Episode 934 ·
Why AI is Actually Artificial Creativity with Brian Alvey, CTO of WordPress VIP
The robots were supposed to do the dishes. Instead, they're writing poetry.
Today, we're talking to Brian Alvey, CTO at WordPress VIP. We discuss why AI should be thought of as artificial creativity rather than artificial intelligence, how modern CTOs no longer need large teams to build and ship at scale, and why embracing shortcuts is the only rational response to the pace of technological change.
All of this right here, right now, on the Modern CTO Podcast!
To learn more about WordPress VIP, check out their website here.
About Brian Alvey
I build software that makes creative people more powerful. I've built software platforms used by hundreds of global media and retail brands including Engadget, TMZ, Gucci, AOL.com, BusinessWeek, Netscape, Tourneau, Mashable, Michael Kors, Capgemini, Dow Jones, Red Bull, MySpace, Uniqlo, Ogilvy, Nationwide, Dell, VidCon, Discovery, the BBC, Best Buy and The Ellen DeGeneres Show. I’m currently the CTO of WordPress VIP, the hosting platform used for high traffic sites including CBS, Nexstar, NBCUniversal, Al Jazeera, Capgemini, Samsung, the White House, AccuWeather, Salesforce, Facebook, TechCrunch and News Corp.
Transcript
(Intro Narrator at 00:00:00) Today, we're talking to Brian Alvey, CTO at WordPress VIP, about how he sees AI not as artificial intelligence, but as artificial creativity and more. You're listening to Joel Beasley, Modern CTO.
(Joel Beasley at 00:00:21) Josh called me and he said, "Hey, I've got the CTO of WordPress, and you're going to get to talk with him." And I was very pumped up about that. So how did you get that role?
(Brian Alvey at 00:00:31) Yeah, so it's CTO of WordPress VIP, which is a smaller piece of the company Automattic. So WordPress itself is a massive ecosystem, and then Automattic is the company created by the co-founder of WordPress that has a whole bunch of consumer products like WordPress.com, journaling apps, Tumblr—a lot of things that aren't enterprise. And then there's an enterprise division called WordPress VIP.
(Brian Alvey at 00:00:54) So I'm the CTO of that, and that's the one that hosts thousands of sites you've heard of. So the White House, NASA, Salesforce.com, tons and tons of stuff. The New York Post, Rolling Stone, The Onion, Savannah Bananas, all sorts of things. Enterprise WordPress, then?
(Brian Alvey at 00:01:12) Correct. So CTO of that, not CTO of global WordPress or consumer WordPress, but enterprise WordPress. Yes.
(Joel Beasley at 00:01:18) So if I need a WordPress instance and I need it to have serious uptime and security, then I would be signing up for WordPress VIP.
(Brian Alvey at 00:01:29) Especially if you have tons of traffic, lots of DDoS attacks. You're a global brand. Yes. So we are definitely the ones who can't say no to people.
(Brian Alvey at 00:01:38) So if you run into extreme WordPress challenges, you have a terabyte primary database—you know, most people have a gig, three gigs. 100 gigs is a lot. We get people with 900, 1,100. Or you're under attack all the time, like a White House or a NASA, or need to be federally FedRAMP authorized, then that's WordPress VIP, enterprise WordPress.
(Brian Alvey at 00:02:01) And how long have you been doing that? Four years. And before that, I was definitely a WordPress outsider, like a friend of WordPress, but just built my own things, always competed with WordPress. So I knew Matt in year one of his thing, but when they started, I'd been building CMSs for eight or nine years myself, and it was always custom.
(Brian Alvey at 00:02:22) And I just thought WordPress was interesting. You know, it's like, I get what it does, but I don't think it's going to be around five years. And now I think it's 23 years old, which is pretty remarkable.
(Joel Beasley at 00:02:33) So you weren't a Joomla person who became a WordPress fan. You were, like, a roll-your-own?
(Brian Alvey at 00:02:37) Yeah, exactly. No, I installed all those things, but I always built a lot of brands, and I built a lot of platforms, but it was always the platform for that thing, very custom stuff. And I didn't think—I would actually tell you, even today, I don't think that something you build for everyone—the way I used to say it was, if you build something for everyone, it works for no one. And that's kind of the truth. But the way that WordPress kind of made it through that little rule is it's open source, and it's got a plugin architecture.
(Brian Alvey at 00:03:13) So pretty much anybody running WordPress VIP, these enterprise WordPress things, New York Post, Rolling Stone, those brands—their sites are so radically different because they've customized them so heavily. So WordPress works more as a framework or a scaffolding. It gets you a couple years of a build, but everything else you do is yours. So no two WordPress applications we work with are the same. And I think that's why WordPress has survived, because it's like that Ship of Theseus thing, right? Early WordPress, all PHP. Today's WordPress, a ton of React and JavaScript and more modern things and a build process. And it's, you know, strangely, while the whole software stack has changed in that time and it's modernized, and the AI stuff they're doing is incredible, it's got a lot of backward compatibility.
(Brian Alvey at 00:04:01) The basic data tables haven't changed. It takes them years to vote on adding a ninth or thirteenth or fifteenth table to core WordPress. It's bonkers how much backward support there is and how much support there is for "download this and run this on anything you want." And I don't really have a stomach for that, but my gosh, I'm glad they do it.
(Joel Beasley at 00:04:23) So one of the things I really want to hear you talk about—everyone talks about AI to the point where it's like, okay, here we go. You have this really interesting idea about artificial creativity. Can you share that with me?
(Brian Alvey at 00:04:40) Yeah, absolutely. It's why people are freaked out, and why they're freaked out about jobs and the future of work. So everybody expected artificial intelligence was going to be the decades of compute that led up to this moment where we're just sort of making a calculator better. Instead, it's not. It's not solving the things that we thought technology was going to solve. So the example I give is, I had a dinner in LA ten years ago with some cool friends, and we would always ask an icebreaker question. And we could do this now. Like, when you look around the room, which one of us is the last person where robots are going to take your job? And now you would say it's agents or AI. But back then, we would say robots are going to take your job. And these were people who were managers of comedians who were on Saturday Night Live and famous people, right? And they're like, "Okay, well, first off, bots can't do that. Robots can't write a five-minute comedy routine. They can't write an Adam Sandler movie. They can't write a 22-minute sitcom. That's just—they'll never do that. They can totally stock a warehouse, drive a car, you know, do these sort of oil pipeline tasks, things like that." And what happened was we ended up creating AI that was language-based, and the first thing it came for was writing. And nobody was ready for that. And it does it well. That's the problem. It's not like it doesn't do it well. It does it really well. And every six months, it's going to do it ten times better than it did six months ago. So the joke is, you know, we all thought robots were going to come in and do the dishes and drive the cars, and we were going to sit and write poetry. And instead, robots are writing poetry, and we're doing dishes and washing the cars. Like, what the heck kind of world do we live in? And so that's it. It was artificial creativity. It was the things that we thought were off limits for technology. Technology came for first.
(Joel Beasley at 00:06:31) It's got an advantage for distribution too, because we all already have the platform. Like, when you build a robot, you have to physically manufacture it and ship it.
(Brian Alvey at 00:06:41) There's a way to look at it. Yeah.
(Joel Beasley at 00:06:42) Yeah.
(Brian Alvey at 00:06:42) Correct. There's a way to look at it where, in the '30s and '40s, if you look at the Computer History Museum and go back in time to what was built then, when they were deciding how computers would work, John von Neumann, the inventor of game theory, said, you know, people were like, "Oh, we should store things in binary because that makes sense hardware-wise. It's hard to have five levels, but you can have an on or an off. That's easy and clear, right? So we'll do things in binary." And they also said, "We're going to basically build computers that work like smart calculators." They knew about neural networks. They knew about these more complex, fuzzy things, but the hardware was decades or a century away from catching up. So we basically spent 60, 70, 80 years training people to work in a browser on laptops and moving things and digitizing things for sort of the dumb computers.
(Brian Alvey at 00:07:34) But because we did that, we now have the infrastructure, the framework that was like charging stations-type things like Teslas need, right? We have all that ready so that now that neural networks are here, everything's already digitized. They can just absorb them and go. So the rate of transformation is going to be ridiculous. If we hadn't laid all that foundation for 80 years for the bots, this wouldn't matter, right? But because we have, and everything's already digitized and all these jobs have gone to browsers and laptops, AI can come in and just say, "I'm sorry, can you move aside so I can just do your job a thousand times better than you can?" And it's pretty—that's really frightening to people who thought that they brought something to the party, that they had a unique skill that couldn't be replaced.
(Brian Alvey at 00:08:15) And then the other thing that it turns out, the only thing that people really had was context. I just know where to go, what tool to grab, what thing to do next, what things people like. If you can document all that context, you don't have a job anymore.
(Joel Beasley at 00:08:30) I like—my favorite thing is the people who still hold on to "the AI won't do that." Because it's fear. They don't even want to look at it. It's like sticking your head in the sand.
(Brian Alvey at 00:08:44) But, "I have that divine spark, and a robot will never have that." And I'm pretty sure an AI can approximate whatever output you have from that divine spark. They can do a really good job of that.
(Joel Beasley at 00:08:56) Well, if you don't think the AIs can—whether—all right, here's where people get weird. I think having the spark and being able to emulate the spark are fairly equal in my ability to navigate the container that I live in, right? So because I'm just observing it. And so I can interact with an AI that seems to have just a personality and a spark just the same as when I'm interacting with a human. And to me, that is fascinating.
(Brian Alvey at 00:09:25) And it's all about resolution, right? So a long time ago, the voice of a robot, the answers they gave were kind of rough and crude. And now it's smooth and perfect. And I remember—I mean, I'm old, so I remember dot matrix printers where that was kind of the best you had, and nothing looked good. It all looked kind of terrible. But then the LaserWriter came out, laser printers where it was 300 dots per inch or 1,200 dots per inch, and the dots were so small, you just couldn't see the dots anymore. And the resolution became so magical and amazing that you're like, "This is a perfect circle. This is a—" and you can zoom in. I mean, there's still dots. You just can't—as a human, you can't see that resolution. So now you have bots where in the '80s, you know, "Shall we play a game?" You know, those kinds of computers, those voices, they were all low resolution. It is very easy to tell what's digital, what's a bot. And now the resolution—I mean, you'll get an AI calling you on your phone that says it's a family member, and it's believable.
(Brian Alvey at 00:10:23) So the resolution is just now imperceptible for humans. Although my kids would say they can spot AI a mile away.
(Joel Beasley at 00:10:30) Okay, so when you're not Brian WordPress Brian, and you're out at a kid's soccer game or just out in public making friends and people who are scared about this, what do you tell them? Do you tell them, "Yeah, you should be scared"?
(Brian Alvey at 00:10:44) Well, yeah, you can't lead with that, right? Because I would say it's 80% threat, 20% opportunity, right? Which is not—nobody wants to hear that. But you just—what you have to do is just say, "Look, this is just a natural evolution, right? All our life is shortcuts. Oh, there's a cool graphic I made about how to make a sandwich." And it looked at it from across the decades. And a long time ago to make this—it's all, my point is it's all shortcuts. So to make a sandwich, you know, 200 years ago, you had to make your own wheat and then get a mill and grind it all and do all this stuff to put things together to make a ham and cheese sandwich. And then now you just order it on your phone. But in between that, you had to invent refrigerators, electricity, stores where you could go up and down the aisles. You didn't use to shop yourself. You used to go there and tell them what you wanted. They would hand it to you. So each decade, the ability to make a sandwich went from weeks to days to hours to minutes to, "I can just say something to my phone and a sandwich appears at my door," right?
(Brian Alvey at 00:11:44) And AI is just that. So you're living with so many shortcuts that you just don't think of as shortcuts—how you eat, what you do—that this is just another shortcut. And I think it's weird to fear shortcuts, and I think it's healthy to embrace them.
(Joel Beasley at 00:12:02) Do you think there's been a rate of change in the arrival of shortcuts?
(Brian Alvey at 00:12:08) 100%. I mean, it's—even acceleration. So velocity, yes. Acceleration, probably too, right? You know, things have really changed in a crazy way. But it's just—it's going to get faster because six months from now, the things that models were slow at—I mean, I think somebody said to me the other day that Bitcoin cryptography can be cracked in seven seconds with quantum computing now by Google, right? The point was it was supposed to take you 200 years, therefore it was safe. And then a few years later, they're like, "It might be 40 years." And then a couple years later, "We could be a couple months." And now they're like, "Yeah, we got it down to seven seconds." So everything only accelerates, but I think shortcuts are healthy.
(Joel Beasley at 00:12:52) Yeah. And there's a little bit of nuance to that. I don't want to get derailed into the Bitcoin thing, but I think what they came away with was to update the algorithms by 2030.
(Brian Alvey at 00:13:05) Mhmm.
(Joel Beasley at 00:13:05) I think that was the—
(Brian Alvey at 00:13:06) You have to move the chains. You have to find the thing and—yeah, yeah. There's a way. Yeah.
(Joel Beasley at 00:13:11) What do you think the next wave of LLMs looks like?
(Brian Alvey at 00:13:14) That's funny. I think about this a lot. In science, there are kind of levels of resolution, right? You have quantum mechanics, and you zoom out, you get chemistry, and you zoom out, you get biology and sociology. And you keep zooming out, and you have astronomy. I think that people will just keep stacking, right? So the first thing was you go to a chat window, and you say something to ChatGPT. The next one was, "Oh, it's a more sort of on-your-laptop interactive thing," Claude Code, Cursor, things like that, right? And then they're like, "Oh, wait. You know what's better than that? A never-ending running thing. You don't have to ask it. It just runs on its own." So they're just walking through the motions of automating and the derivatives and all that. And so the next one is, "Wait, so once it's running on an agent harness and I have OpenClaw and I have one of these never-ending tasks on repeat, you know what's better than that? Oh, sub-agents." So one agent controlling ten agents. "You know what's better than ten? A thousand." They're just going to keep going like that. So I think it just stacks to where you just automate more and more and more to where you never actually think again of the weird old days where we used to go and ask one thing of an LLM and wait for it to respond. People are going to forget what that was like. They, you know, the magic of the company—
(Brian Alvey at 00:14:22) Correct. They're just like, "I remember the sounds. I remember the—" Nope. I don't really remember it. My grandfather probably did that. It's like punch cards, right? And here we are.
(Brian Alvey at 00:14:31) So I just think they're just gonna keep automating, exploding what you have and finding new things that you can do because you're exploding them. For instance, if I have five agents, but I give them five different personalities, I can approximate kind of a jury of decision making. You know, one person is more practical. One person is more a dreamer, right?
(Brian Alvey at 00:14:53) Bots again. But the fact that you can do that and put those personalities in there just lets you approximate the results of what we would have done with a team of humans with different skills and different diverse backgrounds.
(Joel Beasley at 00:15:07) I've been using the MyClaude thing and building this. So I do on nights and weekends, I tour with a national headlining comedian. And so his audience is primarily people aged 60. Like, his average age is 60.
(Joel Beasley at 00:15:25) My audience, if I do a show, my average age is like 40, 35, 40.
(Brian Alvey at 00:15:30) Sure. CTOs. Yeah. We're old.
(Joel Beasley at 00:15:32) Yeah. Yeah. And so for this comedy tour that we do, I'm always interested. Like, I'll write a joke, and when I do it to the 30 or 40s, I'll get a laugh. And then I was doing it to the 60, 65 year olds, and I was getting a groan.
(Joel Beasley at 00:15:50) And I was trying to understand. So I took all my jokes, and I ran them through, and I made an agent that was 60, 65 years old and all of their things that they have going on with that. I made one that was in their 20s. I made one that was like 30, 35. And I said, tell me how you interpret all these jokes.
(Joel Beasley at 00:16:07) And it was really illuminating to see how they could go in there and embody these different perspectives and then give me how they interpreted it.
(Brian Alvey at 00:16:18) Well, I'm glad you got valuable results out of it that you could learn from, kind of pin new beliefs to, because I also know people will do a thing where if they have to go on stage or they're about to do a presentation or they're about to try to argue something or sell something, have a bot kind of counter argue.
(Joel Beasley at 00:16:37) Uh-huh.
(Brian Alvey at 00:16:37) And they say, like, take the other position and just argue this with me. And but what people find is that whole sycophantic thing where, like, whatever you're asking it to do, it just gets really good at that. So I don't know. I haven't done enough of this myself, but I wonder about that. You know, could I take this idea of mine and battle test it in both directions? And do you actually get a good answer out? Does one ever win, or are the bots both so good at arguing you just walk away going, I don't know, I can't pick?
(Joel Beasley at 00:17:08) No. I use this all the time. In my personal life, I mean, my wife and I, we're talking to some sort of AI on a daily basis. And when we make decisions, whether it's regarding our kids or our finances, making choices, we have these saved prompts that we share with each other in our notes about how we can set it up to, because I'll find out, like, oh, if you prompt it like this, you're gonna get a better result.
(Joel Beasley at 00:17:34) And so that's what we're doing as husband and wife, but we will do that exactly what you said, have it take both sides and figure out which decision we wanna make by having it argue both sides. And we just do it one on one. We don't have two bots argue each other and watch it. We just say, hey, this is my thinking. Show me if there's any gaps or what the counterargument would be.
(Brian Alvey at 00:17:57) Right. That's fascinating. I wonder, because we use it a lot. My wife uses it. I use it in work, but my wife uses it more casually like a consumer would, right? Go to ChatGPT, have it rewrite an email, do things like that. But our kids are very much, I don't know if it's the school system, very anti-AI. And they're like, mom, you use ChatGPT. Like, I don't wanna, and I'm done listening to you. You know? Like, it's like, but it's helpful, but it's very funny to see the younger generation's reaction to this stuff.
(Joel Beasley at 00:18:26) Have they called you clankers yet?
(Brian Alvey at 00:18:29) No. Haven't heard that one.
(Joel Beasley at 00:18:30) That's a derogatory term for old people.
(Brian Alvey at 00:18:32) Idiot aroma, probably. Yeah. Yeah. Yeah. They love Idiot aroma. So maybe they like the term too much to waste it on us.
(Joel Beasley at 00:18:40) The old rusty robot. Yeah. Yeah. That's fun. Wow. Yeah. It's interesting. I wasn't expecting age ranges of your kids.
(Brian Alvey at 00:18:48) So older, 17 to 23, one's about to go into college, and the oldest one is done with college. But they all, like, I'll say something, then they'll be like, uh, that AI thing you did wasted six bottles of water. You know, like, every query, you know, waste. They look at it for two different reasons.
(Joel Beasley at 00:19:03) Are you serious? You got it. No.
(Brian Alvey at 00:19:04) That's what they say. I'm like, oh my god. And so I have to tell them, like, what regular data centers are versus AI data centers, GPUs versus CPUs. And the other thing too is just a general disgust for how it's ripping off creators. So maybe I don't know. Maybe I was too nice to the kids, and they really, they love art. They love artists. They wanna be artists, and they really do see this as just a massive threat to art careers, creative careers. And it is. But, you know, I mean, I wanted to have an art career and I ended up doing technology stuff.
(Brian Alvey at 00:19:34) Like, you can adapt and run with it. You can't fight against the biggest change ever, right?
(Joel Beasley at 00:19:39) It's funny. They're complaining about the AI responses and the six bottles of water while they're also simultaneously streaming Netflix all afternoon.
(Brian Alvey at 00:19:48) Correct. Correct. Exactly. And I tell, like, you wouldn't write an email without a spell checker. You know, like, these are just enablers, shortcuts. They make you bionic. Just give in.
(Joel Beasley at 00:20:00) Alright. So you're ready to get the implant? If Musk had an implant, would you go that route, or would you wait?
(Brian Alvey at 00:20:08) No. So, yeah. No. I, that's, I think it's a different question. I think it's funny. So Musk is just specifically such a lightning rod. Like, I know people who had a Tesla, but they were sick of every time they drove somewhere being constantly aware of who owns the company that made their car. And they just didn't have, it was, like, mentally fatiguing, right? Just have to think about what's he saying on Twitter, whatever the thing is, right? So Musk is the one. I would probably do it. I don't know what brand I would choose, but there might be brands I trust more than others.
(Joel Beasley at 00:20:38) So in your social circles, Musk is a lightning rod?
(Brian Alvey at 00:20:42) So, no. I've, people who are like, really good friends with him. Like, good friends of mine are good friends with him and, you know, do things with him all the time. So no. I know people who love him, but there's a lot of people who hate him, and it's pretty crazy. So think about that. Like, if Apple said we have a chip for you and Facebook said we have a chip, you'd look at those two offers, and you'd assess them differently. You might say, you know, Apple, they make good quality products. Probably not gonna—
(Joel Beasley at 00:21:06) Apple doesn't need the chip.
(Brian Alvey at 00:21:07) Yeah. That's my point, right? So a lot of people would look at a Neuralink Elon Musk chip—
(Joel Beasley at 00:21:14) Mhmm.
(Brian Alvey at 00:21:14) As more like a Facebook product than an Apple product. And that's all I'm saying. So I'm totally up for, I'm fair game to whatever kind of bio enhancing thing you got for me, but I will probably be picky about the brand.
(Joel Beasley at 00:21:26) Okay. Well, let's make it brand agnostic. The technology exists. Chip exists. What is the requirement for you to see in society before you get the chip implanted in your brain?
(Brian Alvey at 00:21:38) Yeah. I wanna see the results. I mean, well, think of the Ozempics and stuff like that right now, right? Like, people are making these decisions based on the thing that they're gonna kind of probably have to live with for the rest of their life, right? And I've talked to, just park the Neuralink thing and just go for example for a second. And I was talking with a friend of mine. And if you're, like, if you're a teenager, you're kind of committing to a thing for a long time. It's like getting a tattoo when you're 12.
(Brian Alvey at 00:22:03) Right? You should probably think about it. But if you're 60 and getting a tattoo, like, kind of who cares? Right? Like, do what you want to do your body. So there's a bit of that, like, which part of the hill or which side of the hill are you on? And I was telling him, like, you're gonna have to do this forever. And he's like, dude, if I don't do this, I'm gonna have diabetes. I'm gonna be on diabetes medicine. I'm gonna lose a foot.
(Brian Alvey at 00:22:23) Like, all these real world things, I don't think that having to be on Ozempic for the rest of my life or finding out that there was something that's gonna crop up later that we just hadn't tested enough for, like, I'm old. Like, I'm just gonna do this thing. So I think the same thing on the Neuralink. If you're over a certain age, I mean, what's, you know, it's gonna give you cancer in 20 years. You're gonna have cancer in 20 years already. So, like, you know what I mean? Like, I mean, it's a very, like, probably a morbid calculation, but I just think I'm not that worried about that kind of thing. Like, I would get a tattoo in a heartbeat now, or I wouldn't when I was 12 or 18 or something.
(Joel Beasley at 00:22:57) Oh, that's fun. I like the way you process data, my friend. You're fun.
(Brian Alvey at 00:23:03) It's all analogies. That's all it is. And if you can, so, I mean, thank you. That's a really interesting compliment. I like it. But I would say, I use mental maps and pictures to store things and think about things and then convey them to you. So I think something like a business model canvas is, like, gold. I think that's so cool because I can put an idea on paper. Visually, it doesn't map. You understand up, down, left, right, big, small, bold, not bold.
(Brian Alvey at 00:23:30) And I can convey something to you. And so I just find whether it's my kids, whether it's customers, whether it's my team, all I do is talk in analogies. So it's hard to make a decision on an implanted chip without comparing it to tattoos and Ozempic.
(Joel Beasley at 00:23:47) That makes sense. And I learned something new about Ozempic. I didn't know it was something you would use long term.
(Brian Alvey at 00:23:52) Oh, people thought it was a pill. It turns out, like, it's like an injection between your toes. It's like, it's a, you're, yeah. I know. You're committing to something pretty crazy. You can't just, like, buy it on the street or at least it was in the early days, right? And it's pretty nuts, but I mean, what heavy TV show, comedian, whatever, isn't, like, really looking slim these days. They're all on it. It's pretty crazy.
(Joel Beasley at 00:24:17) As we start to get to the back part of this interview, I wanna know what your thoughts are on what being a modern CTO means today.
(Brian Alvey at 00:24:26) That's a very good question because I would think in 2025, 2015, totally different answer. I would say five years ago, being a modern CTO was having a team of 50 or 100 and adding 20% a year and having a fleet of engineers and a whole bunch of different people working on things for you. And I just think that's all gone out the door, not just in the last four years of the LLMs, but since December with the really good ones. So I am watching people who older than me, younger than me, both kinds, who are just so red pilled, Claude pilled. And they're so deep into this, and they're like not that they don't need designers anymore, not that they don't need all the other things you need, but they just need it in different doses.
(Brian Alvey at 00:25:14) And they're all going like, I'm back, baby. I had retired from coding. I was in management, but now I'm coding again. And I had a taste of this with React and React Native six or seven, eight years ago where I was working on, I was like, oh, I'm an app developer again. You know, I'd built stuff in the '90s. I built stuff in the 2000s. And then I started running teams. And so you fast forward, you know, what would that be? Fifteen, 20 years. And like, I kind of took a long break, but then React Native for making mobile apps is kind of just like the JavaScript and CSS and—
(Joel Beasley at 00:25:46) Mhmm.
(Brian Alvey at 00:25:46) DOM type stuff of the web in the '90s. So I was an app developer. It was really cool. That was just a taste. That was one app that we worked on for six years, right? We just rebuilt that app in four days in Tokyo two months ago. I took two people and we rebuilt like a six year company in four days, which is insane. And we did it better. And it's now a WordPress plugin, this like social video thing we had, and then a native Mac video rendering app that lives in the menu bar that works with your WordPress site.
(Brian Alvey at 00:26:17) And like, we couldn't have, it would have been another six years to try to build all that stuff before. So I think that it's pretty remarkable how the team you need as a modern CTO is you don't need a team. I think that's the biggest change. And I feel like if you and I did this interview a year ago, I don't know that I knew that was coming. And now that we're talking in 2026, you just don't need a 100 person team anymore. If you have a 100 person team, you're spending your time managing them. So I think you just need smaller teams and you could do more. Anyway, I think that's, I don't know if that answers your question about modern CTO, but like, that's a big transformation, right?
(Joel Beasley at 00:26:56) I do track with you. Like, if, let's use round numbers. If you have a team of 100, but now you have this new efficiency, you have two options. The first option is to reduce the total count, right, so that you can just have a small condensed team to achieve the outcomes because there's a cost to taking the outcome and having it communicated to 100 people—
(Brian Alvey at 00:27:22) Mhmm.
(Joel Beasley at 00:27:23) Especially when you don't need that.
(Brian Alvey at 00:27:25) In the exploding network number.
(Joel Beasley at 00:27:26) Yes. It's a wild thing. Or you can keep the bodies, keep the people and reallocate them to other investable projects. And I saw Dorsey, Jack, he's like, I'm just, we're just gonna make the cut and make it really efficient. I tend to lean towards that model.
(Joel Beasley at 00:27:46) Like, if we can cut, like, let's make the cut. Get really dialed in on what we're doing because there's always the post cut shakeup, and then invest in new projects and hire more teams to do, like, figure out what one team can really do today. And then as we do new projects, we can bring on more teams and bring people back in.
(Brian Alvey at 00:28:06) I think that's the only way you can go mathematically because the effect that I've seen is that everybody says, I've got so many side projects. I have so many ideas. I have so many apps I wanna build. I have like 30 ideas. And I can never get to them. I can get to like two or three per year, right? But then once you can build an app a day, you find out you didn't have 30 ideas. And certainly, your company doesn't have 30 good product ideas, right?
(Brian Alvey at 00:28:31) You have five. And before you could get two out a year, well, now you can get all five done. There are no other ideas, right? There's, that's kind of the limiting factor.
(Brian Alvey at 00:28:41) So I think that you're right to say, like, what if we just had 10 or 20 people? And that's still probably too many, but let's start there. Let's get done the things we need to get done. Let's get done those stretch goals, and then let's find out that actually we didn't have the long backlog we thought we had. A couple new things creep up.
(Brian Alvey at 00:28:57) So I think you're going to settle into a rhythm, a size that is more appropriate to the speed at which you can move now.
(Joel Beasley at 00:29:04) Yeah. And I love the new skill sets that are cropping up. The skill sets, everything from prompting to, you know, how to delegate the ratio of human attention to AI clusters, I guess, for lack of a better way to explain it. Because the AIs have to run, but there has to be some human responsible for monitoring that the outputs are what we expect them as far as an organization. Right? And so you're going to have a lot of these AI babysitters, if you will.
(Brian Alvey at 00:29:37) Yeah. It's like the person, you know, in a winery or in a wine factory, whatever. I don't know what a wine factory is called. But where they make wine, like, you're spot testing things. You can't test everything, but you test. So, yeah, the example, the funny example I heard was, you know, FedEx trucks and things like that that are driving through your neighborhood. There is so much automated in that that the person, like, at some point, the trucks are going to drive themselves. A robot's going to deliver things to your door. Everything's scanned. Everything's inventoried. Everything's on a GPS. And there's no driver choosing where to go next. They're told these are the 11 stops, you know, and your package is at Stop Number 9, and they're two stops away. And so at some point, that person isn't in control. They're not in charge. They are there for the responsibility of if the car gets into an accident, like they had the driver's license or whatever, or they're there to just, I don't know, keep the thing warm. It's pretty weird. Or, you know, because people are freaked out about self-driving trucks and robots delivering packages. So it's kind of depressing, kind of strange to think about, like, what the human, you know, keep a human in the loop, but the human's just sitting there playing Candy Crush while the truck drives itself. It's pretty weird.
(Joel Beasley at 00:30:44) Well, luckily, there's all types of people because there are people, like, to me, that's hell. Like, the idea that I'm going to drive this truck and every, and I, no. I need to have, like, autonomy. I need to have creative, I need to be solving new problems. Like, that to me is what being alive is like.
(Brian Alvey at 00:30:57) Mm-hmm.
(Joel Beasley at 00:30:58) Like, that to me is what being alive is like.
(Brian Alvey at 00:31:00) But I—
(Joel Beasley at 00:31:00) I know there's tons of people that want to sit on the couch and play Candy Crush.
(Brian Alvey at 00:31:05) Or they're studying for the LSAT as they're driving by, or they're learning Duolingo languages. Right? Like, there's things you can do because you're not driving anymore. But, yeah, I go out of my mind. Yeah.
(Joel Beasley at 00:31:14) I know. I think learning languages is a thing that's going to get crushed by the AI too. We're all just going to be able to talk right through the apps.
(Brian Alvey at 00:31:23) Yeah. It's, uh, that's sad because they talk about, like, brains atrophying and the decisions of things we used to do. I mean, Americans are terrible at languages already and don't travel enough. Right? And, you know, anybody in Europe knows three or five languages. I mean, so, and, like, no offense to Americans. But, like, the team I work with is all over the planet, and I find it fascinating that there is this one guy, Lucas, in Germany. And, like, we're clearly his second language. Or there are people in India where we're, like, their fifth language. And I just think it's, I think how much, like, I think I'm a smart person. Like, I've done well and, like, people think I'm smart. But how much, how many times smarter than me are these people I work with who have, like, if I had to live in Japan and do my job as my first language in Japanese, like, I don't know enough. Like, I could learn it, but, like, could I give a speech? Could I do a comedy routine in Japanese? Could I do any of these things? It's really hard. And to watch people around me do this all day long in their second or third or fifth language just makes me feel kind of dumb, and it makes me worry about, I don't want to give that up. You know? Like, I think that's an amazing skill. I think it's a good use of brain muscles.
(Joel Beasley at 00:32:30) It is interesting. I mean, we in America, we've got this landmass per language ratio that's different. Like, if you go over into Europe and other parts of the world, landmass per language, I mean, you have to develop this skill. Here, we can travel all over the country and, yeah, with the exception of a couple different dialects that you'll run into and the South, you can pretty much understand our building.
(Brian Alvey at 00:33:00) I mean, there's plenty. Yeah. Exactly. Boston. Right? It's funny.
(Joel Beasley at 00:33:03) Yeah. Yeah. We're great at dialects. They're great at entire languages. Well, look. Thank you so much for doing this, man. We made a podcast. How do you feel?
(Brian Alvey at 00:33:12) I feel great.
(Joel Beasley at 00:33:14) Thank you so much for listening. And if you found this episode useful, please share it with a friend or colleague who you think would get value from it. And if you have topics that you'd like to hear discussed on the podcast, either add me on LinkedIn or send me an email, [email protected]. Every time I get an email or LinkedIn message, it absolutely makes my day and inspires me to keep going.