Chad (00:19)
Greetings and welcome back to Zac Cast. Finally, we are back in the saddle after a very long hiatus because Patrick refused to talk to me on this podcast last football season after his Aggies started off what, like eight and He every week we would talk about podcasting, and every week something would come up. And I know one hundred percent in my heart the reason is he did not want to talk about college football and jinx his season.

Can you confirm or deny that, Patrick? thank you. You have denied that, but now now it's on the record.

Patrick (00:48)
That's correct. I I won't deny that. I jinxed.

Now it's on the record. I it is it is true. I did not want to speak during college football season. The Jinx was not the podcast though, Chad. The Jinx was me. because the two games that I spent money to go see, both Buku Bucks, went to Texas to see us play Texas at Texas, which by the way, your fans are legitimately terrible. I mean, just I I I don't even know which do y'all train those people?

Chad (01:10)
Bucku Buck's going to.

Patrick (01:26)
I I'm not sure. The the frat daddies at Texas were a lot for me. we lost that game. The drone show was cool. and then I paid for, yeah, the college football playoff game. Yeah, the Miami game. So

Chad (01:39)
Playoff tickets.

Which one was more

disheartening? The playoff game where you didn't score a touchdown or losing to Texas again and getting knocked out of the SEC championship game? Again.

Patrick (01:53)
Yeah, you you probably know

the answer to this, but

I wasn't that upset about the loss at Texas. I was pretty upset about the loss to Miami. I I was upset about the loss to Miami because I felt like our offense was a lot better than that. And to basically be held I can't remember, I mean, look, we're so many months away from this now, but well, we were in a red zone, what, like five times and we couldn't score. I mean, it was it was pretty rough. it was it was not our greatest offensive performance. it was a fantastic defensive performance, but

Yeah, we were just I I felt like we were better than Miami and I felt like we had the team built to go win. the Texas game was a rivalry game, so I mean, it was as expected. It it is what it is.

Chad (02:41)
Let's not belabor

this. We've spent now, it's been literally it's been almost an entire year since we have been on this podcast. Let's not keep everyone waiting for the actual content. If you want to keep reminiscing about your the season that could have been, we can save that for the end. How's that sound?

Patrick (02:59)
Yeah, no,

we're good. I I will say I did find something out weird about Chad. We've been friends for how many years, Chad? Since two thousand and seven, right? Six, two thousand and six, so twenty years. Wow, are we gonna celebrate our twenty year anniversary of friendship? I feel like that's special.

Chad (03:07)
Six twenty years.

You're coming up on your twentieth anniversary of your wedding, aren't you?

Patrick (03:19)
I am. The wedding that you weren't invited to is that you're about to bring up. no, but more importantly, after twenty years of friendship, you announced to me this morning that you don't eat breakfast.

Chad (03:21)
Ha ha ha.

You didn't know that already? We worked in the same office for like seven years.

Patrick (03:33)
I knew that you did some like intermittent fasting,

but I felt like you ate breakfast.

Chad (03:38)
I almost never eat lunch either.

Patrick (03:41)
Man, y'all live a skinny life out there, people, if you're like Chad. That's that's wild. Well, we have an audience today. We've got a couple of dogs in the background, so if you hear some, you know, clicking and things like that, yeah, some barking.

Chad (03:51)
If you hear some barking, I have an we have a new

dog. She has three legs. We call her Peggy, because she needs a peg leg. But my old dog, my old island dog is not the biggest fan. And so there is a non trivial probability that he will just start barking at her at some point. So I apologize for that in advance.

Patrick (03:59)
Yes.

Nice. and what rescue did we get these dogs from?

Chad (04:17)
the three legged dog was a foster. She got hit by a car and then take someone found her, took her to the vet, they had to amputate, and then they they placed her in a foster home. So we found her on Facebook somewhere.

Patrick (04:22)
Okay.

Okay. Yeah. Y'all are great people for doing that. I it's incredible. Y'all have four dogs again now? Four dogs again, yeah. Okay. And you have a hard rule in your family that four dogs is the limit?

Chad (04:38)
We have four dogs again. We lost one a few months ago. Yeah.

Well, five dogs has been the limit.

Patrick (04:47)
Okay.

Chad (04:49)
I will not go over five dogs. I'm happier with four.

Patrick (04:53)
The compromises we make in marriage. Okay, cool. Well, let's get let's get onto our first couple topics. today's episode is kind of all about AI. We've got a lot going on in that world. I I wanted to kind of dig into a couple of things. So I don't know, Chad, if you want to start talking about our use of AI and kind of like the good use of AI from from a business perspective and like how that's changing our world a little bit, or if you want to go into some of the

like city related AI conversation first, kind of give me a direction. Where would you like to go first?

Chad (05:27)
Well, we both sent each other some articles about AI related topics. if you want to talk about kinda how it's transformed our work, we can we can go there too. I hadn't really done any prep for that, but I'm happy to ex speak extemporaneously. you tell me.

Patrick (05:48)
Yeah, I think I get a lot of conversations from our clients asking questions about how we're using AI. You know, obviously we have a lot of functions inside of ZacTax that are very similar to cities, and that we're doing a lot of city work, especially like in our finance functions or economic development functions. you know, some some of those areas. then you have kind of your world of AI use, which is, you know, I would say significantly transformative.

in the way that technology businesses operate from that standpoint. So I'd I'd really like to hit on that. I'd really like to hit on what do you see? You know, let's start there. What do you see in your side of the world? How much more efficient has AI made you? How has it made you more efficient? and what does that look like from an overall world of I mean, it's changed how we hire.

Chad (06:21)
For sure.

Patrick (06:47)
Right, specifically. We've got a new employee coming on here in a couple of weeks. but just kind of talk that through a little bit. Give your indication and then I'll talk a little bit about how we use it behind the scenes in some of our like quote unquote city functions.

Chad (06:59)
Sure. So obviously one of the most this may not be obvious, I guess. One of the more powerful places where AI has transformed things is obviously in software development. there's so much open source code that has been ingested into these language models, right? And there's a lot of just computer related tasks.

That that they're quite good at when they're running on your computer, reading files, finding files, editing files, things like that. So software development was a natural place for it to really become more ubiquitous. And there have been discussions about, you know, within the next six months, no one's gonna be writing code anymore. And all that stuff's a little bit overblown, a little bit marketing hype.

But there are a lot of tools now for people who do what I spend most of my time doing, which is which is writing code. you know, you've got Claude Code, which is obviously Anthropics, you've got Codex, which is open AIs. And these are software applications that run on your computer. They have sort of cloud friends that can talk to each other, which is which is helpful at times. but for me, I mean, I would say AI is.

Writing a large chunk of my code now, especially the boilerplate stuff, the the what we call refactoring when we you know, we take an existing code base and modify it to make it, you know, more efficient or restructure it, things like that. A lot of that stuff is sort of mechanical and the AI is really helpful for that. but the biggest thing for me is

Getting out of the like web-based chatbots where everything is just sort of in my browser and and getting it onto my computer where I have a file system. I can store notes as I'm working. you know, as you're using these these AI systems, you're still operating in like a chat scenario. So you're talking and then it responds, and then you say, okay, go do this, and then it will call some tools that it has access to to, you know, to write to files and things like that.

having access to the file system means that I can save decision points and intermediate, you know, context information that if I need to start fresh or work with a different agent s session, I can say, Hey, go read this file. This is what we're working on. And these are the decisions that we've made. So that all of that context can can stay fresh, but within like the the good part of the context window.

If you don't know what that means, we have a bunch of blog posts on our website, zactax.com, where we've kind of been trying to do an AI 101 for cities. so there's a good article about context windows, and I'll link to that in the show notes. one of the cool things that you'll see about our use of AI is in those explainers, having Claude or Codex like build visual interactive widgets, right? That kind of so you can read the article, but it's also

more for visual people, it's it's it's a pretty cool little little tool there. so a few weeks ago we had a little lunch and learn internally where we just talked about how we're how everyone's using you know these AI systems and what can we learn from each other. And out of that conversation, I learned that most of us are using just the web chat bots. So

So we've built some tools around those things to provide some of the benefits that I get using these systems on my computer. Right.

Patrick (10:52)
Right. I think I think most

of us have gotten into cowork now on the desktop from that conversation. But

Chad (10:58)
Yeah, so Cloud Cowork is great because it is a desktop application, but it has it can do a lot more stuff on your computer. So you can it's still in like the the web-based style chat interface, but you know, you can mount it into a folder and then it will kind of work in there and it can do the same thing where it kind of saves notes as you go and it can reference those things. It can also schedule tasks, which is kind of nice for y'all. but

The thing that I notice is if you already know what you're doing and what you want to do, it can make you a lot faster. But you have to really work for it. Like you have to there is a bit of a learning curve. There's a bit of a learning curve where you have to figure out how to nudge it in the right direction. You can't just like tell it, hey, go make me a budget document and don't make any mistakes, right? And then expect it to come back and and do it on its own.

because it's it's still a prediction tool, right? It's still making the next prediction based on what has come before it, based on a probability sample. And so it's gonna revert to the mean if you don't keep it on track. And so breaking down tasks into really small units and then keeping those context windows clean is is is really, really important if you want to get good results. But like it's been really fun for our team too, because

If they have obviously like I'm the only main programmer right now. We have another one coming on board. And we have Gavin who does some programming, but it's a lot more data science oriented, right? He's not like an app developer. so there are a lot of tools that you know maybe someone would create a spreadsheet for that can mostly do the job, but it'd be really nice if there was more interactivity or you know, if they could do other things. And so now they can just kind of talk to these AI systems and build.

Prototype applications, right? Like just little one off prototype applications for certain things. And then see how those work. And if they they work well enough for them, then they can pass them to me and I can convert that into something that we can actually maintain long term. So the from a an innovation standpoint, honestly, I'm gonna tell you this, Patrick. We probably shouldn't even be talking about this, but from an innovation standpoint, it's helped sort of democratize.

on our team the ability to to be creative and and try to build new tools for yourself.

Patrick (13:35)
For sure. It's changed the way that we operate as a business. Cause I mean, it it used to kind of be, you know, the the idea sessions were very much kind of you and me, right? and lots of things kind of thrown and talked about, maybe sketched out at times. but generally it usually always had to go back to you to get like a functional prototype, right? Not really an MVP. Like let's let's kind of

Chad (13:58)
Mm-hmm.

a

proof of concept. Like just especially for for internal tools, it's almost always like what's the bare minimum we can do just to get something working. And then if it works, then we can kind of keep it going, but try to keep it as minimal.

Patrick (14:03)
Yeah, kind of just a proof of concept, right?

Because it's not it's you

know, these things are not client facing. And I'll, you know, just for the clients out there, like the things that we use internally sometimes are significantly uglier than they are for the client facing side. So but for sure, you know, things have changed quite a bit. The other side of that too is that a lot of the things that not necessarily from a code standpoint, but a lot of things that we would do to come up with ideas, those ideas can be visualized significantly quicker than they used to be.

Chad (14:22)
Yes.

Patrick (14:44)
Right. we're building an application. We talked about bringing a new you know programmer on. And so we're building a a new application within Zac Tax or within the Zac family that is outside of the, you know, I call it the legacy Zac system, but it's outside of our existing software system. And the community that we're building that with, you know, it used to be that we would basically have to like describe things in words, right?

And now we can really visualize what things will look like or how things will function. There are like small functionalities that can be built in there so you can show these types of things that that a a new program, you know, a new software will do. So it it kind of changes the way you sell as well on on our side of the the equation because we can easily create visuals for that from a from a business perspective, like going into how we use it from you know the op side.

You, my my side of ZacTax is more of the operation side. Zach's side or or Zac, I just called you Zac for the first time ever. Wow. That's like an internal joke. Chad's side of Zac is the tech and nerdery, right? Like I call it the Department of Nerdery. but you know, on the op side, it's really changed the way that we do things. Now, granted, with a lot of help from your side of the business, don't get me wrong.

Chad (15:47)
Wow.

Patrick (16:09)
Doug has become extremely proficient in the use of AI. but just talk about, you know, our internal, what we call the Zac voice, right? So our internal LLM that that we use for our staffing so that they can speak in the voice, taking all the things we've ever written, taking all the conversations we've ever had, putting it into a single LLM so that.

We know exactly how to speak for the company. Right. That's a that's a tool that really honestly can be custom built for a lot of different communities. And I know there are companies out there that sell some real crap. I'm saying this out loud because I I run across this all the time with our clients who talk to me about it. So it's just not that very it's just not very difficult to build on your own. Right. but you can really control messaging and other things, you know, by

By using these types of tools as well. on Doug's side of the business, you know, there's a lot of do the same function all the time, right? In the finance side. you know, you have to have extremely knowledgeable people in the finance side. But when you're creating a a budget calendar for 30, 40, 50 cities.

And they all have the same statutory code of when things have to be done and when tax rates have to be adopted. And you have to give it a couple of variables of like when that city council meets, so forth and so on. It's pretty easy for AI to go build those budget calendars for you. Right. there's like little things that can be built that just greatly increase the efficiency. I I think one of my favorite things is having a conversation with you. I I probably had a conversation with you like six months ago where we talked about

Okay, now that you're really using AI and coding agents, what does that mean from an efficiency standpoint? I think this is one of the shocking numbers, Chad. Like something that used to take you a month to code now takes you how long?

Chad (18:17)
I mean, depending on what it is, it could take an hour. Maybe it takes two days. and that's really just because no, it's just b that's because once you go through the process of figuring out how it should work, that used to be a really iterative process where like, I mean, you may spend a whole day just sort of setting up the framework for what you're trying to do. And then then you can start tinkering with it. but it's a very slow process.

Patrick (18:20)
Right. But either way it's not taking you a week.

Chad (18:48)
where you have to translate from your brain to your fingers typing. and so now when you don't have to translate to the typing part, because it's just, you know, generating a bunch of tokens per second just writing the actual code, then you can you can iterate a lot faster and and try try new ideas. but it means that my brain actually has to work a lot harder now because you can't just rely on the output. You have to

Spend a lot more mental energy making sure that it's done correctly, not just that it's done. for me, I think one of the biggest things that people need to kind of wrap their mind around is these language models are not deterministic. So for example, on the budget calendars, like you can give it the statutory guidelines and things like this. but it's still generating probabilities.

Right. So like not only if you're gonna use the the probabilistic side of these AI systems to do things that have a very specific deterministic outcome, you have to spend a lot of time making sure, like double checking everything, triple checking everything, building in validation steps to make sure that what you're getting out is what you expect. but for me, I think the the the best way to for people to approach it is not using these probabilistic systems.

And hoping that they're gonna create a product that is correct every time. It's using the probabilistic systems to build tools that will always give you the same output given the same input. Right. So writing scripts, writing spreadsheets that you can put the data in and then get the output that you expect every single time. And you know that's gonna be right because it's that is a deterministic tool. Right. So

using the AI to make tools that are deterministic, I think, is the real the real trick, rather than relying on the probabilistic tools and hoping that they're gonna get the right outcome and then having to spend all the time double checking it.

Patrick (20:52)
Well, it also it also requires and kind of the the next thing to chat about here, it also requires you to hire people that know what it's supposed to look like, right? So specifically in in your world, right, in the programming world, I think we always thought our next programmer would be kind of a junior programmer, right? Like somebody you would train up and somebody that would work within our system and somebody that, you know, really would

at at some point, kinda kind of, I mean, you know, for lack of better term, but kind of what we've done with Gavin, who's a rock star of an employee. but kind of get somebody on the, you know, right out of school standpoint, teach them the game, you know, send them a Jane Jacobs book, you know, all that jazz. But and the reality is is AI has changed that tremendously. And I think it's changed it in in a lot of different aspects of our business. Whereas the other side of our business has always been very different. The other side of our business

Kind of the client facing operations side, whether that's CFO, strategic services, you know, that that side of our business has always been we have to go hire the brightest, the best, the most seasoned, because the the stuff that we do over there is so complicated. We need somebody who's done it 20 times in their career, right? and so we go look for people that make those things look easy so that we can solve client issues. AI changed that.

Chad (22:03)
The best people we can find, yeah.

Patrick (22:22)
On your side, I I would say quite a bit, right? Because, you know, if you go hire a junior programmer, they still have to learn all the ropes. And if you go hire somebody who's like a full stack developer, they know how it's supposed to look at the end of the day. They know what the code is actually doing and functioning as. They can be a really good check and balance on that. Or they can at least use another AI agent to be a check and balance on the first AI agent, right? Or two or three, or whatever that may be. But I I think there's a there's a shift in the employment world.

that's out there. And I think this is really interesting for city governments as well, because there is going to be just an enormous demand for people who have ten, fifteen years of experience at something. Right. And there's gonna be I don't even know what you would call it, you know, light demand for entry level employees. Right. I I think that's the I think that's the hard thing. you know, we take on

We take on interns as a company every single summer. every year we typically have an intern that's kind of in the office, you know, it's quote unquote we're remote, but that is working for us in some form or fashion pretty much all year long. we've had really good luck with our interns. We've got great interns that are working in fantastic communities. Our goal is to train them up because in our world, in the finance world, you know, with with a city when you're coming up, you're gonna go do one budget process.

In our world, you're going to do forty two, right? Through a summer budget process. So if we can teach you how to go through forty two budget processes all at once very quickly, we can then export you to the city of Fort Worth or the city of Plano or Flower Mound or wherever that may be. And you're going to be really helpful day one on job. There's not a lot of environments for that though at the entry level standpoint. And so talk to me a little bit about like where you see that going in in your world as well.

Chad (24:17)
Yeah. So for I mean, for me, the AI, the coding agents are basically entry level interns where you have to check the work constantly. And so if you're just trying to vibe code something, you're you maybe make something that works initially, but it's not something that can be maintained. And there's definitely going to be either security issues or inefficiencies or things like that.

Design choices that are going to be really difficult to maintain long term. So that's why I think we reimagined what that this new developer role would be. Is it it probably does need to be someone who can has taken a project from start to finish. You know, they have maintained a project for years, and so they know what it takes.

Rather than someone who's brand new and then just giving them Claude code and saying, Go for it. Like that's a very dangerous option for us. So, so yeah. So we started looking at someone who or looking for someone who had 15, 20 years of experience, and who's been through through those scenarios. a lot more like what we look for in the finance and strategic services and planning world. so yeah, the the challenge is gonna be.

In the software world, there's a lot of discussion about is this the end of software engineering? Are programmers not going to be needed anymore? And the truth is, good programmers will be needed more than ever. There's always room for more software. There's something called Jevon's Paradox, which is similar to the concept of induced demand where you know, w like the the demand for for tools is not going to get lower just because you can make more of them. It's going to get even

Greater. so

the challenge is going to be in that entry level market where if you can't get those jobs initially where you can gain that experience, then it's going to be more difficult to hire, say, in five or 10 years, because you're not going to have anyone or not going to have as big of a pool of people who have done the work before if you don't hire them now, right? Because you're using A AI to outsource some of that.

Entry level work.

Patrick (26:42)
Well, I want to talk about that just before we jump into the next topic. I want to talk about that just generally from like an internship standpoint, because you brought this up as we went into our summer internship side of things. You you specifically told our team that when we bring in interns, we need to make sure that they're doing quality work to build their basically aptitude, not necessarily just knocking out busy work, right? Which, you know, yeah.

Chad (27:09)
Right.

Patrick (27:11)
Correct. So

you brought up to our team that we needed to make sure that our interns were specifically on projects where they could grow their aptitude and they could grow their ability to take things from start to finish. I was extremely blessed in my first internship when I was in my under Grad at AM to work for Linda Huff and Kevin Russell and and Hugh Walker over in Brian. And they gave me project management. In fact, kind of took me from an internship into a project manager to where it was.

Hey, here's a big project. Go be one of the project managers, work side by side with some of our engineers, and go take these projects from you know start to finish. Now, I wasn't long enough there to get all the way to finish on a couple of my projects, but it did teach me kind of the project management side and like what things are supposed to be done and when they're supposed to be done. They had really great systems in place. and so, you know, when when cities are bringing interns in.

When grad schools are wrapping interns up, project based learning, like capstone style project based learning, is going to become extremely important for folks in these programs and those internships, in my opinion. Right. Taking a project from start to finish with a senior level employee who's overseeing that start to finish, right. Like right now we have a planning intern. And so that planning intern is going to be taking things from kind of start to finish on some of the projects that we're doing.

and some of the work that we're doing, but she's overseen by, you know, a one of our senior level employees who is going to teach along the way. Same thing in the finance side. We've got a an accounting student, finance and accounting student that's there is being overseen by one of our senior CFOs. So I say this all to say that if we're going to prepare people for that next stage in city government, it is not going to be a slow burn like it used to be. It's not going to be, hey, I need you to come in and scan a bunch of engineering plans for the next five

you know, weeks of your life as an internship. it's now gonna be it's important for them to be involved in a day-to-day project management because there's not a lot of room in organizations for people at some point to be at that entry level. Like we have to we have to advance them quicker is basically what I'm saying. Which means you have to well round them in the opportunities that you have where they're not getting paid a ton of money or they're still in school. so

Just a little different mentality that I think you brought up to our team that we had to have with a lot of these interns. and so I think that's important to see is like where's that gonna go. so anyways, my my two cents on that. I do want to move into how AI has changed from a local government standpoint, specifically. Yeah. So I mean, there's, you know, there's a lot going on out there, but I do want to hit on.

Chad (30:02)
The world.

Patrick (30:07)
you know, just in general what's going on kind of at the city hall level when it comes to AI. You know, obviously all of these tools take a lot of power. A lot of these tools take a lot of compute. and so we have to build, you know, data centers for these things to happen. we're seeing a lot of backlash that's occurring at the local government level. It reminds me a lot of like two thousand and eight, two thousand nine Tea Party discussions.

There's not necessarily a lot of logic behind the discussion. There's just a lot a lot of fear, right? and it's really easy to attach to these conversations. and then you have some bad actors out there that give these people ammunition. Like I like to, you know, talk about or tell people a lot. People with dementia usually have a source of truth that eventually allows them to to get a little crazy.

Right. Like something happens that they dive into, but there's always a little bit of truth to what they say, which kind of allows themselves to capture onto it. And and that's what we see right now in this AI argument is especially when you're talking about sound, light, power, and water, right? Those are kind of your big four things. You do see Bitcoin mines out there that were using a tremendous amount of water. Right. You you do see

Older data centers out there that were using a tremendous amount of water. You see a retail electric grid in Texas that is not really built to sustain the demand level that these data centers are bringing online, which in turn, you know, if not done correctly, can increase the cost of electricity back to a residential house or ratepayer because of the way the demand works within the system in Texas.

It's also what makes Texas kind of sexy for data centers because of the way that our electric system is set up and the deregulated nature of our electric system. it's also why West Texas has become extremely popular because it's so expensive to transport natural gas from the West Texas oil and gas fields. It's a whole lot cheaper to transport that to a data center who's gonna burn it and use it for electricity than it is to transport it somewhere else. So I I think, you know, there's a lot to be said there.

But one of the things I do have to say is I've seen a lot of cities who have just chosen to either not annex a property that's going to be used for a data center, kind of ignore it, release it from an ETJ. We've seen that a couple of times throughout the state.

And I just have to say it

You can't regulate what you don't have inside your city or your ETJ. Right? So there's just a lot going on in a data world right now. And I feel like we probably should take a little bit of a breath and understand that most of these things, most of these problems, most of these bad actors.

type of things, it's really no different than anything else we've dealt with in zoning, like storage facilities, sexually oriented businesses, vape shops, everything else, which I know you love zoning to begin with. So this is payday loans, yeah. It's just we haven't stopped to say, okay, what what do these things need to be doing? Right? If you have it inside your city limits and they're a good developer and and you're talking about

Chad (33:41)
Payday loans and things like that.

Patrick (33:58)
the the value of these things are in the billions, right? I know that, you know, we're working on deals that a city has five billion dollars of total property value on the ground and they're doing a $25 billion data facility. Right. Like these are these are highly complicated, highly complex deals. And I think in cities, some of them have just said, we don't want to deal with it. We're just gonna leave it out there in a county. The problem with that is you can't regulate a lot of things and you can't protect people that are in the county.

and because of annexation reform, cities have kind of gotten to the mentality that it's not their job to protect people out there, right? The state has basically said we don't want you to do that. I think we're probably gonna see a little bit of change in the legislature because of that. from a standpoint of what they're gonna do. I think counties are probably gonna see some type of regulatory guidelines where they can regulate sound and light. I think that's probably coming. but the reality of the situation is

We have to, as city halls, we have to dig into these issues and figure out, okay, how do we how do we do things that benefit our community? How do we protect our fiduciary responsibility to our community? And then how do we also protect our community from the harms that can be caused if they're not done correctly? I wanted to make sure we talk about that a little bit. because data centers are not going to go away. The first part of our conversation, talking about how we use it, what the efficiencies are, how you use it.

I think really how you use it is the one that I wanted to really get across because that's what's driving a bulk of the data center development side is important to understand because the demand is there. These things are going to occur over time. Like it's it's coming. We have to kind of step back, take a breath, and figure out okay, how do we do this responsibly? and there's a big difference between having a conversation about how do it responsibly and just saying,

I want to put my head underneath a rock and act like it's not happening. And we also have that occurring because the politics have been so rough. Now there's been some brave cities out there, incredibly brave cities out there. And some of them have had recall elections filed against some of their council members because they've been brave in the process. So the last thing I'll say on this too is

Chapter 312 of tax abatement agreements that are done on these, I think a lot of people look like their handouts to the AI industry. Why most governmental entities do chapter 312 agreements on these and what they replace that with is basically a payment and limit tax agreement, is because if you don't do a chapter 312 abatement, you are putting your tax rate at extreme risk. and that is because the way business personal property works and depreciation of business personal property works.

So I I say this out loud, and this is a lot more complicated than a podcast can get through. But if you're depreciating all of the assets inside of a data center, which ninety to ninety-five percent of the value of that data center is the personal property, if you depreciate that over four to six years and then you replace it in year seven, and the appraisal district considers that replacement to be old value, which a lot of appraisal districts do, you are going to compress your tax rate and lose the revenue associated with that.

It's extremely important that if you're going through a process or you have data centers going through the process, that you have some professionals in the room that can really work through the taxation piece and can really work through understanding that piece. because if not, you could put your city in a really precarious situation in the next six or seven years. And that goes for cities, emergency services districts, counties, any taxing district with a property tax is going to be impacted by the way the business personal property tax hits the tax rolls.

So there's my soapbox for the minute. You want to talk about some funnies? Yeah, that's a great monologue. You wanna talk about you wanna talk about some funnies? I'll let you roll into the the difference of or I I would say where where AI and automation has gone wrong. This is this is a good one. Yes.

Chad (37:49)
No. It's a great monologue. Sure.

Has gone wrong. Okay, well,

I haven't given you my take on this. I suspect that you are not gonna be too surprised. But okay, so this is an article from the Atlanta area. It's a couple of weeks old. We didn't get the podcast when it when it was new. residents of Buckhand are raising concerns about Waymo self driving vehicles using their neighborhoods as holding locations. So basically a Waymo, which is a self driving car.

Patrick (38:08)
go for it.

Chad (38:32)
Doesn't have a ride. So instead of parking somewhere, they're driving just around city streets. And in some cases, as many as 13 cars are driving through a residential street. this one person said yesterday they had 50 cars that came through within a couple of hour period. and so these neighbors are just apoplectic about it.

So I I'm curious what what your take is first before I give you mine.

Patrick (39:07)
It's a public street.

I mean I I don't are they blocking people's access to the public street? Are they doing anything illegal?

Chad (39:19)
the the residents have tried to put like barricades out to keep the Waymo's from going through.

Patrick (39:28)
Like like cones, like they gotta they gotta weave through the cone system. Wow. I I think the hard thing for me is before Waymo was doing that in that area, how many Uber drivers were driving through?

Chad (39:30)
Mm-hmm.

Patrick (39:43)
Right. Waymo is probably doing it because their mapping and navigation systems are telling them that's the best place to drive through. Right. So how is that any different than Uber's navigation system for a normal driver or a Waze navigation system doing that as well? I distinctly remember a project as a city manager where we had an opportunity to connect a street and a residential subdivision to another street where it was in our CAP plan

For 20 years, and we finally had the money to go do the project. And we started thinking about because it was right off of Interstate 20. We started thinking about if there's an accident on Interstate 20, which happens once a day, right? If there's an accident there, what's gonna happen in the new world of automated mapping and GPS systems? And that kept us from connecting that road, right? Because we would have we would have taken a bunch of interstate traffic.

And we would have put it into residential subdivisions to be able to connect through to get down to another thoroughfare through residential subdivisions. So

Chad (40:50)
It would be a real shame if we could use our public streets for conveyance.

Patrick (40:55)
I I don't disagree, but then there's the political reality of it, right? I it's a public street. so I I mean I don't I don't feel bad for the residents that are on that public street, and I also don't know what the difference between it's you can see a Waymo, right? Because it has the big circle on the roof and you know it's a Waymo. You don't necessarily see the Ubers. I would have been curious to see what is the overall traffic count change that's occurred on that street. That would be more interesting to.

Chad (40:57)
be a real shame if we

So

Going to your example of right, like we don't want to put traffic through this residential neighborhood. to me, I think that is the the bigger thing to talk about or or worry about, right? We designed all of these neighborhoods almost with the idea that these roads are only to be used by people who either live in these areas or are

Going to these areas as a destination. if it's the Amazon person dropping off a package, if it's the trash truck or something like that. but otherwise, you don't belong here. and that mentality is really hard once it takes hold. It's hard to change it. It makes anyone who doesn't actually live in your neighborhood makes them an outsider, breeds a lot of

paranoia almost, where like there's no reason for you to be here unless you actually live here or are coming here for a very specific reason. You should not be here. You're an outsider. That's not really healthy from a community standpoint. Like it's a public road. It should be usable by anyone, and not just as a parking lot, you know, for your for your traffic. You already have two parking spaces.

on your driveway that we require to in your garage that we require. And yeah, we still use these streets not as conveyance methods, but as parking lots. and there's a cost to that. And so

I I don't know. I I just feel like it's it's a symptom of a bigger issue where we almost treat our residential streets as if they're private driveways and not public transportation facilities. And the the the question that I have is like, if that's the mentality that we want to take, then we probably need to think of a different financing mechanism for maintaining and building these roads. Because if they're if we're turning them into de facto private roads,

private streets or private driveways. why are they public roads? W where where is the value for anyone else that's paying for these roads coming from? Or going to, I guess is the better way to say it. But so I think it's both a financing question and also just a

Patrick (43:44)
Are they public roads? Why are they not behind gates? Yeah.

Chad (43:59)
question of like w what type of communities are we building where we have all these this this public infrastructure, but if anyone actually uses it, then we're gonna, you know, lose our minds over it.

Patrick (44:12)
Well, I also say there's there's a taxation conversation there to have about Waymo as well, right? These are automated vehicles that are running on the roads significantly more than a taxpayer would be. So should there be some type of taxation system separate for those automated vehicles than there are for somebody who's driving a typical vehicle? I you know.

Chad (44:31)
We don't wanna have that

conversation though, because that would require either a mileage tax or some kind of change to our registrations, which obviously that's not even gonna help local governments either, because that's a state fee.

Patrick (44:39)
Mm.

Well they've done

they've done that on electric vehicles in Texas now. So, yeah.

Chad (44:49)
Right. And it's

not gone over well politically. Like, why are you charging me more for this this vehicle? But if you're not paying a gas tax, which also doesn't come back to to localities, those electric vehicles do weigh more and they do put more stress on the actual roads. but we don't want to have a mileage tax. We don't want to have I mean, if we really wanted to get efficient about it.

Patrick (44:54)
It it is it is not. Yeah.

Chad (45:18)
Since every vehicle has GPS, you could just track what what roads everyone drives on and then apportion the gas tax or some other kind of mileage tax to every jurisdiction. I mean, you I'm not saying we should do that, but you could do that. but no one, no one's gonna allow us to like take our mileage at the beginning and end of the year and then pay some kind of fee based on how many miles we drove. the gas tax was supposed to be like a de facto version of that.

Patrick (45:26)
Wow, Chad. Wow.

Chad (45:47)
But not only has cars got more fuel efficient, the gas tax was never indexed with inflation. We constantly are temporarily cutting it because the street of Worm Moose gets closed or other shocks happen to the system. and now we're you know, on electric vehicles which don't even use gas. So like at some point, we gotta figure out how to pay for roads.

And I don't know why we're building a bunch of roads that can't be used by anyone. Like there's just a fundamental mismatch in how we view these streets. And that's largely because of the the way that we build them. Like it's just understood. Like the reason that people buy ho their houses in these neighborhoods is because there's not going to be traffic. And I get that. Like I understand. Do you want to live on a quiet street?

Patrick (46:12)
I

Chad (46:40)
Well, first of all, these are electric vehicles, so they are very quiet, but that's kind of tongue in cheek. I get that you want to live on a quiet street, but it is still a public road. And like to lose your mind because someone's using a public road, it just seems odd to me.

Patrick (46:58)
All right, last but not least for today's conversation.

Okay, so this actually annoys me personally. So when you order from Pizza Hut or Domino's or any of the major national chains now, the employees that are delivering to you are no longer Pizza Hut employees, right? They f say what?

Chad (47:14)
I did not know this.

I did not know this. Mm-hmm.

Patrick (47:19)
Really?

They funnel it to DoorDash. I think the only reason I know it I I know you Doordash. Are you Uber Eats or DoorDash?

Chad (47:27)
I prefer Uber Eats when I when I order, but I think my wife prefers DoorDash because she has like the Dash Pass subscription or whatever and it's cheaper.

Patrick (47:29)
Okay. Okay.

Right.

When you order that, so like your your delivery order gets put into whatever so if I order through the Pizza Hut app, which I don't need a lot of pizza anymore, but if I order through the Pizza Hut app, it basically goes through Pizza Hut, pay it, and then the delivery is done by a DoorDash driver and the tip goes to the DoorDash driver, right? basically the franchisees at Pizza Hut are claiming that these gig gig workers

you know, leveraged its AI system for their own benefit, you know, causing it an operational breakdown issue and it it pummeled the sales. And and I'll tell you my personal experience on this is they would they would go and pick up Pizza Hut orders that were totally uncontrolled by the DoorDash app, right? So you couldn't track timing, you couldn't track anything. And then they would go and pick up other orders along the way. Right. So like they would go to Chipotle,

And get a DoorDash order and they would go to Panda Express and get it. and then like your pizza would take like an hour and 15 minutes to get to you instead of 15 minutes. Right? Because they were automatically getting paid. They knew what the tip was gonna be, and they were kind of leveraging the system with their. That was my personal experience. but it was all automated and there was zero control from the franchisee of what they could or couldn't do. And basically franchisees were told, this is the national arrangement that we have, and this is what you have to live with. And then sales

Or in their opinion, sales plummeted because of that. I will give my personal experience. I chose not to order another Pizza Hut pizza for delivery because my pizza showed up cold, I think, twice. Right. And my local Doubledaves delivers from an actual Doubledaves location with an actual employee. And plus, you know, Double Daves is a gem of the Aggie world, so I have to order Double Dave's. But

So

People can game the automation system that's out there. Right. And so I think not all automation is great automation. Not all AI is great AI. and I'm trying to remember there was another one that was inventory management as well. St was it Starbucks? Yeah.

Chad (49:56)
Starbucks. yeah.

I I think there's a broader question here too, which is not just AI, but the vast quantities of data that are now available and like the hyper personalization of everything. So, for example, there were a bunch of articles about how if you open Uber and your phone is like less than, you know, 10 or 15%, they're gonna surge price you.

Because they think that you're more desperate, right? So like you used to be able to do some of this. it was just much more difficult. Like the there was more friction. So for example, the first car I ever bought, I was really, really stupid. And I went to the dealership like four times to look at this car before I pulled the trigger. And every single time the salesperson saw me, right? So he knew that I was quite interested in this car. And it was gonna be harder for me to negotiate, right?

but that took a lot of friction because there was a fit like a personal interaction between me and the salesperson. When when everything that you buy at, you know, at a Walmart or whatever is like tied to your app and then they know your spending habits and and your needs and things like that. When when Uber can price based on how desperate they think that you are to get a cab.

It seems like there's kind of a social contract that has the potential to break down because there's not, you know, the same experience for every single person. There's not the same price for every single person. And this is just a brave new world that we're kind of getting into. And I think, and it's not just that too, it's this this question of friction. I think is is really important because as you add more automation to everything,

Like it's just easier now to just send an email or like, for example, cold emails. How many cold emails do you get on a daily basis for all kinds of things? And when you had to like write that, or at least when you had to go to your browser and copy paste something into the email form and then hit send, like that's just a level of friction. when we when we posted this software job, I got

Patrick (52:04)
Crazy.

Chad (52:22)
about 20 applications within five minutes of the job being posted. Right. These people are not going to get interviews. Because they didn't read the job description. They didn't write me a cover letter explaining why they're good. They just had probably an alert that showed up on some website and said, this is a job. I'm going to I'm going to apply for it.

That's a level of friction that's now gone, just in that one realm. That is great for them, I guess, because now they can just spam a bunch of job applications. But for the person that's hiring, it's like now I have to go through hundreds and hundreds of just completely garbage applications and try to figure this out. so it'll be interesting to see how kind of how that plays out on the other end, right? So, like

We're gonna have agents that are just sending out job applications and then we have agents that are just reviewing all these things and like

At some point, I think we're gonna lose the plot because that the friction that made things a little bit more difficult, in many cases is actually pretty helpful. Right. The like having your own delivery people where you can control the experience of that entire round trip from ordering, like you used to have to call the Pizza Hut and tell them what you wanted, right? Now everything is just

Mobile, you can schedule it for later. But now now like Pizza doesn't even have control over the delivery. and so you're just relying on someone else.

My guess is that at some point we're gonna like hit that end of the pendulum swing and we'll gonna realize like we we probably need to have a little bit more like human interaction, touch points, a little bit more control over how the end user is experiencing whatever we're offering, whatever service that is. and particularly I think for cities, this is a a concern. We you don't wanna have every interaction with a resident or a business owner like just completely automated.

Patrick (54:02)
Right.

Chad (54:26)
w working through some kind of funnel where where you lose control of of those touch points.

Patrick (54:34)
I mean I I think that is a huge issue for cities, right? 'Cause I think there's a there's a big push to automate a website, there's a big push to automate phone systems, there's a big push you know, to

Chad (54:46)
Speaking of

which, Patrick did that and it is the absolute worst.

Patrick (54:50)
Yeah.

Chad (54:53)
But the funny part is if you call him, I don't know if I should tell this to everyone, but if you call his little AI, you know, phone service now, he does get emails summarizing what the call was. And so from time to time, I will call it and just talk about random things or get it to like change the change the language to French or something or Portuguese. The the languages that you chose are really strange because I think it's like French, Portuguese, and Russian in addition to English. Is there Spanish? Okay, maybe that's new.

Patrick (54:53)
So

Yes.

They're Spanish. Th they're Spanish, yeah.

Yeah.

Chad (55:24)
yeah, so I'll get it to like change the language and then we'll have a conversation in in French, which is fun because I can say about three words in French. but then he'll get a nice email summarizing.

Patrick (55:31)
Nice. Nice.

Right. you know, look, I think I think the world of automation is is helpful. but you are correct. The touch point is there. And I think there's gonna be a whole industry specifically like what you talked about there, which is like the personalization of that touch point. I mean you go back to your car buying side. My brother just bought a vehicle and used a a personal agent to buy the car for him, right? And and somebody who goes through and does and and deals with all of that friction.

Chad (56:01)
So like negotiated with the

Patrick (56:05)
along the way. so I think there will be a whole industry that comes out of that. I found it really interesting last time we were on a business trip that I think it was Allison. Allison had an Uber Black price that was thirty dollars less than my Uber black black price. So, you know, I think I think that's a a a tough a tough thing that like we're just dealing with, you know, this personalization of pricing and situation and and all that different types of things that

Yeah, we've seen that in airlines. I think people really realize that on the airline ticket side, but I'm not sure we've seen that in everything that we do. From like you order DoorDash and so DoorDash gives you different deals than they give to other people or whatever that may be. So Chad, we did a pretty good job on our first one back for a long time.

Chad (56:53)
Hope so. I guess we'll see when I actually go back in and and edit it, if it was worth listening to. it's a much different world today than it was last August when we, you know, pushed our last episode. So certainly a lot for us to cover here. But yeah, man. so now that we are back into the groove of things, even though it's still summer.

Patrick (56:55)
Yeah.

Ha ha ha.

Chad (57:22)
I gotta tell you this. I've told my wife that next summer we might just instead of having, you know, trying to coordinate camps for all the kids and all their activities and things things like this, where basically, you know, I may get like one or two hours a day of just solid time where I can work and the rest of it's just trying to piecemeal. Like for example, in about an hour, I'm gonna have to go pick up kids from camp and then take them to the next activity. we might may maybe we just like go to Europe for a month.

like all these people coming for the World Cup. Well just go somewhere else and at least then we can have fun and not work instead of driving all over DFW and still not getting a whole lot of work done.

Patrick (58:08)
Yeah, I mean, I've I've yeah, I was I was actually joking with with David and Pagosa that in August I was gonna go live in Colorado because it's just man, you can already feel it outside right now. It's getting so hot. yeah, my next summer's gonna be worse because I've got like we're wrapping up baseball right now with my boys. We're both on the youth side, but next summer I have a high schooler who plays baseball. And so youth baseball ends like the second

Chad (58:08)
So y next summer you may not see me at all.

Yeah.

Patrick (58:35)
week of June and then high school baseball kind of starts the first week of June and runs through August. So yeah.

Next summer's gonna be even worse than my last two summers have been, I think, when it comes to that. So it should be should be thoroughly fun and entertaining. So got a couple weeks till I take vacation. So hopefully we can get one more podcast in before I jump out of vacation. and you know, we'll try to try to get back to doing it on the norm. And of course, when college football season rolls around, we'll start talking again. And I promise everybody I'm not gonna do that. It w it had nothing to do with the podcast now. Now I know.

Chad (59:06)
Depending on how his Aggies are doing.

Patrick (59:12)
No, I know. I just can't attend football games. So I've decided I'm going to a it. I apparently am the new Ted Cruz when it comes to Aggies. we'll know because I'm attending North Texas football games this year. So we'll know at that point. I hope to make like two. Whatever works within the schedule. We'll see. So anyways, but that was fun, dude.

Chad (59:17)
Are you the n are you the new Ted Cruz? I'm gonna start calling you Teddy. Okay.

Patrick (59:39)
Really enjoyed getting back to it and and having a conversation. As always, if you got any questions for us or you've got stuff going on that you need to chat about, you could reach out to myself or Chad. most people who listen to our podcast probably have our cell phone numbers or my cell phone number. And you can always get a hold of us via email if you need to get a hold of us there. Just our first name at ZacTacks dot com. Not super hard.

Chad (1:00:00)
All right, Pat. Well, you have a great one and I'll see you next time.

Patrick (1:00:03)
All right, man. Bye.