Beau Button: The Unemployable Founder Who Paid Players $13M and Built His Own AI Rig

Episode #26 - Beau Button

===

[00:00:00]

Matt Klein: All right, folks, welcome to another episode of Beyond the Noise: Signals, Stories, and Spicy Takes, the show where we dig into the stories of the people shaping the future of app-based computing. I'm your host, Matt Klein, co-founder and CTO of bitdrift, as well as the founder of Envoy Proxy. Each episode, we'll talk with engineers, founders, and technical leaders who've transformed the way their companies build and understand what's happening inside their systems.

We'll dig into the challenges, the breakthroughs, the lessons learned, and we'll wrap it all up with their hottest takes. So let's dive in. Today, I'm thrilled to have Beau Button with us, who's been writing code since he was 10 years old and building software companies for most of his adult life. After years in enterprise software, he co-founded Atlas Reality, the company behind Atlas Earth, the largest virtual real estate game in the [00:01:00] world, which has paid out over thirteen million to its players.

He's also the co-founder of TesoroXP Incorporated, an Atlas Reality spinoff whose patented platform runs real world card spending into in-game currency. He serves as president and CTO of both. Welcome, Beau. Thank you very much for joining us.

Beau Button: No, thank you for having me. It's a pleasure.

Matt Klein: Yeah. Fantastic. Um, I have no doubt, based on the last time that we chatted, that we'll have an interesting conversation.

Um, but how I typically like to get started is, uh, what you had in your bio. You said that you've been writing software since you were 10. So tell us a bit about how you got into technology.

Beau Button: Yeah. So I grew up in the South, particularly in New Orleans. It was hot and humid, and honestly, I was always a heavyset kid, so I found myself gravitating towards things that you could do in the house.

And I was fortunate enough, uh, my neighbor's dad had access [00:02:00] to an old IBM PS/2. Uh, we had no idea what it was and how to really use it, but that was kind of, like, my first intro into personal computers, and it had Quick BASIC, and we were just dabbling. You know, just playing around with it, and I kind of, you know, got enamored with it.

And that evolved into, we had dial-up internet, AOL, GeoCities, Tripod. Like, that was where my, my brain went, and honestly, I just never stopped. Um, but, like, I think the real impetus was is, like, the alternative was go outside and play G.I. Joe with the, the other kids and sweat my ass off. Um, and, uh- ... yeah, I wasn't interested.

Matt Klein: So what, so what types of things did you program when you were a kid?

Beau Button: Honestly, at that age, I, I would argue I wasn't really writing anything useful. I was enamored with the sound command in Quick BASIC. Not that I was a musician back then, but it was one of the first, like, instructions that I had learned where you could make, I, I hesitate to call it music, you know, like sirens [00:03:00] and things.

But, uh, we... I, I, ironically, considering where I am now in the gaming space and, uh, how outspoken I am about not playing games, I, I can't stand wasting my time on games. I, I tried making games. Um, I was never really successful. It was basic, like, again, it's Quick BASIC, so a MUD, you know, a terminal-based game.

Matt Klein: Right.

Beau Button: Uh, it wasn't anything I think anybody else would have played, but it was more so just trying to figure out what I could do with the machine. When I got access to the internet, that changed. That's when, like, HTML, and back then it was VBScript and JavaScript and Netscape and Internet Explorer.

That's when I really got into trying to build... Uh, we didn't call them web apps back then. I called everything a website back then. But yeah, I mean, that's, that's really it, just hobby things really.

Matt Klein: Yeah. I mean, I think we'll get into it more when we talk about your companies, but i- if you hate games now, did you play games back then?

Or, or, like-

Beau Button: No

Matt Klein: ... if you were never a gamer, I mean, like-

Beau Button: I-

Matt Klein: ... we're, we're probably fast-forwarding, but, like, how did you start a bunch of [00:04:00] gaming companies?

Beau Button: I know. It's weird. I, I- I've often told people the reason I didn't play games is because we couldn't afford the consoles. But the reality is that same neighbor, he was an only child, had every console known to man.

The minute it came out, like his name is Patrick, and we're still extremely close friends. He, he would get the console. So I had access to them through Patrick, but I honestly, uh, maybe it's an ADH thing. I just never could find myself just interested in sitting down to play most games. Now, I can play simulators.

Like, there was this game on the N64 called Pilotwings where you would fly around a glider. I could lose myself for hours in that game, but it didn't have missions. It was just like free form open world-

Matt Klein: Right. Yeah

Beau Button: ... before open world was a term. But yeah, I just... The games that have career mode and, like, force you to, like, progress in the way that they want you to progress, my brain, that's probably why I've had to be a serial entrepreneur.

I'm unemployable. I'm untrainable. I just don't like rules. And [00:05:00] to-- It's like it's too much. It's like work

Matt Klein: All right. All right, so then let's go back in time. So you were a kid, you started dabbling with computers. Tell us a bit about where you went from there.

Beau Button: Yeah. So my first job was at the Piggly Wiggly, which if you're not from the South, you may not know what the hell that is.

It's a very Southern grocery chain, and I remember working there for a year and a half, and at that point, I think I was 15 or 16, um, the language I had kind of graduated to, and I was never really proud of, was Visual Basic, VB6 to be specific. Um, and I was, I was itching to make money another way than working at the deli counter at the local Piggly Wiggly.

And, and this was around the time that monster.com was, like, the de facto standard job bulletin board where you'd post, you know, "Hey, I'm gonna-"

Matt Klein: I remember. Yeah

Beau Button: ... you know. Oh, typical like Indeed. I posted out there, and I literally said, "I work in the deli. These are the techs that I know access to." And someone reached out, and I actually landed a job at the age of 16, uh, at a company called [00:06:00] EWeb Systems, and I was hired as a content management and e-commerce developer doing Active Server Pages with VBScript.

Matt Klein: No. No shit.

Beau Button: Uh, and that

was-

Matt Klein: So you, so you dropped... So w- did, did you drop out of high school or, or did

you not?

Beau Button: No, no, no, no. So no, thankfully I didn't. I dropped out of college, but no-

Matt Klein: Okay. All right

Beau Button: ... I do have a high school diploma.

Matt Klein: All right. Yeah.

Beau Button: Um, this company was looking for... And I'm, I'm very close friends with the hiring manager.

This is 27 years now that, you know, uh, this has occurred. But, uh, I've always said he was looking for cheap labor. This is before we could go to the Philippines or India. Like, I'm sure you could have back then, but it wasn't something that you regularly did, so it was like child labor. Um- ... and, and, uh, they paid me above minimum wage, so I was happy as a pig in shit because I wasn't behind a deli, you know-

Matt Klein: Sure

Beau Button: counter working with customers.

Matt Klein: Right.

Beau Button: But, uh, no, I didn't drop out. I, I got out of school at 2:15, and my mom was, uh, gracious enough to let me use her minivan, so I'd drive to, uh, the building on North Causeway and work for five [00:07:00] hours, and then basically go home, do my homework, and rinse and repeat.

Matt Klein: Yeah.

Okay, cool. So, so you were basically a professional programmer starting at 16?

Beau Button: That's... I, I don't say that because people are like, "That's bullshit." But the reality is, is I was the youngest person probably by about 10 years at the company.

Matt Klein: Yeah. I mean, I-

Beau Button: Um, the company's heritage was really strange as well, how it got into that business.

But yes, I was, I was

Matt Klein: I think there's probably a common thread o- of like people from that era that have had similar experiences because I, I actually started working for a software company, um, not as young as 16, but when I was 17. So it was, it was very similar. And at least back then, I don't know what it's like today, maybe with AI we'll get there, it's all different and like no one-

Beau Button: Yeah

Matt Klein: wants to hire a high school kid or whatever. But I, I remember at 17 I was making incredible money. I mean, like based on-

Beau Button: No, I remember I

showed my mom my first paycheck ...

Matt Klein: what, what I could, what I could make elsewhere. [00:08:00]

Beau Button: And she's like, "What are you doing?" 'Cause my mom is very religious and I, I used to shut my door when I was in the-- I turned the lights off.

It was like a cave and, you know, she-- I wasn't a bad kid, but I had an attitude and I had a, I have a real bad mouth on me. So you know, I, I got in trouble quite a bit with my mom and she was always convinced I was up to no good. Um, and she'd come in and I'd show her a few things. I had like five computers on my desktop.

I had various operating systems I was dabbling with. But I remember showing her my first paycheck and she's like, "This is..." I was like, "This is a W-2, Diane. This is not like under-the-table money in a duffle bag. This is a legit job." But yeah, no, I was, I was-- Obviously the money, it made it e- I loved what I was doing, but that made it even better because like I said, I grew up, I wouldn't say poor, we weren't homeless or anything, but like it was, we were very, very You know, modest in what we could afford.

Matt Klein: Yeah.

Beau Button: But, uh, yeah, no, it was, it was great.

Matt Klein: All right. So you work for this local company, and then what, [00:09:00] what from there?

Beau Button: So they, they were, uh, uh, willing to basically keep me employed as I went to Louisiana State University. I had enrolled in computer science, and, uh, when I got there, uh, I knew I, I, I probably wasn't gonna last very long because the curriculum didn't even mention the internet.

Um, we were using VAX/VMS, writing in ANSI C, and, uh, now you-- I, I know your background and, you know, I'm still starstruck because you're the guy that created Envoy, but it's like I don't operate in that space, that low-level C memory management. I'm a user space. Honestly, I'm more of a, a liaison between end users.

I just so happen to be a reasonable engineer, but I knew I wasn't going to fit in because I'd been doing web development for so long.

Matt Klein: Sure. Yeah.

Beau Button: And we were submitting our homework using FTP, and I'm just like: Holy shit, man. This is-- I'm uncomfortable. Like, this is not where I wanna be. But, uh, yeah, no, they employed me, and I, I lasted about a year with college and then that job, and then I kind of outgrew it [00:10:00] because as I learned more, I realized I was severely underpaid.

Um, I was still very happy, but as you start to see what other people were making who are 10 years older that have college degrees, I had to flex my muscle a little bit and basically, you know, go through in one year, like three forced raises. I was like: "I'm out this bitch if I don't get more money."

Matt Klein: Right.

Beau Button: Um, and it just, it made the situation very uncomfortable. But yeah, after that, that's basically when I went into consulting. I, I said, "You know what? I'm gonna try this on my own," and this was when everybody wanted websites. This is way before mobile apps and all of that nonsense. But, um, that's basically how I got into, uh, the serial entrepreneurialism, but most of that was as a con-consultant.

Matt Klein: Yeah. So I, I mean, did you- At the time, had, had you thought about starting a company, or was that not on your radar yet?

Beau Button: Oh, so-

Matt Klein: Like, were you trying to get your, your legs under your feet?

Just, just-

Beau Button: I did, yeah. So I actually at that time started a company with a close high school friend. We called it Web Genetic, [00:11:00] and we were both, uh, consultants really doing our own thing.

Um, my partner at the ti- partner, business partner, his name is William, again, long from middle school all the way up. Um, people always ask me 'cause, like... And they're like, "You, you're the person that I think knows the most about computers." And I'm like, "Well, William's the guy that I know knows the most about computers in, in my world."

And yeah, we started a business, but basically we just did whatever the hell... I mean, we've done some random-ass stuff. Uh, I remember we didn't know it at the time, but we actually wrote software for one of the-- It was a spam vendor that was doing, uh, it was in Windows. I forget the technology where these pop-ups would come up.

It was like a hole in Windows 98 where you could spam the OS, and we didn't know what we were doing. We were like, "Oh, this is pretty cool." And then we realized, oh shit, we just facilitated all of these spam messages, but it paid the bills. But yeah, it was a company, but we didn't-- I didn't really think about building a product.

Like, it was services, professional services.

Matt Klein: Sure. Yeah.

Beau Button: But quite honestly, it prepared me to build a product 'cause we touched all [00:12:00] kinds of crazy tech. N-not as low-level crazy stuff as you, but it was, you know, different technologies, different businesses, and, and we learned a lot.

Matt Klein: Yeah. So I, I mean, so you were doing the, you were doing the consulting thing, and I completely agree that there's a whole bifurcation between, you know, doing what I would call, like, services and support or consulting and those types of things, and then doing a product, right?

Beau Button: Yeah.

Matt Klein: So obviously for a while now, you've been doing product. So I guess take us through a bit h-how did you make the leap, you know-

Beau Button: Yeah, so-

Matt Klein: ... from the consulting gigs to, to doing your first product company?

Beau Button: Yeah. So the transition happened. I was a consultant at the time working for a collections agency in Metairie, Louisiana.

Um, and they had won a contract with the City of New Orleans to basically collect delinquent, uh, property taxes. They call them ad valorem, by-value taxes. And this company had their own in-house collection [00:13:00] software called DebtCon. Um, and I forget the, the language it was written in. It was some obscure-- I can't remember.

It wasn't like FoxPro or any... I, I-- It'll come to me when I'm least thinking about it. But they, they won this contract, and ultimately they were like, "We don't wanna use this old tech." You know, I had been working with them to bring some of their internal software capabilities online in a very strange way, 'cause like we were just exporting CSV files and basically creating portals for customers.

Matt Klein: Yeah.

Beau Button: But it wasn't really connected to a database. It was just like a, a snapshot. And it, it worked really well. But yeah, I, I was there at the time that they were trying to figure out how to build the software that was gonna power, you know, the services that the city of New Orleans needed. And the gentleman who was running that, who had basically submitted the RFP He, he offered, you know, basically to, to bring me in as a partner in a new business venture for that particular contract.

And I said, "Yes, I will gladly." I mean, it was a sweetheart of a deal. I [00:14:00] was making good money as a consultant, but this was, like, guaranteed. You know, all right, now I have equity. We have all of this stuff. And, uh, ironically-- not really ironically, I guess coincidentally, the William Salsam from my first business, he also came on board.

So the three of us, uh, Brian, William, and myself created a company called Arcon Information Systems, and I ran that for nearly 10 years.

Matt Klein: What did that do? It did.

It-

Beau Button: Well, so we started off writing a software called TaxCon instead of DebtCon, and it was, uh, basically a collection software package, uh, tailored to the state of Louisiana for collecting delinquent real estate taxes.

Um, and we actually-- And this is the part of that business that always gave me gas. We lobbied, uh, we had a lot of lobbyists. We changed the constitution in the state of Louisiana because after you exhaust all of your options to collect the taxes, you can sell the property, and the constitution said that you had to have the auction on the courthouse steps like it was [00:15:00] back in the, you know, 18- not 1800s, 1980s.

Um, and what we did was, is we set up a bunch of computers on the courthouse steps with, uh, Wi-Fi, and we had an auction, and no one really contested it, but we realized that at some point that's probably going to be, you know, someone's gonna scrutinize it. So we lobbied to change the constitution to allow online tax sales, and we built the platform called CivicSource, which actually brought that entire antiquated process online.

Matt Klein: And, and was it still very Louisiana-specific or, or

did you-

Beau Button: At that time it was. So yeah-

Matt Klein: Yeah

Beau Button: ... I mean, it, it's, uh, tax certificates or tax liens, um, it's a, it's, it's a niche thing, but it can be very lucrative, but every state kind of has their own interpretation of it.

Matt Klein: Right.

Beau Button: Um, Louisiana has actually kind of, in the last five years, rewritten the laws to be more in alignment with the rest of the country, which the rest of the country kind of made sense.

Louisiana was weird and backwards.

Matt Klein: Right.

Beau Button: Um, but, uh, yeah, [00:16:00] it's, uh, it's, it's still there and they're still operating.

Matt Klein: So ha- had you thought about taking that business or that expertise and actually going outside the state or, or, uh-

Beau Button: We did. Yeah, no, we, we did. Um, the company does as of today operate out of the state, but I exited that company prior to them seeing success in, in doing that.

We had quite a few headwinds. Um, we worked with elected officials. We did have two-thirds of the state as customers. I mean, over the course of those 10 years, you know, basically people were like, "All right, well, this is the future." And we didn't charge the Uh, in the state of Louisiana, the person responsible for collecting taxes is the sheriff.

So those were our customers, but you, you don't charge them. We, we worked on a contingency fee model, so it was free to them. If we didn't sell the property, we didn't make money. But, uh, we, we also had dealt with patent infringement lawsuits. We were paying royalties to some other company that supposedly...

So I never saw it outside of Louisiana, but after I left, they were, uh, they broke through in I think, [00:17:00] uh, five or six different states. Yeah.

Matt Klein: Oh, fantastic. So what, so what made you eventually leave that and do something else?

Beau Button: Well, it's not a great story, but, uh, I went through a, a divorce. Um, and, and, and that business was probably partly responsible for the divorce, and then my marital divorce was probably partly responsible for my mental divorce from business partners, and I'm like, "I just need a complete reinvention."

I was over it. Like, I've learned my attention span with pretty much anything's 10 years. Where I live, not what I do. That's the... I mean, I, I've changed the things that I'm dabbling with. Like AI, I'm into it. I've got ano- another nine years, I'm pretty certain-

Matt Klein: Yeah

Beau Button: ... before I'm like, "Man, screw this. I don't know what I wanna do next."

But, uh, yeah, I just, I was kind of over it. Uh, and there was an opportunity. One of the, the other partners, uh, Brian, had made me a pretty damn good cash offer for my equity, and I said, "I'm out." Um, so it made it pretty easy.

Matt Klein: Yeah. So what did you do next then after that?

Beau Button: Well This is was-- I had some humble [00:18:00] pie, some humble pie.

So the divorce and math dividing by two, it, it, I, I lost my ass. So I ended up being an employee, um, a, a consultant for a company in the state of Louisiana who was working on claims pros-processing for, uh, it was called D-Deepwater Horizon's Economic Claim Center. So there was an oil rig in the Gulf of Mexico that basically blew up and killed a bunch of people.

Matt Klein: I remember.

Beau Button: And all of the oil-

Matt Klein: Yeah

Beau Button: ... basically hit the coastline and, and, and affected fishing, shrimping, and all of that. So I, I was brought on as a consultant to build, uh, or to, to contribute to building a, a platform that helped process those claims. But I was an employee. Uh, I, I've been an employee I think three times in my life.

There was a, a period of time that I didn't talk about before I started Web Genetic where I was a consultant for a company in Louisiana-

Matt Klein: Yeah ...

Beau Button: um, working on all kinds of things. Ironically, still in the oil and gas. Louisiana's big oil and gas, so I was working for a company as a consultant that built, uh, offshore [00:19:00] supply vessels.

Matt Klein: Um, but you know- I, I mean-

That's it ...

what

was it like to be an employee at that point after-

Beau Button: It sucked

Matt Klein: ... after

doing

your own thing

for a long time?

Beau Button: Now I was very f- I was fortunate. The employer, and here's where the real humble pie came in. Arcone Information Systems had leased an entire floor in an office building.

I was the person who signed the lease. And then as we kind of optimized the business, we realized we had oversubscribed and we started to basically subdivide and sublease. And one of the companies that sublet from us was called Lookfar Technologies, and that was the company that I landed a job for or with.

So after selling out of my e- my business, I was going to work in the same building on the same floor. So I'd see-- I was-- it was, look, I It, I got through it. I mean, I, I, I can't say I wasn't depressed to some degree, but I, I knew that this was just, you know, um, a temporary thing. But yeah, it was, it sucked.

Mentally, it sucked, especially the environment.

Matt Klein: Yeah.

Beau Button: 'Cause like I said, it was the same floor. I saw all of the... The thing that really bothered me [00:20:00] is we had spent an absurd amount of money on a 747 engine cowl desk for our receptionist. Um, and again, we didn't do anything right in the beginning. You know, it's a tax collections office.

You probably should be, you know, flying under the radar. You shouldn't have an engine cowl as your desk. It's a bitty, a bit boisterous. But yeah, I saw it every day as I went and sat in my fucking cubicle. Excuse my French. I don't know if I can say that.

Matt Klein: That's fine. Yeah. So, um, so then I guess you did that for a bit, and, and then I'm assuming from there you, you founded one of the companies that you are-

Beau Button: I did, yeah

Matt Klein: that you're still at. So, like, tell us a bit about that p- process, and I, I do wanna come back to what you said, is that you don't play games, yet you are now in the gaming space.

So tell-

Beau Button: Well, this, this is a good segue, yeah ...

Matt Klein: tell us, tell, tell

us how that, how that happened.

Beau Button: Yeah. I like a good problem. I mean, different problems than probably you like.

Um, th- there's a common thread, obviously. I like to solve problems. It's not so much like I love writing software. Honestly, that's why I really [00:21:00] love AI is because I can solve problems, but I'm not in the weeds and I'm not-

Matt Klein: Sure, yeah

Beau Button: ... you know, struggling to learn something and don't have access to the right resources.

But no, the, the way I got into the games was, you know, I had moved after about a year of working. I, I said, "You know what? I'm, I'm, I'm moving from the state of Louisiana. I'm going to Austin," which is like a tech center. Um, a lot of, you know, new and up-and-coming companies were here, and I was like, "That's a good shift."

Um, and I got a call from, uh, the younger brother of a high school friend, and he asked me if I wanted to go into business with him. He wanted to create a game company, and I was like, "No." I, I don't know how to build games. I don't play games. It's, like, so far removed from what I am and who I am." No. And, you know, didn't hear from him for a few months, and then he called me back 'cause behind the scenes he had outsourced the game and went to indie and, like, kind of learned the, the, the ropes a little bit about how to, you know, sell games and things of that nature.

His, his background's in user acquisition. He's like what we [00:22:00] call a growth hacker.

Matt Klein: Yeah.

Sure.

Beau Button: Um, he can get people. He can, you know, have them sign deals. But, um, he came to me and said, "Look, I got a big contract with the networking company Cisco. They wanna build a game for their Las Vegas convention," and he show-- He, he told me how much money was involved, and I said, "Sign me up."

So the business was started. We weren't gonna build our own games. We were gonna be a consultancy building games for other businesses and then offering our services for, you know, user acquisition, growth, and marketing. Um, and we did do that. We did that for about a year. We did it for about five different companies.

We made a little money doing it, and then we pivoted because Pokémon Go.

Matt Klein: Right. S- sorry, but you yourself built the games or you were-

Beau Button: Yep

Matt Klein: ... doing, like, technical leadership and outsourcing the programming? Like, tell us a bit about-

Beau Button: No, I just don't-

Matt Klein: ... a- as a non-game programmer, h- how did you learn that? Because it seems-

Beau Button: Yeah

Matt Klein: difficult, actually.

Beau Button: Somewhere around early 2000s, I had shifted from Visual Basic to C# when Microsoft released something called ASP [00:23:00] Plus, which evolved into something called .NET.

Matt Klein: Yep.

Beau Button: And I, I was, like I said earlier, VB, you didn't say it out loud. It was like, "Hey, what kind of... What, what language do you program?"

Visual Basic wasn't sexy. So when they released the first version of C#, I jumped in, like, headfirst and said, "All right, that's the language I'm gonna use for building web applications, desktop applications." But, uh, the, the game engine I chose, I chose because it's built using C#. Well, the engine's not, but the way you program the game.

Matt Klein: Yeah.

Beau Button: So I chose Unity. Um, when my partner had reached... Well, my now partner, when Sami, who had reached out two times, you know, said, "Hey, why don't we build games?" I started doing a little due diligence. I had never looked at a game engine. Uh, I, I knew what they were. I've read, like, articles on how, you know, John Carmack at id Software.

Like, I, I knew a lot about the moving pieces, but I had never built a game myself. But I, I spent about two weeks diving headfirst into how games were built. I found Unity. I was like, "C#, all right, this will be [00:24:00] an e- an easy transition." I was so wrong. It's a completely different paradigm, uh, coming from writing, you know, desktop applications-

Matt Klein: Yeah, sure

Beau Button: or web applications to, what the hell is a quaternion? Like, I don't know what the fuck that means. Like, how do you rotate th- uh, things in 3D space? Like, I paid attention in geometry and trigonometry, but, like, I haven't used this in so long. But, uh, yeah, so I, I, I knew we had to hire somebody. So I told my partner, like, "I'm not gonna be the guy that can do all of this, but we need to hire somebody," and we did.

We were fortunate, extremely fortunate to, to find a young kid who had just moved to Texas, whose only ever jo- uh, only other job was at a barbecue joint, but he was self-taught, and he had build, been building Unity games for, like, six years, and he was only, like, 18, so a very similar sto- maybe not six years, four years or something.

But similar story-

Matt Klein: Right

Beau Button: ... but, like, self-taught, never worked, never made money building games, but he knew a lot about Unity 3D. I met him at Starbucks, and we, we hired him, and, uh, we've... Actually, in the last four years, we've officially acknowledged that his, uh, he's our third [00:25:00] co-founder of Atlas Reality. His name is Morgan, and he's been with us, employee number one, and he's still with us.

Matt Klein: So, so tell us about those first games. Like, h- how did you pick the topics? Like, I guess you said that they were consulting-

Beau Button: Yeah ...

Matt Klein: in the

beginning, right? So-

Beau Button: Yeah, whoever's willing to pay us. We built a game for Cisco. Um, they had this conference in Las Vegas, and they wanted to build a space shooter game, uh, that taught about their new licensing model.

They were... It was, it was, it was horrible. It was, it wasn't... I mean, it was a game, by all means, but it wasn't something you just pick up and play. It was cute during the conference-

Matt Klein: Right. Sure

Beau Button: ... and then after the conference, it just died. Yeah. I didn't care because we made money off of that. I'd say the, the coolest game we built was for the Santa Barbara Zoo.

It was for the conservation of condors. It was called Condor Country. Um, and they had received a grant, um, a digital media grant that they wanted to, to use in a creative way. So we, we built a game that kind of allowed players to incubate and taught them about [00:26:00] how the zoo actually, you know, helps these endangered species.

But yeah, I mean, it was, it- it... That was the, the, the only fun game we built. We, we-

Matt Klein: Right

Beau Button: ... built another game called, uh, oh, I can't even... It was called Box and Seek. There was a kid from the UK that had a, a, a little money and wanted to build this game that was like hide and go seek, like Pac-Man. Again, money talks, bullshit walks.

He paid, we built the game. It was multiplayer. I did a lot of the back end with WebSockets and yeah, that was it. We re- we didn't choose. They kind of just came to us-

Matt Klein: Right

Beau Button: ... and we said, "How much money can you, you pay us?" And we, we did it.

Matt Klein: Yeah. I, I mean, m- my impression from the outside of the gaming industry is that it's, it's just brutal.

I mean, in terms of-

Beau Button: Well- ...

Matt Klein: the, the competition, the, you know, the speed. You talked briefly about Pokemon Go. So I s- so I wanna hear a bit more about that. But it, it just, it just seems very different style of software development than you would see in other places in the industry. So I guess [00:27:00] tell us a bit a-about that and how...

I, I guess my impression is that when you get into it from a consulting basis, you're a bit insulated because you're getting paid to make specific things from customers. But obviously, as you'll tell us, I mean, you've broken into making your own games.

Beau Button: Yeah.

Matt Klein: And at that point, now you're in the brutal competition space.

So I guess just, like, tell us a bit about what the game industry itself is like and how you went from consulting to making your own games and... I realize I just asked you like 10 questions-

Beau Button: No, it's, it's

good. I, I

got it. It's good ...

Matt Klein: so you can, you can, you can take them in whatever order you want.

Beau Button: Yeah. I, I, I'm spoiled because, like, I haven't had to deal with, like, the horror stories of the triple A game industry where they treat people like shit.

You got crunch. Like, I, I, I haven't had to deal with that. Um, you know, the, the transition from writing enterprise software and trying to be a game developer, you know, I never really stuck that landing. I- and even today, having been, you know, uh, Atlas, [00:28:00] uh, was founded in 2017, we're going on nine years now, I still can't open the Unity editor and really do anything useful.

Now, with AI and the official Unity CLI, I'm a really good game developer because all of the bullshit that I don't fully understand, I can just get done. But, uh, yeah, it was, it was honestly an easy transition, but what drove it, like I mentioned, was, you know, I had built the company that did real estate, distressed real estate sales on the internet using GIS mapping technology from Arc, uh, Esri.

And when Pokemon Go came out, like, I wasn't a Pokemon fan at all. I knew who it was or what it was, but I was just like, "W- what are-- These are zombies. Like, what, what is so different about this game? Like, h-how did they get so many people overnight to just go out to the damn, the ballpark and just start collecting Pokemon?"

And my, my partner and I were, like, dissecting this, and I had probably foolishly made the comment like, "Dude, I've got so much GIS experience [00:29:00] We could easily build a location-based game, 'cause that was like the first location-based game I had heard of despite Niantic having built, uh, uh, Ingress. Prior to Pokemon Go, I had never even heard of it.

I didn't know that was a thing. I had never seen somebody play a game where you actually used real maps and walked around.

Matt Klein: Right.

Yeah.

Beau Button: So I kinda fell in love with it because I was so familiar with what powered that. Um, and, and that's when my partner and I kind of made the decision to, like, go all in and let's come up with a, a game concept that is location-based.

Um, and we did a Kickstarter for a game called TOWAR, T-O-W-A-R. Um, this was when, like, Pokemon was really leaning heavy into the augmented reality. You've gotta hold your phone up-

Matt Klein: Yeah, yeah

Beau Button: ... and you flick the Poké Ball.

Matt Klein: Right.

Beau Button: Like, that was cute. Uh, the reality is, is that's, that's kind of, it's gone. Like, you don't do that anymore because it's, we- there's a syndrome, it's like gorilla arm.

Like, you get tired of doing that after a while, so you disable it and you just play it like regular game. But yeah, my partner was more on the game design side. He came up with the concept that was [00:30:00] originally called TOWAR We got our Kickstarter funded. I think the, the, the... We raised $30,000 to build that ourselves, and then we evolved that into a game called Atlas Empires.

Uh, we-- The name was hard. People were calling it Two War, Towar, so I was like, "You know what? We gotta rebrand this thing." So we built a game called Atlas Empires, which was, uh, the way that people would, would, would r-refer to it as like if Pokémon GO and Clash of Clans had a, a baby. It's location-based, it's strategy.

You build your, your base, and then you can walk around and see other people's bases on the map, and you can attack them. So it was fun, honestly, and it was a very well-thought-out, long-tail game that had a lot of really, you know, just thoughtful progression. Um, it just was really hard to get users, like you said a moment ago as well.

There's so many new games.

Matt Klein: Yeah.

Beau Button: So it wasn't, "Can you build it?" It's, "Can you get users to play it, and can you monetize those users?"

Matt Klein: Right.

Beau Button: Like, once they're in there, like, you've gotta really think about how are you gonna [00:31:00] charge them. Like, how are you gonna make money off of it? Not just ads. I mean, you can show them ads all day long.

Matt Klein: But, but, like, had you thought about the growth hacking and the monetization side ahead of time? Or, or were-- was your first goal to, like, make a game that someone would play, and then you said that you would figure it out later? I'm, I'm just like-

Beau Button: It was, it was more of the latter ...

Matt Klein: trying

to understand how you, how you sequence those.

Beau Button: Yeah, so m-my partner's background is in marketing, so it was a common part of everything we did. We, we, we, we-- When we built games for other companies, if it was an internal game, it wasn't something that we could do.

Matt Klein: Sure.

Beau Button: Like the Cisco game, there was no growth hacking. It was just like build it, put it into Ga- the, the Play Store or Google Play or, uh, Apple App Store, and then you're done with it.

But, you know, we had kind of learned about, you know, the creatives you needed to create. Where did you wanna, like, run advertisements? Uh, and, you know, we, we did talk about that, but I think we kind of lost sight of that because we were so excited about building the game. And, and this is common when people build passion project games.

[00:32:00] You, you, you focus way too much on, like, what you think players want.

Matt Klein: Sure.

Yeah.

Beau Button: Um, and, and there was an element, I think, of, of that that I, I think contributed to ultimately the demise of Atlas Empires, which is we didn't ask other people. We just, we built what we wanted to build and, uh, it just, it didn't, it didn't work out commercially.

But, uh, yeah-

Matt Klein: Yeah. I-

Beau Button: ... we, we talked about it ...

Matt Klein: right. I, I was listening to this, uh, episode. I forget which show it was on, but it was on NPR, and it was recent, and they were talking about Candy Crush and just about how... And again, like, I know nothing about this, so would love to, love to get your thoughts on it, but I thought it was so fascinating.

It was basically about how that game has been being played for, like, 20 years at this point. It's made billions of dollars, and to this day, they don't fully understand why it is as popular as it, as it is, and people keep buying it and keep playing it and doing all these things. So- I don't know, like maybe-

Beau Button: I mean,

Matt Klein: then

Beau Button: [00:33:00] there's-

Matt Klein: Yeah

Beau Button: for, for, for Empires, there was a novelty, you know, you, you, it, like Clash of Clans or, uh, Clash Royale, you know, matchmaking is, is pretty opaque, you know? You, you, you match with somebody who could be anywhere in the world. You really don't know where they were. So, like, a location-based, you know, strategy game where you could attack another player's base, but you had to physically be near it was a-

Matt Klein: Right

Beau Button: I don't know. It felt awesome, honestly. You're out taking a stroll, you know, and then you see in the distance another, a, a, a base, and you're like, "Oh, shit, I gotta walk all the way over there and attack it." But, uh, yeah, I mean, it's, it's a tough, uh... I mean, my partner could describe all of this way better than I can, but I mean, ultimately, it, it was something that we, we lucked into.

A lot of this is luck. In Atlas Empires, we, we were lucky enough to get 100-and-something thousand users. We did run ads. We spent a lot of money. Now, I will say [00:34:00] we raised... After we built the game, we raised to the tune of $5 million. So this wasn't like-

Matt Klein: I see. Yeah

Beau Button: ... that $30,000-

Matt Klein: Right, right, right

Beau Button: ... and we had, you know, a, a successful games.

We pumped more money into the game, and we got the game to where we felt like we were almost there. We were growing. We saw the growth trajectory, but a lot of stuff also happened that kind of, you know, screwed that. The pandemic was, was one of them. Uh, you know, people weren't going outside as much, so we, we kind of had to, to rethink the business completely.

Matt Klein: Yeah. So, so you said that it fell apart. Was it just because that the growth curve wasn't there to sustain the amount of money that you were putting into it effectively? Like, it's, it's a typical business calculation.

Beau Button: It is.

Matt Klein: I mean, is that what happened effectively?

Beau Button: That's exactly right. So you spend a lot of money on UA, and if you don't have the return on that investment, you know, from the players that you're spending maybe $15, $20 to acquire-

Matt Klein: Right

Beau Button: if they're only making you $5, you're-

Matt Klein: Sure

Beau Button: ... you're hemorrhaging. So it was, for us, you know, we, we... Again, I don't think Atlas Earth [00:35:00] would be as successful as it is if we hadn't gone through that and felt that firsthand. Like-

Matt Klein: Right

Beau Button: ... oh, shit, we really do need to be more mindful of, like, how are you going to monetize?

How do you think-

Matt Klein: Yeah ...

Beau Button: how do you think about stickiness and retention? Like, what brings a person back in? Like-

Matt Klein: Right

Beau Button: ... we, we didn't do any of that. We had some things that organically brought people back, but it wasn't at the forefront of everything that we did.

Matt Klein: Right.

Beau Button: Whereas with Earth, everything kind of centers on, all right, retention, how do you get them back in the game?

How do you make it sticky? Now, it's, it's a game we pay people in cash. That's probably the reason why it's successful. We pay people, you know, uh, it's over $13 million in cash we've paid now, so that doesn't hurt.

Matt Klein: Right. So, but I think that comes back to my m- my original question, which is that it sounds like what changed between the first game and the second is before you even designed the game itself, you were thinking about, like, all of the-

Beau Button: No, absolutely

Matt Klein: acquisition costs and the business model and all of those [00:36:00] things.

Beau Button: Absolutely.

Matt Klein: Um, and I'm asking these questions because I know nothing about this. I think this is super interesting, is, I mean, tell us a bit more from what you can share around, like, what are, what are the levers that you were thinking about?

Because, like, to me, what I find interesting about this is it- it's almost like a catch-22. You could spend all day and you can dream up the most amazing acquisition levers and business models, but if the game itself is shit and no one wants to play, it's like, I don't know, you've done nothing. So it's, it's almost like the two of them have to go hand-in-hand a bit, right?

So it's like-

Beau Button: They, they do

Matt Klein: ... te- tell us a bit about the process of what it was like to design the second game and how you were trying to balance it being fun versus being an okay business, I guess.

Beau Button: This is gonna, this is gonna sound strange. As the co-founder of Atlas Reality, it's gonna be strange to hear.

Fun is not necessarily a word I would use to describe the, the, the totality of Atlas Earth.

Matt Klein: Okay.

Beau Button: So we need-- We were inspired, much like we [00:37:00] were inspired by Pokémon GO to build Atlas Empires, we were inspired by the cryptocurrency movement, Web3, blockchain, virtual real estate, NFTs. Like, we were inspired by that.

Like, Snoop Dogg had a mansion in a game called The Sandbox. It was all on chain, but I was like, "All right. I need to learn more." So much like I learned about Unity, I dove headfirst into smart contracts and all of this crazy shit, and I got to the bottom of it and said, "This is snake oil. I'm out."

Matt Klein: Mm-hmm.

Beau Button: But if, if we weren't paying players, if, if there wasn't this, this allure of I can make real cash, not some token, not some crazy intermediary, like real cash, I, I don't know if Atlas Empires would be successful-

Matt Klein: Yeah

Beau Button: ... because it's virtual real estate, so there's an element of, like, you know, Monopoly, you walk around, you buy land.

And we do have a lot of players that love to just buy land because if you own the most land in a city, you're the mayor. If you're the, the, the largest landowner in the state, you're the governor. If you own the most land in the country, you're the [00:38:00] president. So there's some game mechanics that would drive that, but I don't think we'd have the number of people we have today if we-- they weren't earning money.

Matt Klein: Yeah.

Beau Button: So like I said, I, there are other companies out there who have built games that aren't play to earn that I think could answer that question better than I can because we're spoiled. You know, we had a giant, delicious carrot on a string, and people are like, "Oh yeah, I don't really give a shit-"

Matt Klein: Yeah

Beau Button: ... "if this is fun or not.

I wanna make beer money. I wanna make $5 a month, you know, watching ads and buying virtual real estate." So I mean, we, we did, like you had mentioned or I mentioned, we, we kind of- We- it was at the forefront of our decision. Like, I, what is the business model? And one of the things that really proved to, you know, Sami and I that this was viable is over the course of, uh, a weekend, Sami and I, we talked about virtual real estate, we talked about, you know, how we can take the money we make from selling the land.

Um, and I can't get into too many details behind the scenes, but-

Matt Klein: Sure

Beau Button: ... you know, essentially our business is a profit sharing. We [00:39:00] take revenue, we manage our own treasury, our own investments, we grow that, and that is essentially the backstop for the virtual rent that that land generates for people. So it's not like we're just taking your $5, putting it in a bucket, and then just waiting for you to get the $5 somewhere else.

It's like we're growing that money intelligently.

Matt Klein: Right.

Beau Button: So Sami's background came from a company called Acorns, which is micro-investment. So all of this kind of played into what is Atlas Earth, you know, behind the scenes. But, uh, yeah, I mean, it's like I said, it's, it's, without the play-to-earn aspect, it's just the, uh, you're collecting squares on a map.

Um-

Matt Klein: Right

Beau Button: ... but we, we did put a lot of thought into how we can make money, the different revenue sources, in-app purchases, incentivized video ads, and then today we've got about nine different revenue sources. We have a hotel booking platform. You can book hotels through the system and earn. Uh, the Tesoro company, which is the, the spinoff, you know, the, it was born in Atlas Empires really.

Then we implemented it into Earth, and then we said, "Why are we keeping this to ourselves? This is by far [00:40:00] the best idea we ever had. It needs to be its own entity." But, uh, yeah, my, I'm sure there are many other people out there that have built way better games, like brilliant, and, and I've seen passion projects at these, like, game conferences.

I'm like, "Shit, this is incredibly fun." It's just how do you make money? You, what do you charge $9 once and that's it?

Matt Klein: But then do you, uh, m- you know, this is maybe like a meta question. Are you a game or are you something else? I mean,

it's like-

Beau Button: No. So that, I'm glad you asked that because for the longest we were masquerading as a game.

And, you know, if you go to Google Play, we're in the finance category.

Matt Klein: Right.

Yeah.

Beau Button: So we are a rewards pla-

Matt Klein: Right. Okay

Beau Button: ... uh, a, a cashback app. So there's other apps out there like GetUpside and Dwolla, like these things that reward you for doing things. Uh, GetUpside is about fuel. Um, and it's pretty lucrative if you're religious about going to wherever the hell they tell you to go to buy your fuel.

You can get 10% back. If you drive an F-250 in Texas, that's $10.

Matt Klein: Yeah.

Beau Button: All right, who doesn't [00:41:00] want an extra $10?

Matt Klein: Sure.

Beau Button: I'm not driving out of the damn way to go get $10, but a lot of people will.

Matt Klein: Yeah.

Beau Button: But we've, we've kind of rebranded, and we don't really m- mention that it's a, a game. It's, it's a cashback app, and we put ourselves in that same category as, you know, GetUpside.

Um, there's another one I can't remember right now that does receipt scanning, et cetera. But no, we're a cashback app.

Matt Klein: Okay. All right, cool. And then before we turn our attention to AI, which I know you're passionate about, just tell us very briefly about your spinoff company and-

Beau Button: Yeah

Matt Klein: ... just, uh, you know, what was the process of spinning it out?

Why did you spin it out as opposed to keep it as a larger entity? I'm just curious what the story there is.

Beau Button: Yeah. So Atlas, uh, Earth or Atlas Reality with the, you know, the newfound success of, of Atlas Earth, we were profitable, which was great. We had done some series C, you know, uh, safe convertible debt to, to get us there, and it was really, a, a really great feeling to, to look at the, you know, [00:42:00] the finances and say, "Oh, shit, there's more money coming in than there's money going out.

This is-" "... we're, we, we've made it."

Matt Klein: Right.

Beau Button: Um, but, uh, yeah, I mean- we, we had built this technology on an existing company that we had acquired through an all-stock deal. We had a mutual investor. The tech that powers Tesoro existed for several years before we implemented it into Atlas Empires. But the, the company had built this cash-back platform that paid people instantly in cash as a credit back into their, their credit card statement.

So you'd go to a restaurant, you'd pay $100, and then instantly you get a $10 credit back. They had to front that money, and they just couldn't make that work from a cash flow perspective. But they had built the rails. They had already integrated into the credit card networks. All of that work was already done, and the mutual investor was like, "You guys should do something with this, but instead of giving cash back, give virtual currency back."

And we're like, "Oh, man, it's-- Would that work?" [00:43:00] You know, because the, the beauty of the way that the software worked was there's no point of sale integration, there's no QR codes. The merchant, when they sign a deal with us, does nothing. All we need is their VMID, uh, which is a unique identifier on the Visa network and the equivalent for MasterCard, and we're able to basically, you know, tap into their transaction data through the credit card network directly, you know, transparently to them.

So the user experience as an end user, uh, it-- When we built the technology into Atlas, uh, Empires, we had called it, uh, internally it was AMP, the Atlas Merchant Platform. And the user experience was incredible. You could literally be in the drive-through line, watch the, the, the cashier swipe your, your card or tap your card, and then you get a push notification in Atlas Empires in real time saying, "Thanks for buying a burger at Burger King.

You just earned 100, you know, 100 runes." There was nothing that came close to that kinda like dopamine hit for doing something you were already gonna do and getting that, you know, in-game currency, which, you know, allowed you to go attack a base or something like that. [00:44:00] So we, we knew this needed to be part of Earth, so we kind of rebuilt it, integrated it into Earth, and then we were at this point where we needed to raise more money for Atlas Earth, uh, under Atlas Reality, and we ran to a, a brick wall.

People were like, "I invest in games What are you? Are you a gaming company?

Matt Klein: Right.

Beau Button: 'Cause you got this, like, rewards platform you're calling Amp, but that's not a game. And we're like, "Well, we're both. We're a rewards platform and an ad tech platform." So we could never find an investor, um, VC or even, you know, anybody to contribute to a seed round to give us enough money to lead that was okay with both of those.

Um, and it was actually, uh, one of our investors that, you know, casually mentioned, "You know what? We should just split the company and, and treat it as its own entity." Um, and, and my first reaction was, "No, I'm not ... The infrastructure, I'm the guy that has to set up all of the cloud infrastructure-

Matt Klein: Yeah. Sure, yeah

Beau Button: the organization, the mobile device management, the security policies." Like, like, in my head, I was [00:45:00] like, "No, I don't wanna have to do any of that." But after about two months of continuing that conversation and talking to our attorneys, it was very clear that that's what we needed to do. And, uh, about a month after we finalized it, we closed our $5 million plus Series C round for Tesoro, and off to the races for Tesoro.

And then, uh, because Atlas was, you know, profitable, the, the money we were making for Atlas was enough to facilitate improving-

Matt Klein: Yeah

Beau Button: ... Atlas Earth. It wasn't enough to do both of those two things-

Matt Klein: Right

Beau Button: ... 'cause the, the Amp platform was working, but it's, it's, it suffers from, you know, low DAU. So you, you need to have a huge number of daily active users to drive the number of transactions to drive revenue, 'cause we basically take a small, you know, piece of the, the pie, the commission from the sale, and we, we couldn't support both of them with the revenue-

Matt Klein: Right

Beau Button: coming. And it didn't feel fair, honestly. It's like, why is ... The game's making all the money. Why are we taking that money and pumping into this?

Matt Klein: Sure.

Beau Button: We spun it off.

Matt Klein: Yeah. Cool. [00:46:00] Um, could talk about this more, but, uh, would, would love to learn about your, your AI interests, 'cause I know when I talked to you last time, um-

Beau Button: Yeah

Matt Klein: I, and, and that was, that was probably four or five months ago. In the world of AI, you know, things change every two weeks, but I know that you were building your own, like, cluster in, in your office. I don't know. I would love to learn about-

Beau Button: Yeah

Matt Klein: ... your current situation.

Beau Button: I, I, I always say I've got, like, a Spidey sense.

Uh, you know, uh, I grew up in Louisiana. We always had to evacuate for hurricanes. Both my partner and I grew up in Louisiana, and, and this has nothing to do with natural disasters, but, like, it's an intuition thing. It's like, you know what? I know what's about to happen. I called it for the pandemic. We shut our office down three weeks before the state of Texas made it a mandate.

I was like, "Go buy some black beans and rice and just sit at your house. Take your office computer home. I don't wanna have to deal with this when they make this decision." But, you know, the, the reality is What we're dealing with right now is, is grossly unsustainable. Um, I don't think [00:47:00] that's a secret. You know, if you look at the cost of electricity, the cost of GPUs, and the cost of a subscription to any of the z.ai, Codex, a-a, Anthropic, it doesn't matter.

Like, it just, the math doesn't math. And I've always been one... I, I, I've never really been able to realize, like, platform, you know, uh, portability, you know. Like, Kubernetes at one point in time was, I was convinced that was gonna allow me to, to run workloads in AWS, GCP, and Azure, and, like, just drag and drop shit across.

That never happened, trust me. Um, uh, and I went down some rabbit holes, and you know a lot more about all that.

Matt Klein: Yeah.

Beau Button: I was like, "What is Istio?"

Matt Klein: For sure. Yeah.

Beau Button: "What's a service mesh?"

Matt Klein: Yeah.

Beau Button: And I was like, you know what? We're, we're married to Azure. We're never moving. So, like, I'm at that point, or I was at that point when, you know, you and I fir- first met, where I was like, "I see the writing on the wall."

Matt Klein: Yeah.

I do too, for sure.

Beau Button: It's, it's a drug, and I'm not gonna lie. I'm addicted because, like I, I said earlier, it's not that I loved programming. You know, early on I think it was a, it was a curiosity, like, "How does this work? What, what am I doing? How do I make this... How is this machine even [00:48:00] doing anything?"

And, and I can't say I, I, I, I've fallen out of love with it, but, like, I'm more interested in building software and solving problems than I am in, in building the sausage and the inter- so, like, AI to me has kind of reinvigorated my passion for building things because, you know, humans are difficult. That's the most challenging thing for me as a startup co-founder is humans.

Between the two companies, we've got roughly 50 people, and we're fortunate. Right now the team is rock solid, but there's ebbs and flows. Like, you-

Matt Klein: Of course. Yeah

Beau Button: ... you have to hire, you fire, and it's like, "Goddamn, man, I'm going through this again." So having access to something like AI where I, I talk to AI just like I talk, I'm talking here.

I'm very blunt. I do drop the F-bomb pretty frequently. Um, and, and there's all kinds of strategies on prompt engineering. I've mostly read through most of the books and I said, "That's bullshit," because I get exactly what I want from these things, and I don't do any of that. But it's because I know what the, the output is.

I can judge its output and steer it-

Matt Klein: Yes

Beau Button: ... to get what I need.

Matt Klein: For sure. Yeah.

Beau Button: It's not like I'm just saying, "Here, I'm gonna one-shot [00:49:00] facebook.com," and, like-

Matt Klein: Yep

Beau Button: ... "Oh, shit, it doesn't look anything-" Yeah ... "like I'm expecting." But, uh, yeah, I've, I've made a, uh, the company has made a pretty significant investment in local AI.

Um, it pain- it's painful because these GPUs, you know, two years ago were a f- a quarter of the price. But, um, I'm trying to build something in-house that is close enough to Claude. Um, Claude Code was the, the first kind of coding harness where I'm like-

Matt Klein: Of course

Beau Button: ... "Holy hell." This like, you know, ChatGPT was cute.

You, you, you, you ask it to build something, you copy the code out of the actual chat editor, you paste it into Visual Studio Code, and then that kind of made its way into Visual Studio Code. I used a tool called Cline, and that was the first time where those two worlds kind of came together and the IDE and the LLM were actually working together.

We had tools and I'm like, "Holy hell." But I just don't wanna find myself In a position where Anthropic has me by the, you know, the, the huevos. Um-

Matt Klein: Yeah

Beau Button: ... and where I've got-- I'm, I'm in front of a Mac [00:50:00] Studio with five hundred and twelve gigs of VRAM. It's the last generation, not the new one that they just announced, and I'm like, "Oh my God, it's the cost of a Kia."

Uh, honestly, I've, I've been considering putting this thing on eBay because the thing's quadrupled in value. Um, I do have a cluster of DGX Sparks. Um, it's not the greatest machine, but when you, when you cluster them, you can, you know, run a, a reasonably large model. It's got basically a half a terabyte of shared unified RAM between them, but the networking hardware, just to get those things to connect to each other is like three thousand dollars because it's ten gigabits per second.

Matt Klein: Yeah.

Beau Button: Um, and then I just bought two RTX 6000 max Qs from a company called Exact. Um, we were accepted into the NVIDIA Inception program, and there's a discount for startups, so I took advantage of that. But I have not cracked the nut yet. I am in the middle of evaluating harnesses, um, and I've gone off the deep end to every harness, open code pi, deep, uh, seek, uh, harness, and I'm, I'm at the point now where I'm actually building my own because all of those are shit, in my [00:51:00] opinion.

Um-

Matt Klein: Yeah, I mean, I, I, I feel like everyone reaches the point at which you're looking at the harness logs and you're having that thought of why, why are you not building your own? A-and we, we could talk about this for an hour. Maybe we should do a follow-up episode. But the, the first thing that comes to mind, which I've thought about myself, is bitdrift is smaller than your two companies.

We're like twenty-five, twenty-six people now or something along those lines. And I do share all the concerns that you shared about that the current situation is not sustainable. We are increas-increasingly reliant on these tools in some form. I think things have changed a bit even since I talked to you, because now, in my opinion, there are multiple competitors who are almost equivalent at this point-

Beau Button: Yep

Matt Klein: which further drives down prices. But I, I think my question to you as, I don't know if you would consider yourself a small business owner or a medium business owner or what, but I think it's interesting is that how do you decide that [00:52:00] it's worth investing that much of your time and your capital in this super expensive hardware, right?

For stuff that locally is not gonna get the same kind of utilization that you're gonna get if you're using shared resources or paying someone. So I-I'm just trying to understand, like-

Beau Button: Yeah

Matt Klein: ... how do you reconcile the effort and time and, like, dollar capital that you're putting into that versus you could be doing any number of other things to, to build your business?

Beau Button: I would say 80% of the justification is for my own mental curiosity and mental health. Uh, I, I went into this, uh, good and well knowing that I may not be able to actually replace what I'm getting from Anthropic or from OpenAI.

Matt Klein: Yep.

Beau Button: But as an engineer and a, a, a hacker that like, if... I've got so many things.

I mean, I have, I got- I probably do have ADHD. I'm into so many different things, and I, you know, I, I basically like preemptive multi-thread, and I don't get anything fucking done really at the end [00:53:00] of the day.

Matt Klein: Sure. Yeah.

Beau Button: But, like, I wanted to know how it all worked. I wanted my hands on it.

Matt Klein: Yeah.

Beau Button: I miss the days of racking 1U servers and having physical hardware and Screw Terraform and infrastructure as code.

Like I, so my, my partner and I, we, he does some things. He'll sponsor a conference and I'm like, "Oh Jesus, that's a lot of money." And sometimes it pays off, sometimes it doesn't. But, uh, I was, I was pretty adamant like, "I need this. I need to know how it all fits together."

Matt Klein: Yep.

Beau Button: And then I am banking on one thing, and this was something that I'd, uh, I was thinking, and then I read an article from somebody who's way, uh, more, uh, articulate than I am about how the Chinese are going to basically saturate the American LLM market and it's going to ex- it's gonna implode.

Um-

Matt Klein: I

agree with you.

Beau Button: I think it's gonna happen. Yeah. And guess what? I, I cannot fucking wait. Uh, this new version of Quinn that I'm running on the RTX machine is amazing, but it, it's not close enough yet to-

Matt Klein: Yeah

Beau Button: ... what I can [00:54:00] get through Anthropic. But I see, I see every time I replace one model with another, I'm like, "Damn, we're so close."

Matt Klein: So close.

Beau Button: And-

Matt Klein: Yeah. Yeah ...

Beau Button: and, and I feel like I, I think in the next month or two I'm gonna get to a point where I could comfortably replace one of the open source, I mean, forgive me, one of the frontier models with what I'm running. Um, and I'm running an LLM gateway. I've got guardrails. We've got all kinds of things that you just don't get unless you're like an enterprise Anthropic customer from them.

And it's the Wild West in our company. And you know, I've always... We have patents, so Tesoro has like nine different patents and I'm always, uh, I don't like patents. At the end of the day, you know, I kind of share, not that I love Elon Musk, his same fucking sentiments on patents. It's like, no, that's for the weak, you know?

Uh, we, we just get it done first and do it better than anybody else can do it.

Matt Klein: Mm.

Beau Button: But they do serve a purpose, especially when you're raising money. But it's like I, I just, I don't know. I just feel like I'm gonna get there and we're gonna be able to use it. I've got guardrails. I use something called Tailscale.

We've got basically-

Matt Klein: Yeah, sure

Beau Button: ... a mesh [00:55:00] network. Uh, the authentication's all provided through Tailscale. Um, they have their own, uh, it's, uh, called Aperture, which is an LLM gateway built into Tailscale. So like it wouldn't take much. Once I find a model and a harness that gives me what I want, like, and I'm very particular and I've been doing this literally two-thirds of my life.

If it satisfies, satisfies me, I can give it to any one of the engineers that, that we employ, and they will fucking love it.

Um-

Matt Klein: Do you, do you have any of your engineers currently using your internal systems, or are they still mostly using-

Beau Button: Nope ...

Matt Klein: the frontier services?

Beau Button: Nobody has access to the pin on that.

Okay. Not yet. I, I would say By the end of the year, that, that's kind of my objective. Uh, again, I've got a lot of real work that needs to get done, so I don't spend a whole lot of time on this. Um, when new models come out, you know, I'll read a, an article on Reddit. Um, I was using GLM 5.3 from z.ai, and I played with it, and it's really powerful.

But, like, I'm just passively listening. I have, you know-

Matt Klein: Yeah

Beau Button: ... a little bot that goes out and looks at Hugging Face and kind of gives me an update on what's happening out there, and I'll spend [00:56:00] two hours a week playing with it. But-

Matt Klein: Yeah

Beau Button: ... it's just not there. By the end of the year, though, I, I'm fairly confident that I will have some of the engineering team using the hardware that we've invested in.

Matt Klein: Yeah. So, um, again, we could talk about this for a long time, but unfortunately we're pretty much out of time. Where, where i-i-if you had to guess, where do you think things are gonna head in the next few years with all these tools? I mean, I, I think you've already made your own personal opinion clear, is that you think that things are going, uh, you know, more, more towards open weights, which I personally tend to agree with.

Um, but what, what do you... If you had to guess, what do you think in three years? What, what do you think the company and AI ecosystem is going to look like?

Beau Button: I am optimistic that there will be, uh, a legitimate democratization of AI and the hardware cost and all of that will, will come down probably as the result of the whole thing crashing and burning.

Um, GPUs, you know, you went through [00:57:00] crypto and then all of a sudden something-

Matt Klein: Yep

Beau Button: ... else came up that juiced the cost of DRAM and, you know-

Matt Klein: Yeah

Beau Button: ... GPUs or Nvram. Um, I do think that will happen. I think realistically, probably in the next three years, I think everybody will probably have access to extremely, you know, the efficiency of the, the, the models, the, the requirements will come down.

So you may not need a Mac Studio, but I'd love for every home to have access to a device like that for open, you know, for, for AI communication. I, I, I, I think, I feel it. I can't say it's gonna happen because-

Matt Klein: Yeah

Beau Button: ... who knows what's gonna happen. It's the Wild West in this. The, not even just this country, in the world, but that's what I want.

That's what I want for the, the world, is everybody to have access to it because it is extremely powerful. Um, I hesitate to talk about what I think's gonna happen to, to humanity and employment and all of that because it's, it's really bleak. Um, I'm not, I don't, I'm not qualified to articulate that in a way-

Matt Klein: Yeah

Beau Button: that's, does- doesn't make me sound like a doomsday prepper. It's just like, "Oh, shit, everybody's gonna be unemployed." But, uh, yeah, I mean, for us, I think we are at the point where we are a [00:58:00] AI first organization. We're already there with the frontier models. Everything we do is centered around, all right, well, how can we leverage AI to do more with fewer people?

Um, and the one thing I said previously on a- another podcast was, "I have not fired a single human because of AI, but I have not hired a replacement for people who have all of, either voluntarily left or have been fired for other infractions." And we're probably gonna remain this size for a very long time.

And-

Matt Klein: Yeah, I, I, I mean, I would be lying if I s- said that at bitdrift, um, we, we, we also have not obviously fired anyone due to AI, but I think we've hired slower than we otherwise w- would have because of these tools. I mean-

Beau Button: Yep. Absolutely

Matt Klein: ... we can get more done with them, for sure.

Beau Button: Yep.

Matt Klein: Yeah.

Beau Button: So-

Matt Klein: Yeah

Beau Button: ... I, I'm, I'm betting on worldwide accessibility.

I mean, that's maybe a bit, uh, optimistic because, you know, you're not gonna have homeless people using little super AI computers, but, uh, that's another problem we should solve. But, uh, I'm, I'm fairly certain it will be extremely accessible, and I think [00:59:00] that the Chinese are gonna drive that. Um, I, I don't know of another company, uh, especially a, a, a local or domestic company that's-

Matt Klein: Yeah

Beau Button: building open weight models that comes close to what they're doing.

Matt Klein: Yeah. Okay. Well, thanks, Beau. That was, uh, uh, a great chat. I think that's a great place to end. Um, so that's a wrap- Great ... for this episode of "Beyond the Noise: Signals, Stories, and Spicy Takes." Huge thanks to Bo for joining and sharing his story.

You can find this episode and all past ones on the bitdrift YouTube channel. If you had fun, drop us a review, tell your friends, or yell your favorite hot take into the void and just make sure to tag us. I'm Matt Klein, and I will see you next time. Thanks a lot.