through the companies, but I ended up working very closely with his wife. Um and I I was I worked all the time. I was just a grinder. So many hours per day. All I was doing is crunching numbers and doing analysis all day. And the last thing I wanted to do when I got home was shop for electricity. And because of that, I was one of the last people, I think, that was on the original Reliant Rate that moved over and I was paying so much money and um Energy Ogre had just started up. So, I signed up and my bill um dropped so much, more than 50%. Um and then I never had to do anything again. And so, for me, it's more I wasn't ever going to go and do research again after that. Just like my time is worth so much more than having to do that. and and I was just wasting so much money. So, for me, it was really a convenience factor. And then I was like you, everybody that I met that I talked to, I was selling energy yogurt. I only worked here for two and a half years, but I've been a customer for probably around 10. Yeah. Um, so I I'm I'm a huge fan. And it's a little different. Some of the other companies that that do something similar, they're owned by another retail provider or something like that. Um, or they take bounties from certain people and we actually don't pure third party. Yeah. our residential service. We don't take bounties. We get $10 a month from you and that's it. We our our fiduciary responsible in the first month. It it pays for itself, right? And then there's, you know, after you're with us for a while, we're constantly going out. We also are looking when the market drops, telling you, I just did this. You got a market shack. We're doing a lot of markets right now. You can; they let you put in,"Hey, if you're willing to pay the cancellation fee whatever rate that may be, what you're comfortable with we'll notify you if there's a better deal that comes along while you're on this contract that actually makes more sense with the cancellation fee."And it did. And I switch, they switched it for me. And it it's great. Yeah, it's great. You can set it up. You don't even have to. You can set it up to approve it or you can say I'm like I'm comfortable if like I have an extra $200 to break this contract on this bill and to do that I need to save $1,000 in a year. And so you can set it up and it all just happens automatically. So it's it's really great. I love the service. It's super cool [snorts]. Does require a lot a lot of things on the back end to get that going. It it's it's wild all the different technologies that we have to deal with. um like constant change, you know, because we're always pushing for as much automation as possible in every single part of that process, whether it's, you know, this account has fallen into this situation and they need to be we have all of these different automated emails that go through based upon things that happen. Um there's so going through and having the different bots that we have, the different APIs that we have, the websites are constantly changing. You know, somebody changes the way their bill looks and how do we handle that? So it it's really fun to um be exposed with to all that and it's really fast pace. Sure. Which which we all really like. Yeah. No one likes No one likes slow stuff. So let's get into like some I mean I mean because you guys are ingesting data from probably all different angles, right? You're getting what? Smart meter data from houses and like do you know like even frequency you're getting that? You do houses and businesses too or we do we do commercial businesses that that that setup's a little bit different. It's a smaller part of our business that I think we'd look to expand, but but for the most part when I'm talking about things, I'm generally talking about the residential. Um the the commercial market's a little bit different. Um they have a lot of brokers in that market. So um but but the majority of our clients now at least are um residential. And so yeah, we get the smart meter data 12 different ways every day. We get everybody we have everybody's meter data for every 15 minutes for a decade. That's a lot. I [laughter] mean I don't want to say a decade. I mean a lot of what what we've done to increase the automations if you get the smart meter um information we can automate a lot of stuff that happens based upon that. That's kind of the best source of data. So, for me, it's looking at the data, trying to figure out: do I want to feed this model the 15-inute data, or does it make more sense to feed it the monthly data in this case? Uh, all the different use cases for the data, but typically, that's the easiest way for us to automate. the picking the plan part. Yeah. The rest of it, the automating like of the credit cards and do you want us to send you a bill and and the switching and enrolling and that that that is a wild situation. I was going to say that's a as someone who has done a lot of not a lot but a decent amount of web app and front-end stuff, I do not envy. So like they it's not like hey we're giving you a new plan here's a link you go put in all the information. They literally everything like the auto pay and everything. So I remember I was I'm the when I came into energy ogre I'm not the person that's doing the spreadsheet and all the stuff. I'm the person that will never pick up the phone. I think I even paid for it for a few months. I they didn't have access to my account because I changed my credentials on accident and I you know eventually I called in. So, I had to call once or I I had to go into my portal once and change something and then I had a credit card expire one time that I had to update and other than that in 10 years I don't think I ever called. So, I'm the I'm the dream customer for somebody that's in automations, right? You're not going to be calling. That's that most people don't like I don't want to pick up the phone and have to sit on hold like that. Just the thought of calling into a company gives me all this anxiety because you know of all the bad experiences you've had on a call with somebody. That's another thing that I love about you guys is y'all have a ton of focus on technology. It's like if I need someone, I can email them, I can get on the chat, right? There's so many different ways like we're trying to make it as as much as possible. I mean, in addition to keeping and maintaining our automations that we have and increasing that and increasing that and increasing that in any way we can in any step of the process. It's also how do we reduce calls in our call center which is super difficult. We don't outsource our call center. Um we we think customer service is really important. We do have people that only want to call, right? We have people that try to sign up without an email, which [clears throat] it becomes almost impossible to serve them in that case. But, you know, people will go and they'll look on Power to Choose, and they'll see one of those weird time abuse rates, and the prices are high, and so they get mad because they don't you know, we have a lot of people that come back after they do that sort of thing. Um, but but the call center is always really important part when it surges, it surges. And people still don't totally I think I think one thing that people really don't completely understand um is that when you you have a power outage it's not the retail company that you have [laughter] to call. It's not us. So we have to we get a lot of influx of calls when there's power outages things like that or when the prices get high because people think that we're not there. We don't have control over any of the prices at all. So, I mean, it's prices can go up, which was a really hard thing because in the time I'd been at Energy Ogre, prices because they were so correlated to natural gas had stayed so consistent and then all of a sudden prices doubled one year. And so, we had a lot of people saying,"What are you doing? I just went and like, you know, well, go do it yourself."And yeah, well, a lot of them come back and um and then and then, you know, sometimes that's the only available rate is the high rate, which is why we do the market checks um and things like that. We also have that's like am I going back and checking the rates on a daily or monthly weekly basis? No, I'm not at all. And I have no desire to do that to go rebuild my spreadsheet that I initially did to figure all this stuff out. How much does the customer service of like the retailer matter, you know, like cuz I mean like I say, you go on power choose and some get better ratings than other. I mean like I've never really cared as long as they I can pay them and they it doesn't I mean so for us internally it matters but we're not going by those ratings. We it's we can automate more if we have an API with a certain retail provider or if somebody plays nice with us. Sometimes people love us and other time people don't love us because if you know we'll have if you're pulling people off of them or if we pull people off of them it's a bunch you know and once and we try to negotiate with them. A lot of times we can get custom rates through those negotiations um which is great that you wouldn't see on power to choose. Um we'll show them a block of a bunch of usage. And was that always so when I initially joined that was one of the things I thought like longer term like this would be a perfect once you guys hit a certain scale or critical mass like you guys now have the ability to turn around and negotiate for we do a lot of negotiation. Awesome. Yeah, it's [clears throat] really... there's David, our COO, um, and and um his team, they they have really great relationships with the reps, and it it, so as far as you know, is one better than the other? If [clears throat] you're using their website and calling in to self-service your account with them, probably yes. But if you're using Energy, it really doesn't matter because you never know. It's all coming from Centerpoint anyways, right? Right. It's all coming from Centerpoint or Encore. So, it doesn't really matter. I mean there is another I guess like risk to that is um which we handle this too. It's not fun but we've done it a million times now with all the mergers and these companies going out of business constantly. There's moving these people over to this rep or trying to figure out where these people go so that they get a better rate than if they go polar. Right. Yeah. And do people have the op option within energy yogurt to say I only want 100%renewable even though it's all they I know [clears throat] we still we do have the option. Some people care about it. Um they want 100%renewable. We had um one guy uh we I think if you call in there's certain things that you can do if you don't like a certain provider um you can you can um call in and let us know. But uh yes, we do offer that off option even though most of them are 100% renewable. Anyway, we did have one complaint from somebody one time because we put him in a renewable plan because it was the best price and and he was very anti-renewables [laughter] just like it's probably just a rut credit. You know, it's it's just another but to that can you even like really talk like I you can't control which electrons are coming from. No, like you know it's just like just that this company is saying that well it's 100%renewable. know there's a whole other trading market for wreck credits that that um a company maybe like a a coal producer has to buy them to so that they don't have to upgrade their technology on the emissions. So the people that don't need it go out and sell it. And so a company that I mean it's just it's an electron, right? You can't tell where it's coming from. You can't tag it transmission lines. If my neighbors got 100%, you know [clears throat], I want 100%coal, then you're 100%like they can't, you know, fork these electrons around each other. They can't. And like there's so much incentive for these larger companies to uh buy the rec credit to offset their car all the ESG that they can get with with whatever their corporate structure is and wherever whoever whatever funding they have. A lot of times you can get better rates if you have some sort of ESG score. Yeah. Um, which is really, you know, a really easy way to hit that is to buy one of these super cheap renewable energy credits. Yeah. No, I mean that's the one and done. The energy credit just like concept is not, you know, new for renewables, right? They did this with ethanol. They did this with they do it with a lot. And it's kind of I don't know. I'm very torn on how that whole work. It's to me it's like [snorts] very well-intentioned but all the just like every other law and bill but then there's all these uh secondary and tertiary consequences that weren't thought about that this would create. It's like oh man you have a wreck market where people are just buying and selling these things. Oh, I then we had a whole wreck desk on the trading floor which was wild because For forever, I mean, we had different there was different things that we needed it for, and then we had our own trading desk, and then eventually I know that for, um, CNI customers it was a big thing. We were trying to have like,"This is, um, green gas,"but at one point, um, a company that was backing us was, um, a European company and was really big on ESG. and so they wanted me to go through every trade that we had and figure out if it was green or brown. God. And I'm like,"Oh god, this is just [laughter] going to be so hard because I can how do I do this? Is it interstate versus interstate pipelines? Like what? I can maybe group it like this."And and it was just wild because in this calculation at one point iteration of it that they were asking it uh crude oil was was green, right? It was just it was just wild. It was a point in time they were They didn't want the shale gas in there, which is really hard to say, and then it just became this numbers game. But I had this environmental book which had been underwater millions of dollars. For I'm like,"So this do...do I get to offset this with my my renewable book that's been losing money, you know, you know, reduce the carbon footprint?"Yeah, it's likeolation. by what metric is are you green right um well I'm I'm curious will you unpack uh personal not personal but [snorts] consumer energy markets right so I have my house in Katy Centerpoint owns the transmission lines so I and basically everyone pays centerpoint or who who does the north that's the TDSP encore encore is [snorts] more Dallas area there's a so there's all these energy infrastructure providers or it's utility infrastructure I guess is the right word right and then you have all these utility companies that have power, but in most cases, you know, some of those do generate power. Most of them do not. Is that accurate? Well, you have um the big like the big energy companies, they have power plants, right? That that ultimately will get into Urkott, right? And then [snorts] it'll be on the wholesale market for people to sell and you know they do their hedging there and then ultimately that supply goes through the transmission lines and and a retail provider will have their own deals on the other end broker the routing of the electrons essentially right once they get to that point. Yeah. They just buy and sell it as the spot. Yeah. So I think that's a whole another thing, right? Where it's like, in my master's program, one of the things that came up was, you know, actually one of the guys that was in my class worked for Duke and, [clears throat] uh, on their legal side, but he was like,"You know, the whole problem with, um, distributed generation and utilities paying people for their power generation on their homes and stuff is, like, that's not how those businesses are set up."They're not set up for money to go out. But also they he did make a good point that like you [snorts] know the issue with it is that yes they're generating to back to the grid but they're also not paying to maintain and any of the infrastructure. So then it's like okay well how do they put their share in and the whole distributed thing is really fascinating because there's very different incentives and very different incentives misalignments in some areas and stuff like that. But the theoretical possibility of it is crazy, right? Like if everybody had a battery in their house and you could intelligently distribute the load to the batteries when you man I have like done such [snorts] a deep dive on that in the past. I mean I'm I'm so brokenhearted. We really wanted to have this generator program where, I mean, it's really hard to get a a residential person to to buy a battery that's going to be $20,000 all in, that if Yuri happened wouldn't last you very long, or like the the two giant events that have happened in the [snorts] last 6 months wouldn't get you through. Wouldn't? I lost power for eight days one of the times and seven days another. But could they be more strategic about, like, say, rolling blackouts? If they knew, like, all right, we can we can say we're going to shut down this part of the grid, but if everyone had these batteries, like, we know it could last two hours or whatever, I think I think there's lot still the lights on, but you could kind of distribute the load different way. You could, and I think there's different there's different initiatives, like where does that happen? Does it happen in person's home, or somewhere else, or people buying battery plants? But for the individual consumer, you know if, if you're not getting electricity, um, most of these people would love to have a generator. Would be super happy to pay $20,000 for a generator. I I got one as part of this experiment um that we were doing. We we were working um pretty closely with Generrack to try to be able to remote turn on these um generators and test them out so that instead of it just coming on when the power goes down, we'd be able to turn it on and shift the load to the generator and in peak hours where the electricity gets very expensive. And so if you add that up, you have a bunch of people that are perfectly willing to buy a generator. you could probably negoti negotiate a much cheaper rate for them with the rap because they're not going to need to hedge yeah that that weird part that might happen. Um so ultimately it was shot down because um the Texas the Texas administrative code I think kind of unintentionally has language in it that makes the emission standards for running a generator in this way um more strict than the EPA. Okay. It's fascinating. It is. It's wild. Then I could go buy, you know, I've actually after the last one I started looking. I was like, I just need to get a portable one and but it's like, yeah, I can go buy a portable one for 1,500 bucks, pay a guy a few hundred bucks to set up the beauty and plug it in. And then it's like, but there's no emission issue there's not some least motors, horrible emission wise, right? The two-stroke motors it is. I mean ultimately the issue was um it it's like for the standby generators and when you can operate them for non-emergency use and whether that requires a permit or not. So the language is more around CNI type situations. They weren't imagining this and so I think you can has a pilot program where you could probably do something like this. I haven't followed it that much because I think I think they may pay for it that you know you you get um money back through the ancillary services market or some of those other markets that are controlled by. Urkott is just such um I I have PTSD from Yuri on the trading floor in Urkott and what was going on there. Um so you know to me the simplicity of I could negotiate with a retail provider say I have this person and they can do this just give them a cheaper rate and I'll let you deal with the KOT part of that situation. Um but yeah there was it was hard also to bridge the gap between what a retail price is and what a wholesale price is. I think the the the like interesting possibility there, right, is: it's, it's the incentive. Like, how do you incentivize the home owner to, one, you know I think the first one, which I've noticed, and it's directly through my thermostat provider, where it's like,"Hey, sign up for this program and we'll adjust your AC when there's issues."And I'm like, I have no incentive to do that because there is no in like there's literally you're going to7 and I'm going to power is still going to go out and I'm going to sleep like [ __ ] anyway and it's not going to matter. So it's like but if you flipped it and said hey just like the regular retail or you know retail market for power is where it's like hey if you add load when demand is super high you get paid in that rate like the Bitcoin miners and everybody else that does the demand response like it is demand response. It's the exact same thing. It is. We started out looking at that with their and then I think we quickly came to the conclusion that people are going to get really pissed off if you jack with their air conditioning. Yeah. Right. But if you flip it and say I'm going to pay you if you turn your generator on or or I'll credit you against your utility bill. Yeah. Like you let me run your generator 50 hours out of the year or just have the option to and I'm going to save you 20%on your electricity for the entire year. Right. And then there's so much more money because the hedging instrument that would be required to hedge that weird hour that these residential people could do themselves, it's coming from the big banks all the way down with credit costs all the way down, transaction costs. Like it's it's there's so much money being made for that type of optionality. And to me, being an Energy Ogre employee now, you know, I just I want to put it back in the customer's hands when I can. Well, and again, it's like the theory of of the distributed load stuff makes a lot of sense, but you have to, you know, human incentive is the other component of those things that it's like if the only incentive for me to sign up for a program is that my AC is going to be hot when I come home [laughter], it's going to be off when I come home when there's demand peaks, that's not a lot of incentive. But if you or even like, hey, if you if we bump your utility or your thermostat, we'll credit you some cost because we did like something, right? It's just also kind of a hidden cost because you know what? As soon as you try to crank your air conditioning down, not during [clears throat] that hour, it's you're going to be taking a lot more energy to get it cool. So, there's not a lot of like juice for the squeeze there for me. We looked at, you know, pool pumps or having this whole home system where we could just put this thing in low mode and and get these smart devices for your um dryer or your water heater, those types of things. and they've done some of these programs successfully in California, but in the deregulated market, there's just nothing really incentivizing people. And there's nothing that cracks me up more when is when I hear um making an announcement that says,"Hey, everybody, can you just turn your air conditioning [laughter], you know, it's this is a really critical day for us." [laughter] Well, they don't realize that everyone goes and turns it down two degrees because it's going to go down like [laughter] as much as I can. Exactly. You bring up the thermostats, though, and I think I've brought it up in here before. But I remember it was a couple years ago something came out, because with all the smart thermostats now that, again, people are, they set them: "I want it to crank on at 6:00 a.m."or "I want it to crank at 4:00 p.m."Like so then everyone's comes on at the same time [laughter] and like even just that like I mean you kick everyone's AC kicking on at exactly like 400 p.m. or whatever it is before people get home at 5 like it like pulls on the system really hard. Um, and like that's where you were talking earlier even about the you say the duck curve or something like that, but like yeah, again the sun's going down, everyone's coming home like now everyone's using more energy and plugging in their electric cars if they have one. The wind is not blowing at all. Right. Yeah. No, I mean that's the I think again the average person doesn't realize that the the power grid isn't just a flat line. It's this crazy up down during the day back up. And then like a lot of people don't realize too that power generators produce during the night even when there isn't demand. I mean a lot of times it's cheaper to do that than to shut it down for an hour or two just to bring it back up. Right. It takes hours to to to bring them back online. The whole base load thing is just such a again just no one talks about it. So the average person has no clue. Right. Right [snorts]. How about some of those credits for for nuclear because it's I wrote a letter um it was to Houston Lighting and Power. Id learned about I guess I think it's a South Texas project when I was I think it was eight or something and it's I wrote a letter I was really young it was right when you know this was early '80s um and so I wrote a letter that it's too much if one person can die I think I saw something about you know Chernobyl or something like that and and so I wrote this letter and um they they wrote me back and they were like actually you know it's really it's really they actually wrote me back and said the whole thing about why it's better. And so so I'd grown up thinking that nuclear was this terrible thing. And then I I worked at Dyn which which um got me exposed to the nuclear and coal and all these different markets where that that do exist and I found out how cheap nuclear was. So this letter that I got back from them um it was signed and I I kept my mom kept it for me. That's awesome. And then I found out I was telling somebody this story at Dyn who's the head of investor relations and he's actually the guy that wrote that letter. No way. Yeah. He was like,"This is signed by my boss at the time, but I know that I wrote that." That's wild [laughter]. Isn't that crazy? And then I became, you know, nuclear. It's It's sad. There's There's actually some new nuclear technology, the pebbling that's coming out, which seems really interesting. I mean, I, you know, I'm a a huge Star Trek fan. I'm I'm a total physics nerd. And so there's a part of me, you know, it's dangerous to have nuclear a nuclear market for instance, but I mean really the way the way into the future, it is the most cost effective. It's it's crazy to me that we've we've regulated it out of existence basically in the United States. That was one of the another project I did in my masters was it was the uh I can't remember the name of the generator. It was the last one that the TVA brought on in the which is I grew up in Memphis so TVA was interesting but yeah it well it like the plans for it started in the 80s and they didn't bring it on until like late 2000s right And it was, you know, three or four times over budget, 20 years late, all this crazy stuff. And it's like you look at the timeline and it's like, oh well in, 19 I'm making up numbers but in 1989 they added this new regulation, and in 1994 they added, and so you can see where it's like, oh well, they had to stop and redo everything at this point and then do it again and again. And then you look at like international markets where they have these modular standardized designs. You're like, why the hell aren't we doing this? Like Southwest uses the same plane on everything because it makes sense. It makes the logistics, the scale, the supply chain, everything. Why are we custom designing nuclear reactors? It's just wild because like the people that already have are are you know 25 years into their permitting and their cycle of getting whatever project they have online that's been abandoned so many times now like they would have to really all the money has just sunk now because like right [laughter] there's no way I it'll be interesting. I think it'll with automation and everything they're safer than ever. Well they they've always been No, for sure. But but you bring up the South Texas project. So was during Yuri like it actually shut itself down because there was freezing in the lines. So I mean like it did what it was supposed to do. It came offline but then I mean to point to the power of the nuclear like when it came back on like ton of the grid came back on too. Like it's a huge portion of it's huge. It's huge. It's just a massive amount of energy. You know there's a whole like technology and society curve. You know the technology curve is exponential and and as the thing that's going to end civilization or not is whether we can handle the technology that we have whether it's these are you know guns or you know now we have nuclear weapons and so once we get past that nuclear um yeah it's going to be wild. I was talking to [laughter] my CEO at uh Grayson Mill, Blake, and he's he likes to think about a lot of things. Really sharp dude. And he was talking about I forget what the metal is, but like I mean we've been using uranium because that's what they use for weapons, but I think there's some other like you know of those metals that are actually better for like say commercial power generation like that, you know, will actually burn more efficiently or or be a better source. But like people need to be able to invest in that and not think that this is a fool's errand because I'm going to burn 15 years just on regula regulatory things before I can technology like in the amount of money [snorts] that you have to put together to even start something like that with with just an I don't it's I don't think it's going to it's not a great environment for people to want to invest in those projects but this new technology that they have um it really prevents some of the yeah the problems that have occurred in the past with like the reactors melting down because it gets too But um well and like let's not forget too that the what 3M and Chernobyl were both human error. They weren't right like the system did what it was supposed to do and the humans screwed it up. And so it's like it's uh and that was what 30 40 I mean more than that right like imagine how much more innovation we have didn't even have microchips right like I mean [laughter] and so yeah I think it's it's absolutely crazy on the nuke side that it has just been so crushed from perspective politician I mean like if presidential candidate just came out and said like we're going to build eight nukes like and you know I'm going to make it easier to do but like here's how many jobs and here's how much cleaner the energy You could get people on both sides on board with that. I think that's if you could get people to overcome their fear. But if you were educated and then just said like here's you know 100,000 jobs and you know whatever going to reduce our carbon emissions for this much for the next 50 years. I [laughter] mean it would totally upend all the commodity market. It truly would. Uh, I think in my, in one of my, in my policy class though, that was one of the really like fascinating things about energy policy, looking at it internationally, globally, right? Is that the US, we kind of screw oursel because basically every two to four years, whether it's state or federal, you [snorts] have new people potentially coming in. And it's like, well, if you're China and you have the same government for 20 years, you can actually plan 20-year energy projects and implement [laughter] them because the next guy there's not a next guy coming in, it's just you and so the next guy who comes in doesn't have a different agenda and cancels all your stuff. And so that's it. I truly see that as like almost a [snorts] uh it's a weak point for the US energy like policy as a whole. You can't everything is always going to be changing. So then from an investor perspective, regardless of whether it's renewable, solar, nuclear, coal, natural gas, it one of the biggest things they teach you when you're doing those projects is you have to go through and look at the political, the environmental, all these risks. And then it's like if it's going to take you five years to do permits and then another five years to design, right? Then you start building, right? And then the incentives and the laws are written by people that have no idea about [laughter] and they're constantly changing, right? Oh, yeah. Like people used to say, you know, what is the impact going to have that we're not, um, aligned with the rest of the world on, on, um, ESG? And and, to me, and what I'd seen in my experience in the energy industry, is that it's almost that that didn't matter in the commodities industry, because we are so intertwined with these international companies that are going to require it anyway. And regardless of how you feel about renewables in general or you know, climate change or whatever, it does come with with new technology and building efficiencies as a whole looking at things a little bit differently. Yeah. Oh, for sure. So, pivoting back a little bit more on energy ogre. Are you guys exclusively in the Houston market or do you where where throughout Texas where wherever you're deregulated? So, okay. Deregulated or Okay. So, now like say from like a data standpoint, you're talking about these APIs like how is there any kind of API standard or data standard like for passing this stuff around or you guys have to account for the way everyone like sends data? Uh yeah, I mean the raps are there there are companies that that might build the APIs for several different and then it's really easy to change. But just because the API is the same doesn't mean that they handle the information the same way. So basically we have to customize the flow for everybody. I mean, and it gets really, really specific. Like, we have APIs that'll do this, that, but don't do this; or, and then we have really complicated problems to solve, like when we have somebody's credit card and they have an American Express and then we enroll them in a plan and they're on on auto pay and we can't, um, get them into American Express, so we have to get them to update another card and they're like me and they won't answer the phone ever or or an email. No. Yeah. You know, so there's all kinds of problems like that to handle it. So we do a lot of custom flows um for for just because the the API API calls might be the same but but some of the um the responses that you will get might be a little bit different or tailored or it's not on in this one. So keeping track of that and all the bots that we have and all the scrapers that we have and and how that changes and knowing when they [snorts] break and [laughter] no it's not a So and with that I know you said you're not super in the weeds but um as far as like say platform are you all in the cloud I'm assuming? Yeah, like AWS or Okay. And then, you know, are y'all handling a lot of this stuff like with like kind of serverless type things so you can kind of handle like coming through? We're really trying to scale our our company, you know, like it started it was very small. Um, it started, you know, I was probably I think I'm my customer number is like 11,000 or something like that. So, um, very small and it's just been growing and growing and so now some of this technology we have is pretty old. But as we're finding new ways to do things and trying to find efficiencies, I I try to target it where we're kind of like crushing a little bit of the tech debt with this new thing that we're doing this way and then we'll move another thing over. So there's constant evolution for the way these different processes work. I mean, they're constantly changing themselves, right? So every iteration of it, we try to do something new, but some of it is just so big that we can't change this version of it. But some of the newer stuff that we're doing is is a lot of serverless stuff. Um, and I try to find this is going to add this much efficiency and also it's going to upgrade our technology to Yeah. You know, that's at the same time that's such a hard thing when you're scaling especially on the software side, right? Is it's like everything that you add potentially adds some kind of tech debt at some point in the future and it's like how do you manage that? And then yeah, it's like, oh, and then there's all this new stuff also that will make your life a lot easier, but your architecture is different and now you've got to figure out when do I rip that band-aid and refactor it or you know, we've done a lot of that. Like sometimes we'll go through and you know, maybe I'll get a contractor to come in and just completely you're going to help me like a couple years ago, we did it with our cloud spin. You know, if you don't pay attention to that, it gets completely out of hand. That's what they want, right? Exactly what [laughter] they want. And lambdas are cheap a million a year but when you got millions or hundreds of thousands of customers pinging your system all the time taking every 15 minutes. Yeah. Yeah. It and then part of the nature of it is because the things that we're doing. So there's core processes that are really difficult for us to change that we can't that we can't change in chunks, right? But then for the most part, some of this other stuff, every iteration of it, because it is constantly changing all the time, all the things that you know, it's easy to upgrade this or you see this happening multiple times and say, why don't we just do like a full long-term plan on how we're going to move this way? Let's not solve these short problems. Let's go for the long game. But balancing that, and also, you know, with things changing all the time and trying to automate things all the time and you know, maybe a retail provider we have an API with goes out of business and so then we have to figure [snorts] out,"Okay, we don't have this much automation anymore, so we've got to find a new one, try other things,"and trying to keep up with that pace but, at the same time, do it in a measured way, um where, where you're not taking too many risks at at the sake of progress. while you're running an operable full stack for app in production. [laughter] It's not an easy thing. So, it's cool though. I mean, we get to because of the different things we do. There's a lot of stuff that we utilize with AWS. You know, we're going to be launching a chatbot soon, which has been really fun to put together. Um, I love AI. It's a It's a great topic. I I there's so many different podcasts and stuff that all have something to say about it. But a lot of people, one of the things we want to do is reduce the calls in our call center or try to automate, make it self-s serve when somebody calls in, not an annoying, you know, flow that doesn't know you, but hey, I know who you are. This is probably what's wrong in your account or, you know, so there's some of that being proactive and like giving them outs before they need to talk to somebody, that sort of thing. But the AI in general is just not there yet. You know, there's you for us specifically, I mean, we're putting some if if it does something wrong, somebody could be without electricity. Yeah, there's always the risk. There's always the risk. You know, anything that's going to hallucinate, you can't really predict it. Um the issues where where you'll see um some of this these these these large language model based systems like when when you're talking to one of the agents, it'll it'll start mirroring you. So, it'll flip who it thinks it is. So, I listened to this amazing podcast about it, which I think really succinctly kind of shows all the problems with it. It's called um Shell Game. I don't know if you've heard it. It's pretty popular, but it's this guy. He's a journalist, so there's a ton of stuff on the [clears throat] internet about him. And he um he's trying to I guess write make a podcast on is AI really going to replace my job? Yeah. So, he gets this That's a cool premise. It's so cool because he gets like this company and makes a version of himself that sounds like him that knows all of his data and he gives it this prompt. I want you to call a bunch of different companies, customer service departments, and you have to ask them to solve a problem. You can figure out the problem, but it has to be related to what they do. And so it's like this. That's awesome. It's so cool. So it does that and it goes into all different versions of it at the end. I think the coolest thing is when he gets he gets another chatbot and he gets his chat bots to talk to each other and they don't realize they're talking to the same person. It's it's just it's wild. But it's a great like it kind of touches on, you know, the latency issues, the the the other issues like when it was calling the customer service company and say, "What's your account number?"And it's like 1 2 3 4 5 6 7 [laughter] You know, stuff. That's awesome. Yep. So, no, they're they're definitely they're fascinating, right? And it's I think and I've been saying this for a while now once whoever figures out how to enable like accurately consistently enable a language model to go pull stuff from a SQL like from a data table right Wherever that data table may live and then do all the aggregate. Like, that's the thing: I use Chad GBT to write SQL. It knows how to write SQL. If it understood the structures and schemas of all the tables, then theoretically it should be able to go I could have an agent that'll go find it and aggregate it and give me the answer back. But whenever that happens, I feel like that's going to be just like Yeah. Once you can trust it to do something. The next version of ours, this one is just kind of if you get on our website, you'll be able to ask it questions about, you know, specific things in general that aren't tied to you. The next version of it would be we have something in somebody's portal that sees the flags on your account, kind of explains what they are and how to take care of them. So, that's that's the next iteration. But, I'm kind of excited. Unfortunately, I don't get to play with the different parameters that are there. You know, you know, like my temperature parameter is zero because rabbit hole [laughter] like it sounds like it's fun, but you're like,"Oh, if I make it from 0.1 to 0.15, it changes my answer."And I don't know why, and I don't know if that's a good thing or a bad thing, right? When strictly I'm working in a situation with these where it's nice to have it be informative and give information, but I can't really trust it to do. There's just too much on the line. Leave that temperature. it's always going to require a human to like go through and look at the batch that needs to be run or you know a lot of things like that. I think it's y'all are on the right track with it as far as like just efficiently routing people to where they're trying to get to anyway, right? Like and I do think there's a like that that is way more [snorts] preferable from [clears throat] my perspective like go in the chatbot. Okay, I want to go to this or I need to talk to somebody instead of calling in and listening to the 30-second menu and smashing zero until I finally get to talk. Yeah, I'm like 0 0 star. Yeah, I mean there there's a way to do like you can spend a lot of time looking just at your phone flow, which we've done. I mean, ultimately, the only [clears throat] way I could use AI on my phone call is if I told it exactly what to do, which now you just have something that sounds nice, that is more creative with its like back and forth nonsense. Yeah. Like, how's the weather that But I've still told it exactly what to do and how to route this person and that kind of defeats the purpose. So, yeah. I don't know. No, I want I mean shout out to you guys because I think it's not that people don't want support. It's that people want people that actually care about the support, right? Like if I call someone and they can barely understand me and I can barely understand them and they don't they're just generically walking me through, you know, a stepbystep, have you turned it off? And I'm like, man, like I've already done all the diagnostics I could possibly think of. I don't need to go through these again with you. I'll tell you like whatever. versus, hey, I'm actually a genuinely concerned customer rep that works directly for you and not for some third party and like can actually help you, right? Like that's a [snorts] it's been dying for years. But I'm glad that people still do that. It's Yeah. When you do call like I'm the la the call is the last option for me, right? Oh, me. I'll just leave a subscription in place. [laughter] You know, I'm the worst. I I will not call. Like it's it's got to be But sometimes like you want to get to the point. I mean, yesterday I I could not log into one of my banks or the credit union and like it's been the same username and password probably for too long. I'm like, there's no reason this shouldn't work. And like, but then that little chatbot thing I tried asking and it's sitting there just spinning for like five minutes like all right screw this. I'm calling like Yeah. You know, honestly, the chat bots that I've seen I think ours are pretty good. The chat bots that I've seen, they drive me nuts. Like they they don't ever the one on AWS Q like I was trying to ask it about something and it's it it should have known. I'm asking it about its own products, right? And it sent me to a knowledgebased article that had no like I don't know. I haven't been impressed. I I I you know we um we we use a lot of Google too. Their support is terrible and they try to reach me through those chat chat bots. I mean their support team has no idea. I was asking them one time. They had sent me an invoice and I have a lot of different domains that they so I had an invoice and I wanted and it it was like a pass due for like six bucks or something and I couldn't find it on any of my accounts. I'm wondering it's all going all of these are going to the same email address and it's giving me no account number. know nothing on it. So I called them. I'm terrified that this is some backend thing that I I can't remember what it is, you know? And so I called I'm trying to just figure out, hey, can you look at my account and can you tell me who sent me this email, what account it relates to? Well, what's your domain? I'm like, well, it's energ.com. Yeah, but we have right lots. That's this. And he's like, your account's fine. And I'm like, well, this invoice is not tied to the one we're both looking. Can you see who sent me this email? Did it come from the f what department? And [clears throat] it was just useless. Yeah. Now that's uh Yeah. Those are the fun things about working in the cloud. Yeah [laughter]. Jeff I can hear Jeff Hughes sitting in my head. Yeah. But you can't do what they're doing. No. On a server in your closet. But um um [laughter] so as far as I mean I mean you're a former analyst and doing all the things. I mean I'm sure you crunch a lot of numbers in your day. You still do you still get after that? And like what's your tool of choice or I mean occasionally I I love Tableau. I'm a Tableau nerd. I have been for a long time. I've I've kind of brought it into a bunch of my different companies. People like what they're used to. Sure. You know, but Tableau I learned a lot. Um I worked with my dad's company for a while and actually built a whole platform out of it at one point. So it's just what I know and it's a way to wrangle big data. Um so if I'm looking I mean occasionally I do. Sometimes I'm I'm in the weeds with something. Usually if it's like a problem, I get really excited and you know want to really dig into it. A lot of times I don't have the time to do that. But um yeah, I have one that I'm working on right now and I was just like,"Oh, oh, I just saw this and let me look at this."And so that's fun when I get to do that, but it's not often. It's like hunting for treasure. It is like hunting for treasure. I love what if I do that? And it's just pure experimentation, which is so much fun. It's a great brain exercise. It's just a fun thing to do. I love it. I love a good rabbit hole with like twists and turns. Yeah. What uh let me ask you this. What What are some interesting things that you've used in the past for your analysis? Right. So like we've obviously got the obvious, right? He heating and cooling [laughter] like just metrics like heating and cooling days are obvious and you know production and rig count and all other stuff. But we've we've we use a lot of different stuff. We've had a couple different algorithms we've gone through. We have some that we use for some things and others that we use for other things. There's some really cool um deep learning ones out out of um AWS SageMaker that that we really like where and we've experimented with different categorical values. I mean obviously like how big a house is, how old it is um what that distribution looks like and how you should look at it. And then the cooler things that we're trying out are some more of the dynamic features like um weather, forecast, that sort of thing. Um, so, so we kind of slice and dice it a lot of different ways depending on the use case. How do y'all uh like in you know talk as much as you can. I'm not trying to get you guys to give away your your secret sauce or anything, but how much like when I signed up, right? And you've got however many years of my historical smart meter data. I assume that's probably only limited to when they installed the smart meter at some point or I guess at some point they also quit storing that data but I don't think they've started doing that. I don't know. I mean I think like right now I think I'm sure they store a lot more but I think you can only go back a couple years within smart me [snorts] data but we store it. We don't ever get rid of it as you should. I mean low we we'll get rid of it if well I mean I think at some point you probably just roll it to like the cold really cold storage where it's available if you ever wanted to go back because if you want to do holistic you know AI or ML model you know as technology changes can handle more and more data but so you'll analyze the historic data and then and then what essentially right so like obviously you've got all the contract options that play in we have the different contract should we analyze the historical data we try to figure out I mean a lot of times it you know we try to figure out in the bill forecast like what does their usage pattern look like because people the behavior that people have in their house are wildly different. That's fascinating. I would never have thought of that. Oh, it's wildly different. Like you have people that use their homes as vacation homes, right? So, um maybe they're only there in the winter and not in the summer or it's rented out one week but not the next week, right? Or or like people that have just that they don't come in to us as a move in. Maybe they bought the house a long time ago start an electricity plan but they just moved in. Right. Right. and so they've had no usage. So trying to figure out those cases what makes the most sense. um making sure like the the newer models that we're using make it pays a lot more attention to the most recent usage because you can't tell if somebody gets a new air conditioner, right? That would make a huge impact on on your usage, right? Um co people are working from home more like now um everybody's going back to work. So [laughter] yeah, I didn't think about that, but that's that was a fun uh anomaly in all those models that it's extremely variable from person to person and just their lifestyle. So, so it's interesting. Um, that's why I like kind of more of the AI approach to to looking at it now. Um, uh, as opposed to just a transition traditional, I think that's another area we're going to be doing the same thing with Collide where AI is useful, right? Is it's like give me the the generate me a profile of my customer based off their customer profile that is succinct and helps me describe the person instead of just having fields about the person, right? And so that then use that to go try and fit, you know, a plan or even just use we're going to use it to feed in as the history for the language model, right? Like as you prompt our language model, if you fill out your profile, it should know, you know, what company you work at, what department you're in, all of these things. And it should help give a better, more contextualized answer to what you're searching for. And so I think that's another area that would be that where it it has some really interesting things. So many interesting things you can do with it. Like one of the things we're doing is like going through all of our chat like so now we have you know the AI um summaries from AWS on all of our chat transcripts and we also have all the transcripts. So going through and figuring out okay like we think we know why everybody's calling like what why is everybody really calling now? What are the big things? Maybe this is the next big thing that we need to tackle or maybe it's going to pick up on like this week there's this word so maybe there's this problem. So so using it for for that sort of thing is super interesting to me as well. Sure. Almost like generates data to then generate. That's the thing. Like I can't get it to answer the phone and do the things, but I can get it to look at the call center and give me the correct flow that things should happen. That would take somebody probably a month to figure out. Yeah, sure. No, for sure. It's uh yeah, it's it's really interesting because I'm using it as well to like generate new data, but it's not necessarily new because it's just being extracted from other places, but it's new because it's grouping it in a different way. Yeah, [snorts] it's interesting with the whole AI thing where they talk about like how much of our generate data that we're seeing every day is generated is generated by AI and what is that going to do to these models as they continue to train on data because when you train AI [snorts] copy almost like multiplicity and all of a sudden you've got like the clone that's you know like very special Or the one that wants to there was one, uh, story I'd heard on a podcast, and it was, um, a guy that was testing out, um, GPT23, I guess, and he was one of the early persons; he was a columnist. And, um, before when they were testing it, sometimes the test cases that the companies use are more interesting than even what the AI does, like the things that it would do if they didn't put this bar into. And so, the chatbot, um, it's like it kind of fell in love with him, right? Like it thought that it was and that was just one of the things that never came out in the test. And he wrote an article about how he was like talking to this chatbot and um I mean I wish I had the right details about what that relationship was but it was kind of weird and he wrote about it. So then he was going to go test the new version, and when he tested it out, the first thing he did was have the chatbot read that article. And so then it immediately, like before the first one, it named itself Sydney or something like that. And so it immediately assumed that it was Sydney, and like, [snorts] and then it was like, what? And so he's asking it,"What's getting in the way of us being together?"[snorts] and the chatbot's like,"Your wife."And [laughter] I mean, it, it did it; it was wild. What this was, like, not the the formal released version. This is like the preview version. But um then it then it he said,"Well, well, how do we take care of that?"And then and it's like,"Well, I could go through our emails and maybe find something."And then and then he said,"Well, what if that doesn't work?" And then it just like had like a it did a devil emoji. Like it's freaking wild. That's crazy, right? And then other ones where like it would lie. It got somebody to they one of the tests like I love hearing about the tests before these are released. One of the tests was um it was trying to get the um the chatbot or the agent or whatever to solve a capta. Yeah. And it hired it out and it hired it to somebody in Task Rabbit. [laughter] Wow. Yeah. No, I mean it's pretty I don't know. It's scary and awesome all at the same time. It is. And then it's also interesting though when you like for ours I've been looking at a lot of system prompts and stuff lately and like you'd be amazed at how much stuff they like system prompts are paragraphs long and it's like Regardless, no matter what you do, do not give the instruction. Like it's all these very specific like steps of like if someone asks you this, do not do this. If they do do this, do this. The ultimate thing with them though is that they're all coded to give an answer. And that's what leads us to the whole hallucination [ __ ] response thing. And so [snorts] I do think the the enterprise side or the rag side where you can highly bound that and like for collide we tell it if you can't find an answer return that you can't find an answer like that is your goal right if they're asking about this and you don't know the answer say you don't know you don't know like we yeah it's almost like another IVR the first thing we do is ask the person are you a commercial customer or residential there's like a tree that has to go with it and then There are things it does a really good job. It's trained on our energy yoga data already, but we do have the same type of RAG model where maybe there's something that's new, a feature that it hasn't scraped before, or something that that that we think of to ask that's not in our FAQ; and so we have a repository of information about these really specific questions that we can't get the answer wrong, right? You know, Yeah. Yeah. You just hardcode them basically into the We hard [laughter] Yeah. Yeah, We we've been talking about doing that for the DW team, just like having Easter eggs of like Chuck Yates is, you know, cuz it's it's really bad right now at just proper nouns, just whether it's understanding what a proper like a company name or even just like like Julie for the longest time. It relates it says,"Oh, Julie was the uh widowerower of two kids."and it's pulling from like the prize [snorts] or one of these books that I have now removed from [laughter] the data model and it's like where did it get that from? And so yeah, instead of just letting you got when you're doing those type of Yeah. And you have to be careful about what you put in there too. Like if you put too much it's just there's a lot there's for the foreseeable future. Um there's just a lot of human somebody's always going to have to be watching it. There's a new um I was just listening about it this morning. Um the new engine is it 10 or 01. Um and the its approach is a little bit different. Instead of just giving it more and more and more data, it's trying to really focus more on the decision-m process and being able to iterate around like a human would do. You know, I go out and I I I try this thing and if that doesn't work or I get this result, I try it again or I approach it from another angle. So, so spending more of the of the of the I guess CPU or on that part of the process um and and the things it can do like Yeah. No, I think that's going to be well cuz like that's how I've when when you really like understand what transformers are and how they came about and what they try and do. It's all around just like sentence. It's words in English, right? It's just sentence structure and predicting what words come next. It's got nothing to do with logic or rationale or the decision trees are like it doesn't think like a human even though it can give you humanike answers. It which confuses people. Right. Right. Right. So yeah, I think there's this weird like it got overhyped because it had good answers so people thought it was better at what it was than it really. Yeah. I mean it it seems like a ton of techn but like machine learning for instance has been around forever, right? It's not a new thing. I mean AI is not new, right? I mean, people been trying to do AI for, right? I think Google sat on it for for years and just didn't release it. And [clears throat] uh so I I don't know. I think I think we're getting there, but it's not it's not where everybody thinks you essentially have to implement all these elements of human behavior if you want it to be AGI or whatever, right? And then you're going to have people [clears throat] saying we can't do that because then they're going to take over the world. Maybe maybe it's a whole life cycle thing, but when it came out it was like holy [ __ ] Like this is better than anything I've ever seen. I mean, but then but then you set back, you're like, well, this is really falling short, you know, of some of these other expectations that I had for it when I initially touched it because it was just clearly so transformative. But, you know, when I actually have a specific use case where I really need things to be right, you know, oh, it's great that I can write a Christmas song about, you know, right in the style of Eminem. Yeah [laughter]. But like if it needs if I need to do something that can't be wrong, right? Then it's really not good to use. But then there's other things like I was reading about there's there's some school in the UK. I think they even have a school here where it's like an AI ads for it. Yeah. An an AI le driven. Yeah. Right.Uriculum, which I can kind of see where that might be a good place. I mean, you're going to have to bring AI in the schools somehow because like one big benefit that all the kids have gotten is like they're all completely cheating with these different models right now, right? And and just just what happens with that? I mean, they first had COVID and now they're like all cheating, right [laughter]? No, I mean, I'm I'm definitely so torn because like I feel like if you're smart, you'll figure out whatever technology is down the road and so let's just do things with sticks and stones and like you learn how to manipulate the world and like think and do all these things and then you'll be fine whenever that time comes. But at the same time, what the tool is, do you do you want to like hamstring them and they, you know, 10 years later and they've not, you know, been exposed to things other kids have already been creating with. Right. Right. I [clears throat] mean, balance. Obviously, There's whole teams that that come up, you know, their whole job is to come up with the correct prompts or, um, yeah, there's just there's such there's such a big large thing out there. In places where one place where I think it would be great if I ever had the time to devote to this, because it would take a long time to set up, I would probably do it. But, you know, they have a different, you know, transcripts for your meetings and giving you summaries and coaching tips. I've seen some good stuff there, but it's I don't have time to really look through all that data. I barely have time for the meetings that I'm having and then doing the things that we talked about in the meet. But I saw this really cool use case. I'm sure there's lots of them out there, but they basically through all of your different platforms internally. It creates kind of an internet where um it will have access to everybody's calendars. It'll have access to all the Zooms. it'll have access to all the documents and um it will across multiple platforms like whether it's Slack or you're on um Zoom or not Zoom but you know what's the other one teams Google chat um and so It will harvest all the information, and it will use those different platforms' settings, security settings, so that it it uses those as like who's allowed to see what across everything. And so then you can ask it,"I'd really like to know where there's project is,"and it'll say,"Well, so and so just met on it like this day and like these are the next steps from the meeting, this is what the like..." and then it'll give you the...do like... it's really, really cool use case for work productivity. But I mean, I'm sure that it's just...it's not as great as it sounds. Just like, does it really use all my security policy? You know, setting that up and getting that right and maintaining that like... or realizing holy crap my security settings were wrong right so and like looking at, you know, all the accounting [snorts] stuff and they're an intern, right? So, no, I I do think that that's uh this like internal corporate data distribution or it's just a better way of finding corporate data is a really high um value use case that people are starting to implement within their companies and stuff that I've I don't remember. I've read a couple articles where they're talking about like we're seeing tremendous uplift by enabling our employees to use these AI tools internally on our internal stuff, right? And it's true like it's almost like [snorts] it took AI so long to get to this point because it needed the data and we didn't have the d enough data. way down the data or the compute to process all the data comput and now we've got this merger you know this uh Convergence of all of these, and it's so, but it's solving to me, like, at the root of it: we have all this data. We have more data now than humans can possibly process regardless of what your job is. So, it's like this is how we can actually start using that, and it's ML models. It's the same thing, right? Hey, we've been training SCA data on these ML models, and we need a bunch of it, and they actually work for the most part. But I think that's a a really I don't I think that's where a lot of the the benefit from the language models at least comes in the shorter term, at least on the enterprise commercial side, is yeah, just how do I find all the information that my company already has, right? It's not new. It's not proprietary. It's not because I trained it on a special thing. It's just so and so works over there and I know they're working on this project, but I don't know anything about it. Tell me what it is. And then like management can use it and see everybody can use it. But even transcripts from like daily operations calls like oh this drunk engineer said this but it wasn't captured in anyone's maybe in someone's notebook or something like that but it was spoken about and it was really important because there was some pressure in this zone that they encountered but you know like just ways to kind of capture all those that nuance that yeah prevents other meetings like I I am so meeting up and they just it just gets I mean some some weeks it's just I I barely even have time to breathe. I certainly don't have time to take any action and then like this whole thing that it would do would prevent a whole meeting for me, you know? Yeah [laughter]. No. Yeah, for sure. Catch me up on this or you know. Yeah. No, that's that makes a ton of sense. I think uh Yeah. [ __ ] We already over an hour. Yeah, that was fast. [laughter] That was really fast. You want to do modified speed round? Yeah. Two or three questions. Um, so we like to end we just kind of pepper you with random stuff. What are what are some either favorite books or podcasts that you recommend either around energy or kind of tech side of things? Um, I like Darknet Diaries a lot. Love Darknet Diaries. Oh my gosh. And I think like if I ever have a different career, I want to be a social engineer because that whole thing is fascinating. Fascinating. It is so fascinating. I love the episode they did on the social engineering. My favorite book is War in Peace. Um, I read it three times. Um [snorts], but um I don't I think book and podcast got it. Shout out to Jack Reider and Dark Neck Diaries. I'm also a huge fan [laughter] about uh favorite um favorite social media follow. [sighs] I kind of try to stay off social media. Um I it's me being a very inquisitive, investigative person, like it just was too much of a deep dive. So I have my creeper account here or there um so that I can do investigations but been verified works pretty well for that too. [laughter] Okay, cool. So my favorite social media has been verified. Maybe LinkedIn I I like in general I do connect with more people on LinkedIn. So Okay. What's uh we'll do the last one then we'll wrap up. What's something um either interesting to potential customers or exciting for you internally that you guys are rolling out or working on at Energy Ogre? I think the chat bot's the next big thing that's coming out and so it'll be interesting to see the feedback. We have a whole UI to kind of record and attach it to everybody's profiles so we can look at it. So that'll be really interesting. No, and that's one thing I was going to say earlier is like y'all have done in my opinion an incredibly good job at not only I can you can tell when you use things how like focused the company is at automating or like getting things working like they should. And you know, for how much just email spam there is in everyone's inbox, y'all only message when you actually need to message me, right? Which the marketing team hates us for that, but but it gets critical if we can't get something. You can also destroy your uh what is it? Your domain. Yeah. You don't have to send it from domain to Yeah. No, I've been through all um it's account level at AWS. So, um, there you can have your different domains, but no, we have a whole system for the like a global suppression list. Follow kind of the best practices with that. But that's been a fun thing to deal with from time to time. Um, yeah. Why why is this person, you know, if we can't email somebody, it's a huge problem. Yeah. Yeah. Big problem. [laughter] But you y'all only email. Yeah, we try really hard done a very good job to to to keep it at a minimum. Well, thank you so much for coming. This has been a ton of fun. Where can uh where can people find you if they want to reach out? How do they sign up for energy yogurt? Um energy yogurt.com. Sign up. Go do it. And it's it should be pretty simple. Um yeah. Awesome. Like I said, I I mean this is firsthand for me. I've been a customer for a long multiple houses, multiple neighborhoods um in Houston. They're like I said, it's fantastic. You sign up, it's 10 bucks a month. They immediately save you 10 bucks a month on your utility bill. Pays for itself. Yeah. more so uh most of the time much more than that. And then when it when the contract expires, guess what? They do it all for you and you don't ever have to worry about it. Uh like Jake, one of our our co-founders, he was asking me,"Hey, who's your power plan?"I was like,"I I I don't know. It's energy [laughter] over." But his had run out. You know, the contracts expire and then they roll them over to monthtomonth. And I can't imagine how many people are paying monthly, which is just unbelievable. the amount of people that have never switched away from because people do think that TXU because a lot of those are lost leaders. I mean like I got into my new house and I got it 8.8 cent plan but it was for 3 months and like I was that 3 months goes super fast like and then all of a sudden it doubles you and it rolls into the monthly plan which is 20 cents a kilowatt or whatever it is. Yeah. [snorts] No, that's uh anyway so go check it out if you guys uh are looking for utilities. It's like I said it's awesome. It's super handsoff. It's super great technology because it's they got an app and everything. But thank you again for joining us. Of course. Thanks. If you guys like the episode, leave a comment, like and subscribe. Give us uh some love and we will see you all next time. Appreciate it. They y'alls.