I "Vibe Coded" A Tool For Homeservices | C. Suite EP 111

actually make real utility with these tools still comes down to human ingenuity and understanding the problem. We have made the solution simple to build, but most people don't understand the problem enough to actually build the correct solution.

It's finding that line for most people is hard when you're being bombarded by it. If you're going to use AI tools as a defensive measure, that's the wrong decision.

I don't think you need to take 4 years now because there's so much information and learning that I've tried to put out that people can follow the path. The same thing is true with AI. Yes, there's the leading frontier, the edge, is going to be the people that have money. The rising tide of AI allows software in general to get easier to build. And I believe we just need to make sure that we are part of the boats that are not chained down. All this fear gets us like so wound up versus being like every single technological revolution, we have figured out a way to solve better problems.

We are live. Welcome back. Another episode, The C-Suite Unfiltered with Mike Andes.

On the HSM Podcast Network.

On the HSM Podcast Network, brought to you by Mike Andes and Mike Andes Media. How you doing today, sir? Fantastic. All right, well, we're going to jump straight into it. Lots to talk about. Okay. We're going to talk about a topic that I am definitely an expert in. Okay. That is AI. I'm not an expert. I am a novice by all means. But I wanted to talk about AI and not just hypothetical, but actually something you recently built with AI tools and then get into what we have all heard, this buzzword of vibe coding. But I don't know to the extent what people have played with it, what, how that applies to small businesses and should people be concerned about it? Should they be doing it themselves? So we're going to start specifically. You built the mowing quote website, the Instant Mowing Quote website. What was it called again?

InstantMowingQuote.com.

InstantMowingQuote.com.

Perfect. Keep it simple.

Keep it very simple. Good SEO. InstantMowingQuote.com. And you built that via Vibe Coding. Is that correct?

You could call it that, yes.

Okay. So explain what that is to the people that maybe don't know or have heard the term like How did you just build a website via vibe coding?

Yeah, so it's not just a website, it's an actual application. And that's, I think, the difference between something like Lovable, I think was what vibe coding was the original thought, like even 2, 3 years ago. Now the tools themselves like Manus or, I'm thinking Xcode, but it's not Xcode. Cloud Code? No, well, Cloud Code, yes, but Codex, there you go. Oh yeah. These platforms now inside their tool, you can actually just build the whole thing. And what's interesting is it is full stack, which means it's not just the front end, which you would see visually, it's actually the backend. So user management, authentication, 2FA, all those other aspects. And so that really took a big turn, like late 2025 is when that really evolved a lot more. And so it's been really useful for me, not only just in like building the tool, but also understanding more about the technology, even at a surface level. So when I'm talking to engineers, et cetera, we are just way more, I'm way more informed and understanding some of their lingo. And also even at HomeWorks, the ability now for the support team to take really basic issues or bugs they find, and then we actually have an open claw that is basically doing the lifting of here's what the fix should be, then it goes to an engineer to validate it, 'cause there still are errors. And that's the part where it's a very slippery slope. Everyone's gonna hear about the business that did GOP ones, that did like a billion dollars with a few people. Well, like, realistically, they're the middleman for a GOP1 product, and they just used AI to, like, generate the Facebook ads, make their website. They were a label on a white-labeled product of GOP1s, and they made up, like, fake influencers and fake ads. Is that, like, good? I don't think so. Um, but all that to say, I think the news will hype the 1% or 0.1% of products that are actually landing and getting product-market fit. The vast majority don't. So like yesterday I saw an app that was just like, I'm like, this is just a wrapper on ChatGPT of like a prompt that they're putting on the backend. Yeah. And that's useless. And so how can we actually make really utility with these tools still comes down to human ingenuity and understanding the problem. I think what we have done with AI and let's call it vibe coding is we have made the solution simple to build, but most people don't understand the problem enough to actually build the correct solution.. And so this is why like even the App Store is getting flooded with all these vibe-coded apps and Apple's like cracking down on it cuz they're useless. Like it's, it's just useless because it's not solving a problem. And so, um, I am very optimistic that there's still going to be a place for really good software that can scale. But I also think vibe coding is super useful to be able to plug into those, that infrastructure and that foundation to build your own bespoke little thing that you've always wanted to build. So I've, I like, in my mind, I think I've seen several people do things like tracking door hangers. Right now it's like, you know, you put a Strava or like an Apple Watch, like track your crew, but then it doesn't like speak to your Apple Map or your Google Maps, which doesn't speak to your CRM. This is the type of thing that you could build yourself. You're— if you have a very bespoke way of doing estimating or upsells or tracking reviews from your crew, these are things that you can now build and then plug in using an API or MCP access to your software, into your CRM. And that's the part where I think people are going to try to go and and they're gonna get lost and it's gonna cause damage, whether it be authentication, payments, security, and yeah, so that's my general thesis on it.

Okay, so that went a lot of different ways. I wanna pull back a little bit for those of us who aren't experts, who are novices, maybe we're just dipping our toes in the water that is AI, Vibe Coding, OpenClaw, all these different things, all these different tools, APIs, like all these terminologies that everyone's learning on the fly. 6 months ago, I didn't know what API was, right? And like now, okay, we understand that that's like this backend plugin and I, I can explain it on a basic level, but I don't actually know how it works.

Mm-hmm.

So let's just, to use a, a very bad analogy, explain how you built the Instant Mowing Quote website, like to a 12-year-old, like it didn't exist. Yeah. So what were the steps you took to effectively build it in simple terms?

Yeah. So like first, The part that most people don't have is the past 4 years of building this for Augusta Lawn Care and fully understanding like what every single field does to conversion, testing so many times. Like I knew the problem like so well. And so that's the part that most people don't because like some people are like, oh, you should do this for window cleaning. It's like, no, no, I don't have the experience to know exactly what the problem is. So my solution is going to be half-baked. And so the first step to your point is like you, you want to use something that allows you to be able to just speak with common English that can build for you. And this is why most people can't define clearly what their problem is. And so the English that comes across to the software is not great, it's not clear. And this is why the best prompts, like prompt engineering is a thing because now if I can clearly identify the problem, what I want it to do, what the edge cases are, now the software can build it. The solution part is, is commoditized. Most people don't understand the problem itself. And so, like, even when I was talking to Jonathan Patochnik on last week, people missed the fact that, like, he had 50 pages of prompt for his little estimating tool he was trying to build, and he's never going to plug in or use in his business. Like, he was playing with it, he's just experimenting. But 50 pages of logic and what works and what to— don't do this— and, like, the parameters. And most people, that will be a one paragraph, like, build me an estimating tool where I can do X, Y, and Z. AI is not going to be useful. And so understanding the problem and having a clear prompt is the first part. And so I went into it first off being like, okay, I had like different levels of the problem and like where I can create a solution. So the first problem is like I see a lot of people that require a form and then I go measure the property, whether that be in person— worse— or they just want to go onto Google Maps, measure it out real quick and send it to the customer. So like, okay, first thing first, can I have it where that person just puts in the address of their customer and it spits out a price? Based upon using AI and just simplifies, or even where they do want to measure the property, but it's in one spot. It's console, like it's more of a backend admin tool. And so that way I could refine the AI to be accurate on square footage of the turf and using satellite images, using Regrid, which like scans, you know, has 140 million properties and all their information and like balancing. How do I weight between Regrid, which knows the property data, and then I can like look at an LLM and figure out, oh, Okay, everything about this property, how old it is, how close is it to a park, what are the size of the trees around it, what do the properties around it look like? But then also sliding towards, I can see a satellite image and pretty much figure out what the square footage is. And then where do I, how do I balance that algorithm? That's what is on the backend. And that, that is actually the backend sauce that most people would be like, why is it accurate? And, or why is it close to being accurate? And that's the actual backend sauce that I kept measuring with and playing with. To get as accurate as possible, but it really starts as an admin tool. Then it's when, okay, now I'm going to allow users to come in here and create forms. Whole different level. You have to make the backend, build the stack so it's full stack. And then the third part is like, now I want customers to see this. So now a whole nother level of like user interface. How's it going to look on the forms? You have a transparent background so it can work on any design. What's going to happen if there's white on white? Like, does it tell me like that this is not going to be visible? These are all the little edge cases that most people will not go through. And so I was able to get the product probably where it's at in about 2 weeks. Um, 90% of it, but the last 10% took several weeks and like iterations with the team and like throwing it through support and like let people beat it up, give it to people that like, or I know would use it, but like they want to have see what, where, where's the flaws and like getting all the little edge cases. So for example, everyone's like, okay, great. It's connected from Instant Mowing Quote to my CRM. Well, what happens when someone submits an existing customer, submits a form? You don't want them as a duplicate inside of your CRM. So what's the protocol to look at? Okay, if there's an existing email, do I create another lead inside your CRM? Do I block the webhook? Do I keep the lead inside Instant Mowing Quote but mark it as a duplicate and then email the owner? What happens if the email is different but the address is the same? So these, all these little edge cases no one thinks about. And those are the problems that people will say, oh, this is AI value coding trash. And the vast majority of people, that's what they will build because they won't keep pushing it and pushing it and pushing it. I'm also in the unique position where pushing it to that last 1% makes sense because thousands of people are using it. For most people, it would be better. They'd be better off spending those 5 weeks getting more jobs, creating systems in their business, figuring out better protocol for advertising. Like, that's going to be a better return for their business. And so that's why people hear me say what I just said. I'm like, I've got to go spend 5 weeks. It's like, no, no, just go get more customers and make sure your crew is trained up better. All right, real quick, I don't think I really clearly defined what instantmowingquote.com is inside the video. Essentially, it's a way to be able to give quotes to your customers from your website and build custom forms that you can customize exactly how much AI you want to use to get the quotes. So you can have it where it scans the property using satellite and gives them a quote via AI. You can also have it where they have to actually measure the property with the pins. You can also have it where it's small, medium, and large, and they choose an image that looks similar to their property and give them a price. So it's super customizable. You can add signature blocks. You can designate certain zip codes that you do allow service, so you don't get leads outside your service area. There's a bunch of really cool stuff, and I believe that this is actually just the start of what people are going to be able to build on in HomeWorks. As we are about to open up our MCP, think of an MCP like a USB-C connector. It connects different devices. Same thing is true with like AI tools and software. It's the MCP access or the APIs that connect everything. And so as you build out something in Claude Code or Manus or in Codex, the ability to connect directly into HomeWorks is going to be super powerful. It's coming soon. I can't wait to share with you. It's already in the hands of some of our users that are technically inclined, and it's showing great results. And so we'll refine it, make sure there's lots of good parameters in place. But as you build these tools, connecting it to something like HomeWorks will be super valuable. So definitely, if you haven't, go check out HomeWorks, become a free member, check out Instant Mowing Quote. It's free also. And, uh, let's jump back into the, uh, podcast.

To extrapolate a lot of what you just said, if I were to oversimplify everything, is that vibe coding is seen as this like disruptive new phenomenon, financial like goldmine of I can build this, I could sell this, I could use this to optimize, right? Whatever that means. But what you're saying is very interesting to me, using JP as an example as well as yourself in Augusta, is that you both had to master this like bespoke thing that you wanted before the technology existed. And what I find interesting based on everything that you just said is that a lot of people are looking for that end result. Oh, I want it to look like this, but they haven't put the time or energy into figuring out how to build it without vibe coding, without using AI tools. I think that's a very interesting takeaway. So do you feel like, like this— let's just use Joe Blow, someone who doesn't exist, who just got into lawn care. And maybe they have a tech background, like they know how to use AI and these tools, do you think they could have built the same website you built in 5 weeks?

Uh, potentially, yeah. But it would, it would miss all the edge cases because like I know which ones are going to convert the highest and which one's the most accurate. Like I already have all that data. So could someone build it? Yes. Um, would I suggest someone build it? Because like I'm giving it away for free and get a couple thousand users so far. Like it's, it's growing. But like, people fail to understand just how hard it is to get $1 from people. And so when people talk about AI and these things that can vibe code one person, build a billion-dollar business, I just want them to go build a product, solve a problem, and try to get one customer to pay them. It's not easy. And so, uh, the level of like how good you have to be now in order to exchange dollars with someone is just getting higher and higher, which is good for the customer. And it is good for the end user. And that's why, like, software will become better in time and the better companies will get better even more. Like, it's just going to compound. I think the winners will continue to win.

Well, and that's what I was going to switch to. Not necessarily macroeconomics, but it feels like it's almost the rich are going to just get richer. So what are your thoughts on that? Like, you built this tool because you have this mastery— I'll use that term— this mastery of home services, lawn care, quotes and estimates, and how to look at square footage and properties and maps. Like, all these systems you built out over Augusta before you had these AI tools. So you had that mastered. Now you can plug it into an AI tool in 5 weeks, get it 95, 99% of the way there to what you want, tweak it along the way, keep constantly improving it. Like hypothetically, someone can make the argument that like, oh cool, so Mike is just going to get richer because of this. So what does the average person do? Someone like myself, someone that maybe doesn't have a business or wants to get in business, like what do they do to keep that delta as small as possible from the rich getting richer, the, the mass, the people who have mastered these skill sets now plugging into AI tools to keep expanding that gap. How do they make that gap as small as possible? How do they keep up?

I think just constant improvement. Like it's like saying like someone that knows algebra is definitely going to have a better ability to be able to do equations and make progress in mathematics than someone just learning how to do ABCs and learning their first few digits. So should that be that the person that's just learning how to do their numbers and, and the alphabet should not start? No, but they can't skip those steps to be able to ever learn algebra, right? And for the person that's learning calculus that's ahead of the person algebra, it doesn't mean the person with algebra should be like, oh man, I gotta give up. The guy with calculus is just gonna keep beating me up. No, you gotta keep learning. And so I would hope that people, if they're, you know, 5 years that they've, they've done the work that I've done and ideally done it faster. Like I took 4 years to build a business after like beating my head against the wall to finally make money. I don't think you need to take 4 years now because there's so much information and learning that I've tried to put out that people can follow the path. The same thing is true with AI. Yes, the leading frontier, the edge is going to be the people that have money. I also spent hundreds of thousands of dollars developing an AI tool that we scrapped and just started using open source models. In terms of answering the phone at Command Center. I did all that work to fully understand all the problems with that technology before anyone else did for the, for the advantage of the owners so that we weren't just like blindly like plugging in random off-the-shelf tools. That being said, I was able to take my resources and distribute it to the people through my knowledge. And so I think the same thing will happen with AI is like, yes, the rich people right now can have the money to spend gobs of money on tokens, gobs of money on training these, these, uh, AI agents and, and building these tools. It's not cheap. Like the, the subscriptions are not, the tokens are not cheap. And so, um, yes, but also it's, it's up to the individual to take those resources or take the learning from those people and then go execute on it. And so the same tools that were $1 million 3 years ago are now literally probably $100 a month.. And so take what, what is being given to you that other people are paying gobs of money for, i.e., private equity and venture capital are pouring money into this like who knows what, and they're spending billions of dollars and trillions of dollars. Take the tools, learn them, and don't stop learning math because you're not algebra yet and you're still learning your ABCs.

Yeah, I think it's an interesting point. And to clarify, or to help the person listening right now, is when when do you decide, do I outsource this quote unquote to AI? Do I build the tool? Do I invest the 5 weeks? Because I do see there will be a good return on that time spent on outsourcing to this AI thing, Agentic, Claude Code, whatever you're using to take certain things off your plate or build certain tools that you think will have a return versus just focus on mastering the task at hand. Like where do you think that line is for most people or how do you define that line? Do you see what I'm saying?

I think that vibe coding in general is probably 1% of the people listening.

Yeah, fair. I think that 90, but we're all hearing it. Like we're all getting ads for it. We're all like, oh, I could use that. We're all getting pounded with videos.

It's the same thing with cryptocurrency. It was the same thing with NFTs. Like there's always a hype cycle.

So yeah. Where's the line for someone to be like, hey, that's noise. Put your head down, master the thing you, you're currently building. Don't worry about building this tool, this agentic thing, this coded whatever. Product, just master the thing you're doing, and then eventually you can build that tool if you need it.

Yeah, build a bigger business to where now the pain that that is causing by not having the tool actually becomes dollars and cents large enough to go fix it. Like, you should not spend 5 weeks trying to figure out some tiny little thing. Like, no other software is going to build it. Like, I got to build it myself. If you spend those 5 weeks, figure out Facebook ads, you could triple your business next year. Figuring out that tool might optimize your, your admin flow by 10%. So a lot of the things I'm seeing people build are very defensive measures. I gotta save here, I gotta save, scrimp here. So they're, they're spending all this time on defensive measures versus they should spend that same time on offensive measures like hiring and creating systems and advertising and doing a better marketing flow and creating a landing page. Like spend time on that, then they could triple their business. And then when they triple their business and they have a larger organization, they have multiple people reporting to them, multiple management. Now all of a sudden the AI tool makes sense to spend more time on. Time with to be able to get some return on it. Mm-hmm. Um, so I think the vast majority of people right now are very much overhyped on AI, and as everyone's being pushed, like, they almost have this FOMO. Mm-hmm. And whenever you feel FOMO, you should basically just always slow down.

Yeah. That's interesting. I, I just, we're just getting bombarded with it. Like, I, I recently got an ad, you know, targeted ad that was someone, can't even make this up, someone's selling a course on how to prompt AI tools to effectively code on your behalf. Like a course on how to prompt AI tools to code on your behalf. Like we're talking, this is late-stage capitalism at its finest, but it's, we're being bombarded with it. So like there are times where I'm like, do I just need to do more outbound? Like from a sales perspective, do I just need to be quiet, put my head down and do more outbound, reach out to my warm leads, make more phone calls, send more text messages, Or do I need to build this really cool tool that will be great for prospecting? This kind of dashboard that I've showed you before of like, look at prospecting. And if we can just figure out the metrics behind our sales, then it becomes this tool that we can plug into some code, whatever, right? Like it's— this is all half-baked. Like, I'm not the expert, but like, it's finding that line for most people is hard when you're being bombarded by it, like the crypto example and everything. So I think your overall point is if you're going to use AI tools, as a defensive measure, that's the wrong decision to make. Like a bold claim based on everything you're saying.

I think it's like, what do you— how do you define AI tools, right? Like, I think again, Vibe Coding is very—

if you think you're going to build something, yeah, that's going to build—

is different.

Take something off your plate, right? As a defensive measure, you think that's the wrong move? Generally speaking.

Generally speaking, yes. I can— I'm generalizing because there are people that have built a bigger business. They can get a lot of leverage from saving themselves time and 50 other people.. But for the vast majority of people, building a better business means get more customers, do higher quality service more efficiently. Do that and you will have a business that doubles in profit very quickly. Mm-hmm. Doubles in revenue very quickly. And I don't believe, I do not see right now a vibe-coded tool allowing you to do that. Mm-hmm. And if it exists and it's so easy to build, eventually it'll be where you can get it off the shelf and you're gonna pay a couple hundred bucks a month for something that you spent weeks, months. And then once you, you're like, I'm going to build my whole soft, build my whole business on top of this thing I built. So what happens when it goes down? And you're going to go get an IT person, like you're going to get someone that's an engineer and they're not going to cost $200 an hour like they do on payroll. They're going to cost $1,000 an hour to come be your emergency resource to pull in for a day. So you're going to pay $8,000 for them to come fix the problem because now AI can't figure it out or there's some sort of issue, some authentication error, security breach. Something, all of a sudden things went down, it— you can't fix it. You're just going to swap what you're currently paying $50 a month for, for an $8,000 once a year bill. And I think for the vast majority of people right now, it's a distraction and they don't clearly— they do not yet know their problem enough to actually build the solution. I think the vast majority of people should use the tools, not necessarily build with them. And what I mean by that is like the vast majority of people just have not yet figured out how to Give your financials to it. Ask where the holes in your financials are. How can I improve this? I'm having this problem right now with this employee and I can't figure out how to be able to solve it. How do I communicate better with this employee? How do I follow up on this estimate? Here's this email from a customer that's freaking out at me. What do I say? Most people aren't even there yet. So to think building tools is going to solve your business when you're not even AI curious enough just to communicate with it and become a better leader, a better marketer, a better entrepreneur, like What are we talking about building software for?

Yeah, fair. So pivoting off of that and more specifically towards like Homeworks, you have the, the context of like you have built a CRM. So how disruptive is this new vibe coding, cloud code, open source type technology that someone can go and build a CRM? Right. And we've seen it. We have someone on our team that's basically building a bespoke CRM for the franchise to see if that's something we can use. So what do, what do you see as someone who actually owns a CRM, uh, or as the founder of a CRM? Like, is this going to disrupt you and your industry drastically, or is it more the next piece of the puzzle that advances all CRMs?

Yeah, I, I think a rising tide floats all boats. So I believe that AI allows the solution space to become better. I believe I have inherent advantage when it comes to data and figuring out the problem, and I believe that is the, I think that's the real root of creating great tools is understanding the problem. I have that data. I believe though the rising tide of AI allows software in general to get easier to build. And I believe we just need to make sure that we are part of the boats that are not chained down. And so we are absolutely using OpenClaw with our support team. We are actively, every ticket before it gets touched by an engineer is using AI in some way. Um, but I can, I can tell you like 30 to 40% of those tickets still need engineers. Like, nope, it messed up here, it did this wrong. And so if you're gonna go just use AI, I know all those bugs will get through, those little glitches that eventually will cause a cascading issue, or all of a sudden you're get— your software is being held, your, your customer data is being held ransom. Like that's a problem. Mm-hmm. And so for example, inside the franchise, we can build something because we have a closed off system. We give access to an owner. We give access to a member of Command Center. We give an access to employee. When you say, hey, open up my, my website and log in, sign up now, that is a whole bag of tricks that most people are not ready for. And it becomes a whole security thing, like payment data, like everything. As easy as to build software with AI, just think about how easy it is to attack that same code with AI. And so the more resources and data and stuff that you put in there, you're just basically more and more vulnerable to someone else that's like, the bigger you get, the more vulnerable you are to someone else being like, ah, I think that's a good target.

Mm-hmm.

So yes, you build a $10 million business on the back of that software that if you build on vibe coding, very good chance an engineer's like, okay, fantastic. They're my competitor. Mm-hmm. I'm gonna go get an engineer, spend a few thousand dollars and have them hack into my competitor's software. Mm-hmm. Take all their customer data and lock it up. Mm-hmm.

Yeah.

Like why do you think Mythos from Anthropic had to go to the government and 6 of the biggest tech companies months before it could go to the public? These are the type of things that they're going to go lock down all their systems, but it's going to get in the hands of someone. And because you have— don't have a whole team making sure you're actually up to date with this stuff, you're going to be vulnerable. And so I don't necessarily think it's like a bad thing or a good thing. I see this as a rising tide will float all boats. We're going to go through this SaaSpocalypse where valuations will get reset. But if you look at like Salesforce, they have like the best numbers ever. And so even though their valuation is going down, like revenue's up, number of customers joining is up. Stripe, like everything is up.. And so it's a hype cycle. And in every hype cycle, I've always learned to slow down and do the basics. Yeah. And so we're just trying to make it more reliable, make it faster, build the features that someone will otherwise will say copy Homeworks dispatch feature and it'll do its thing. And they won't think about the thousand edge cases that we've had to go. What, what happens when someone selects 5 crews? What happens when someone selects 4,000 visits to the backend and then it crashes? And it's halfway through loading all those visits and the change, does it go revert? Does it go forward? Like, does it warn them? Does it like execute on the ones that went through versus not? Like, these are all the edge cases that people like Salesforce have fixed millions and millions of edge cases. And that's why you pay them because I'm willing to pay the $5,000 a month for Salesforce versus going out and trying to figure out those millions of edge cases that make their software work. We build custom websites for home service business owners just like you, and you get a 15-minute free consultation about your website because our team, on average, when someone joins as a client, we increase the number of leads they get by almost double, 80%, within the first 12 months. In fact, if you go to our website, you can actually go through and see all the different websites that we build for a whole bunch of different home service businesses that have taken their business to the next level. Go check it out and book a free consultation call today.

It goes back to the mastery point that like HomeWorks is, is mastering a home service CRM and you have that and you have the security protocols and the edge case studies that it works for you. And we use a software across the holdco that I'm intentionally not naming, but I had a call with kind of our liaison between that software and the holdco and I was like, hey, what are your thoughts on this whole vibe coding thing? Like, how are you guys handling this?, and he said the same thing. Hey, we have the source material, we have the data, we have the security, we have the edge cases. We think if we can open source certain things, it will just amplify features we've already built or make our, the features we've already built more bespoke or more custom to what you want. And I showed him this dashboard that I was building and he effectively was like, oh yeah, eventually, like I can't give you an exact timeline, but within 12, 24 months, like you'll be able to take X data that you want from this software and move it and build this custom dashboard that you just showed me. Effectively, he was very confident with that. And so that felt like a rising tide floats all boats moment. But he also did express concern that, hey, the companies that move faster are going to also be the winners in the end that can adopt the more open-source, bespoke, customizable features.

Yeah, the ones that move fast also will be the ones that likely you'll hear about in the news where all of a sudden their security, their data is gone. Right. So there's a double-edged sword. Like, that's why, like, yes, there's all these companies that are big talking all their big game because they're on the leading edge. But what you're not seeing is like Google shuts down most of their initiatives. These big companies like, uh, what's the stinking, uh, uh, Sora. Oh my word, we're all video editors are toast. And now they've closed down the app, right? So like these, these, this is a hype cycle and staying levelheaded in terms of like how does this actually solve problems for the customer is always the thing I fall back on and just like get really religious about that. And so I do think one of the things that we are really focused on right now, and we now have like an MCP, We have just several users using it now, but like our goal throughout 2026 is open that up so that as you build your little bespoke tool, you can connect it into your HomeWorks account. You can do your special referral program that you have with your specific neighborhood. You can build it and then plug it into HomeWorks. And so like some of the stuff right now is so cool. Like you ask Claude like, hey, who's behind on their schedule today? It's like, so-and-so is. Hey, I think you should move someone over here. Here's how many budget hours you have this week. You're understaffed. Like Do you be able to talk with your agent to be able to, when you wake up in the morning, it has a brief of the jobs, which ones it thinks are you're behind on, which employees need more help, more training. This is the type of thing that like I think will, uh, unlock the imagination of people to be able to say, I could build something and now connect to something. The source of truth is solid, but there's so many edge cases in that. And we've hit it in our testing of like, okay, what happens when someone says do this function 1,000 times and it hits our database. Like, it's just an MCP. It's like plugging in a, like, live wire. It just, it's gonna happen. And like, how it affects our databases is important 'cause it can, it can slow the whole thing down.

Yeah.

So, um, that's been really fun and interesting. And Dan, the, you know, CTO, he's like very, very pro AI. I'm very thankful for whether it be like figuring out how the support team can handle more tickets, whether it be, um, you know, which is even last night we went from like Gemini, we tried Gemini Flash 'cause it just came out., and a lot of its benchmarks are higher than what we were using with Opus, which is Anthropic's model. And so like we tested out last night, it's like, okay, it's good. But like there's a couple things it doesn't understand yet. Let's try it again in a few weeks once they've ironed out some of the kinks. So I'm very thankful I have a CTO that's like so pro AI and I'm not like pushing them to do it. And so yeah, he's awesome.

Nice. Well, we'll switch topics a little bit. One thing I have to say, I've been holding back a smile because we'll have to clip this. But when you, when you were really getting on a roll, you were like, save 10% on admin cost, optimize your, your admin flow by 10%. You really hit a trump there. You really hit the 10% on admin cost by building your custom tool. But anyways, it's going to be huge. Now it's going to be huge. Not to get political, but let's get political and stay on AI. Let's talk data centers. All right. So, or data, data, however you want to pronounce it, vase, vase, you know, if you're one of those people. The 3-member Box Elder County Commission approves a massive data center in Utah. The project is being developed by O'Leary Digital, Mr. Wonderful, and Stratos, with investors expected to fund the buildout while critics are objecting to the use of a favorable tax arrangement and incentives through Utah's military installation development structure. The energy use of this project is reported to be up to 9 gigawatts at full buildout with the power generated on site, an amount described as more than twice of the state of Utah's current power use. So any initial thoughts on this data center, or would you like me to keep kind of setting the scene?

No, I think like every time we've ever built a new technology, there is, there's trouble, there's, there's issues with it, right? Whether it be building cars like, oh, we need to like regulate how fast these things can go and we need to build lights so that people stop when it goes red. Like there's going to be certain legislation that needs to come out to put safeguards in place. And I think the biggest thing that I see as a problem is energy consumption. And we are not, like our grid is not built for in the US. Like what are we gonna do? We are like way behind on nuclear. We're not really good at any sort of like power. We've shut down so much coal mines, et cetera, like our coal plants, et cetera. So like, is that good or bad? I don't, I'm not really here to discuss that. It's more a matter of, the models will get more efficient. I don't think we're gonna be building data centers at this like incredible rate that we keep hearing about. Like most of these data centers get canceled by the time they go from announcement to actually building. And so all these infrastructure plans and like billions of dollars and like, okay, did it actually materialize? And it, it's not happening as often as people think. Mm-hmm. Um, and so a lot of it's news and it's hype. I also think the models just get more and more efficient. And if you look at the curve of like Moore's Law in terms of how much we can, you know, storage of memory, I think the same thing's gonna happen with compute, which is like in 10 years it'll be like 1,000 times cheaper than today. And like, well, people are using it more. Yeah, but it's like, I do believe we'll catch up with the rate at which we are developing. And so if you even look at like Cerebras, what they recently, like in their IPO is like, can we make a larger wafer? Can we pack more into this thing to compute better?. And I think that, uh, that will outpace even our demand, which is like, it's, it's like doubling every year. But yeah, if we can make the, the technology 10 times more efficient, 100 times more efficient, 1,000 times more efficient, it's the same thing. Like this thing used to be an entire, uh, room, right? So if you looked at like how the, the adoption of computers, oh my word, we're gonna literally need to fill the entire state with just com— like a computer. No, like it's now., what, 200,000 of these could fit in this room? Yeah. And so that, that form, that function comes down. The, the efficiency of this goes down. The fact that it has what, like, uh, I think like 4 or 8 gigabytes of RAM, like just, just not long ago that was incredible speed on a computer. And so I think they'll just get better and more efficient and the market is positioned in a way to give reward to the people that are figuring that problem out. And so when all that money and attention and mine, like everyone's brains go in that direction. I think we're just gonna get so efficient. And that's why I see, we also seeing a lot of these data centers just being closed down. Like the projects, they're not gonna get built out. And by the time they get built out, it's like 4 or 5 years, some of these things, they're not gonna be like needed as much. So, oh, that's my, my thought. I think I could argue it the other way.

Yeah.

In terms of like the energy consumption, I do, I do have a problem with a big company coming in, building a data center and then like energy rates going up around them. Like supposedly we, there's now the bill been passed where like that can't happen. But like, where are they getting their energy? There's a finite amount of it. So like, who's using it? But it's funny because even like the whole like water thing, like that was the dumbest argument ever. And so like, people just love the news. Yeah, it's just another thing that gets us all emotionally tied up.

Yeah, 100%. And I'm not like overly invested into it, but I thought it was an interesting topic and I appreciate you sharing that perspective. Just to share a little bit more of the details about the specific one in Utah, you know, people are mad. You hit the main bullet points of like power consumption, water consumption, subsidies to big tech versus, you know, water consumption.

Things like the science behind it is so faulty.

Yeah, generally, like, I don't have a hard and fast position on it. I've heard many reports saying that, so I'm not getting into it. But this is just like, why are people mad? Big tech, local disruption. So this is, this is an important point. Before I get to it, though, like the builders are saying, and Kevin O'Leary's point, I listen to a long-form podcast of him effectively defending why this is important. And like one of the things that kind of bothered me was like he kept going back to AI competition, national competitiveness, and that like we have to beat China. But he didn't say why we have to beat China. And it's like, that's such like old school kind of— I hate to say it, don't yell at me— boomer mentality of like, well, it's all about national competition and national sovereignty and our security. It's like, but how is building a data center like going to help us beat China? Like, I need you to get me from A to B. Like, don't just say like, China is bad, therefore, like it's a false premise.

Well, let me, let me explain their positioning. Yeah, please. 'Cause I agree with you. Um, but their, their positioning on that and how he would like, if you boil down a little more is like, okay, if I don't have a data center, I can't govern more compute towards the model. Correct. For example, recently, uh, Anthropic dialed back one of their models in terms of just how much compute it was having. 'Cause they were limited, like they didn't have enough compute. Now they just gone, done a deal with xAI. To have a whole bunch more compute. But before that, like people were, you could tell it was having problems, it was struggling, it had a lot of outages, it would do way more hallucinations, it would do dumb stuff sometimes 'cause it was limited by compute. So having the data centers, I need the data center to be able to run the compute of the model. Like when you type a thing in, it's doing something at that data center.

Yeah. To beat China in a way, in a technological advancements in face development. That's why I think it's smart.

These people are using, again, our emotions being, okay, well, I know my power's going up, but I want to make sure that I don't have a nuclear bomb being passed over here from China. So like, that's the mentality. And it's stupid in my opinion, because like, if you see what they're saying, it's like, they are like, AI is going to beat us. And then they're going over there with Jensen Huang or whatever his name is. Jensen Huang. Jensen and Elon, and they're making deals. So like, they're speaking out of both sides of their mouth. It's a big political thing. But if I can get you to be afraid, of them and then say our only defense is to beat them in the AI arms race, then, then now you will give me moral high ground to be able to do stupid stuff like building a data center in your backyard that hinders your view and wipes out all the wildlife.

Yeah. So what would you, if you were in charge of, say, if you were the mayor of, uh, Box Elder County.

Oh, did they say that?

I don't think that's how it works. Like, what is your position? Are you, are you demanding certain, like, requirements from the data center, from the company's investment back into the local community. Because one of Kevin O'Leary's, you know, I would say more moot points is like, well, it's going to bring jobs. Well, a lot of temporary jobs for the buildout and then a handful hypothetically. And I don't have an exact number, so don't, don't come at me on this, but a handful of people would be there to maintain, run, operate security, you know, of the building and of the center. But if you're a local person, you're saying like, you're disrupting our, our power grid, you're disrupting the nature, like it's only bringing temporary jobs, and temporary jobs typically can harm communities. If you look at certain data on like when there are these massive buildouts, they come in and they almost like ransack the local community. And then when they leave, that community is somewhat in shambles. And that's just like anecdotal, but there is some data behind that. So what would you do if you're in the position of like, okay, we have like generally a strong local economy, we don't need a data center. Like, what's your position on that? Like, bring it, bring it home.

I'd be tough. Like, if I was the mayor, and like I had my family growing up there, I'd be also like, okay, is this data center gonna become abandoned in 20 years? And now we've built our whole budget on the fact that this thing brings us millions of dollars of tax revenue and now they're bailing cuz they created this new model that fits on a phone instead of a huge data center.

Mm-hmm.

Or doesn't require a millionth of the compute, or they use open source stuff outta China that's just like, is just fine. Will they even be useful in 20 years? The other alternative is the other idea is like, well, compute, like the usage adoption rate is going through the roof. So yes, even if we get way more efficient, we'll use it. I would be mostly worried about that, like, is it gonna become a ghost town? And then I build my budget based upon their tax revenue. The second, the more immediate thing would be energy consumption, and I would be looking for like, you gotta prove to me you're creating that much energy. So whether you're gonna do a wind farm or a solar farm, or you're gonna do energy stuff that powers the grid during the night for us, whatever it is, like, I would be looking for some sort of net-net neutral to our, power as a, as a community member, I'd be wanting some sort of guarantees around my energy bill. And so I'm like, hey, you can move into town, but just FYI, you're going to make sure everyone around you that you're plugged into their grid and you're guaranteeing their prices don't go up. And whether that be you are supplying the power, like, hey, my other thing is like, why are they on the same grid as all these people? Get them out. Get them out of here. Like, if in the middle of Utah, there's a lot of places in Utah where you will not see that data center, right? And there's plenty of really good spots, like even in Arizona where they've built these, where it's like you have your city, but then you go like 10 miles out of town and then you see the data center. No one wants to be there anyways. It's the middle of the desert. And if you're gonna build, spend a billion dollars, you can trench some power lines or build some power grid 10 miles back into town. And so, um, I would love them to become more energy dependent, energy independent. So that this isn't even a conversation. I do think it's necessary to be able to progress this technology right now though. So like, I don't really, I think, I think it's good that we have democracy where this is gonna get fought out at the local level. And I think that a lot of local residents have a right to be a little bit worried about what it's gonna do to the spike in jobs and what happens in 20 years. Is this, is this gonna be around in 20 years? Do we, will we need this data center? And they also are hearing all these negative, illustrations of what AI does in the news. So-and-so laid off 30,000 people 'cause of AI. No, they didn't. They overhired by 200,000 people and they're using this as a scapegoat 'cause now when they fire people, their stock price goes up. In the past, when you fired people and laid people off, your stock price went down. So what do we say? We're getting more AI efficient. No, you're not. You have a bunch of people doing nothing. They've been twiddling their thumbs and working 4 other jobs on the side. So like you're just getting back to being efficient, but you're gonna use the AI thing as a way to, booster your stock price. But, but they're hearing that and AI is the bad boogeyman that's going to come take all our jobs away. And it's somehow been the brand when in reality every technology has brought more jobs and like it just doesn't make sense. So like we're all buying this fear. And so either A, you're going to be the politician that is going after everyone's fear, which is energy is going to be crazy expensive. All of our wildlife is going to get killed and we're not going to have any green spaces. You're gonna lose your job and we're gonna lose the AI, we're gonna lose the AI race to China. And all this fear gets us like so wound up versus being like every single technological revolution we have figured out a way to solve better problems. And so like, do you wanna go back to tilling the ground with a hoe to be able to get food? Or do we just use this thinking tractor that will be a million times more efficient and you'll be in an AC vehicle pushing this like this instead of like, rack-breaking labor for thousands of people. And so we will just get more creative as humans to figure out better and bigger problems. And that is like how we evolve and how we get better and how like our lives generally improve. But again, all the news lines, uh, sell and click much better.

Yeah. Yeah. And I think from, I mean, to, to use a, a word that can have certain connotations, but from an evolutionary perspective, like the rate at which technology is developing is like the human brain genuinely cannot keep up with that. Like we cannot address to the environment around us from an evolutionary perspective as fast as it's happening. Like quite literally, I was listening to this podcast and I wish I could credit it, but it was like, you realize that just over 100 years ago, the only thing we all cared about was having enough food and not dying.

Yeah.

Like that's like it as an economy. Like just like roughly 200 if you want to be more realistic. It's like you just needed to have enough food for your family to live and like Hopefully you didn't get a disease that killed you. Yeah. And now 200 years later, like airplanes, AI, like all these things, all these technological advancements, you have the Industrial Revolution, the 5-day work week with Henry Ford, like all of these jobs, the tractor, it's like, oh, it's gonna take away farming. Now AI, we're doing the same thing. And I think it's just, it's hard for us to keep up. Yeah. And it's scary. Like it is in a lot of way. I think a lot of people are genuinely scared. I try to be more optimistic, more of like a futurist in that way. Like I see a lot of positives,. But I understand when people are scared, they're worried, they don't want these data centers, whatever it may be. I fully understand their position. Yeah.

Like we have to just look back in time and like look at the data to realize that every few years we have another fear. Yeah. Right. Will it be COVID? And then we look back, we probably need to do some of those things. Whether it be 5G, like that was definitely—

I remember 5G was a weird one.

We are literally all gonna get fried. 5G towers. Yeah. Right. Like every few years we're just gonna find another thing to freak out about. And so this is why like I'm so against listening to the news because like I am not plugging myself into that complete game on my cortisol levels. And like, I don't do well when I have like, I'm like, that is going through my brain. Like I've said this recently to some of our team around the AI thing is like, no one's getting fired. We're hiring more people. And I, the reason I do it is because like fear in your system prevents you from discretionary effort.

Hmm.

And so like it prevents you from thinking creatively. Like when you're afraid, you're not usually thinking very well. Like, you're not— you're like, you're freaked out of your mind. You do dumb stuff sometimes and you're not making logical decisions when you're not well rested. So like, let's focus on like getting rid of the fear. And so like, again, I think 99.9% of people should be unplugging from the news because they just— their job is to make you so afraid. You have to pick it up. You have to listen. You have to read. You have to stay informed. And somehow we bought into this logic like you need to do that in order to be like a good citizen. It's like, no, you don't. You need to take care of your family and help other people. And it has nothing to do with knowing what's happening in Korea.

Yeah, fair. All right. Well, that's AI data centers. That's Vibe Coding. That's all of it.

That was— we just hit all the most, like, controversial topics.

Yeah. Yeah. So leave a comment. Tell us if you hate it, you like it, whatever.

Oh man, you're getting roasted.

We will wrap up with some not financial advice. Do you have any for the people? Would you like me to start?

Go ahead.

I will start. So a podcast that was great, very much a futurist podcast, a positive outlook on the future. Is Modern Wisdom, Chris Williamson with David Friedberg from the All In podcast. They talk all things AI, data centers. They talk, David Friedberg wants to build a base on the moon as like the hub to then get to Mars. It's interesting. Like it was very intellectually stimulating and like a fun conversation. I didn't take it too seriously, but I thought it was like, hey, this is someone who has like a very positive outlook on the future.

I like this. He does.

Let's talk about this.

Like, it's so funny.

We need more of that.

What's the point of getting like building a thing on the moon and Mars? Realistically, there's only one real reason. Just in case Earth gets hit by an asteroid.

Yeah.

Like, can you imagine, like, that is like the premise of a lot of what Elon builds SpaceX around? Like, obviously you actually make money using things like satellites and AI, but like, we're trying to make it as extraterrestrial. What's the reason? The use case is that if this Earth gets hit by an asteroid, we got somewhere else to go. Yeah. That's basically the whole business model. That's crazy. Like, that's why I think it's so funny. All these billionaires are talking about getting to the moon and getting to Mars. Like, so what's the real use case? Like, so best case scenario, I get to survive and live on Mars. Yeah. Woo-hoo!

In a dome with artificial weather.

Let's go. I am so excited about that. I think let the asteroid hit me, bro.

Fair. I'm kind of with you on that, though.

Just take it. But you know what? If they want to spend trillions of dollars blowing things up and sending things to Mars, all power to them. They got more money than You have no idea what to do with.

Uh, so that's my non-financial advice. Listen to that podcast. Okay. Modern Wisdom with David Friedberg on.

Let's go. Non-financial advice, dude, I don't really have anything. Oh, okay. I'm gonna, I'll do it right now. So if you have to sit down for a long time, get a cooling pad. This thing was like $20 on Amazon. You push the button aside and you get literal fans in your butt, just cooling you off, making sure you don't get sweaty. So when I'm doing, you know, a long, uh, recording session, I turn the bad boy on and you can't even hear it. It's on right now.

I respect that. That's a great poll.

All right. Actually, people probably been wondering the whole time and saying you can see it.

Yeah. Oh, that's all right. It looks good. All right. That's all we got. Like and subscribe.

Leave a comment.