Jesse Friedman: Welcome to Impressive Hosting, a podcast about the role hosting plays in shaping the open web. I’m your host, Jesse Friedman. On this show, we go deeper than uptimes and dashboards. We talk about hosting as infrastructure, about ownership, independence, and what it takes to build ethical, high-end WordPress hosting that actually serves creators, businesses, and the internet itself. Before we dive in, head to impressive.host. That’s where you can comment on episodes, ask follow-up questions, and help shape future conversations. You’ll also find links to follow, like, and subscribe wherever you listen. Today, I’m joined by Nick Hamze, probably one of the most prolific designers, developers, and plugin creators out there. He’s drawing the weirdness of the old web back into the new web, doing a lot of fun experiments, and launching tons of new stuff all the time. He advises Automattic, helps advise the WordPress community, and he’s building really cool things. I’m really happy to have him on the show today. Nick, tell us a little about yourself and what you’re working on.

Nick Hamze: Or what am I not working on?

Jesse Friedman: That’s the point, right?

Nick Hamze: There’s just so many things. I always had ideas, but now with all the models and whatever, there’s seriously nothing I can’t do as long as I have the tokens for it. It’s blowing my mind. I’m working on a project with Rich Tabor, and he always asks me, “How’s it going doing this?” I’m like, “I have no clue.” I don’t really care. I’ll have a conversation with Cursor and it will do something, and then I’ll be like, “Great, cool. Let’s change this. Let’s do that.” It’s just blowing my mind. I’ve done a macOS app. I’ve done I don’t know how many web apps. I’ve done WordPress plugins, WordPress themes. As we’re talking right now, I built something that makes WooCommerce themes without me even being there. It just runs. I call it the theme factory. The goal was to make 50 themes in a month. I think I can make 500. It’s blowing my mind how easy it is once you build this little machine. It just chugs away while you’re doing a podcast or whatever.

Jesse Friedman: Well, it’s definitely true about AI. It’s like driving a car. You don’t have to know how a combustion engine works to drive a car. With technology, maybe we’re a little nervous about the lack of understanding, but it’s okay to embrace it. Is that what you’re saying?

Nick Hamze: Oh, 100%. I don’t have any other options. If I had to understand it all, I mean, I have to do it this way, and I’m totally fine with it. I had a conversation with Claude last night and it said, “Hey, we really have to put these guardrails in. You can’t just give your AI total control over the software.” I was like, “Yes, I can.” I had to explicitly tell it, “You don’t have to ask me. You can just do whatever, because I trust you. You’re smarter than me at this kind of stuff. It’s not like I know any better than you do, so I’m cool with it.”

Jesse Friedman: So you’re not worried about it completely trashing your setup or going completely nuts?

Nick Hamze: Nah.

Jesse Friedman: Yeah.

Nick Hamze: It’s the one that wrote it, so if it trashes it, it can rewrite it.

Jesse Friedman: That’s an interesting point. Every time I ask Claude to write me a plugin or some code, and then I ask it to modify it, it doesn’t just go in and edit a few things. It rewrites the whole sheet of code from scratch every single time.

Nick Hamze: Yeah, unless you tell it not to. That’s the thing. When I was paying for my own credits, I made sure to tell it every five minutes, “Be very judicious about your tokens.” But then I found out we could expense credits, and I was like, “Okay, now I can be a little more adventurous.” Because that’s what people don’t understand. AI is cool, but it is expensive when you’re doing it at scale.

Jesse Friedman: Oh yeah. When someone else is paying for it, it kind of unlocks things. But right now, the biggest bottleneck to doing some crazy stuff is the money we have to put into it.

Nick Hamze: Yeah. But the fun thing is when you think about what it costs, this project I’m working on now maybe costs me $10 to generate a theme. That might seem like a lot, but it would cost me $10,000 if someone made that thing by hand. So I’m cool with it. You just gotta keep it in perspective that it’s always cheaper in software development to have AI do it than to have someone else do it.

Jesse Friedman: Yeah. I’m actually at Automattic’s NoHo office lounge in New York City, and I was just talking to some colleagues in the hallway about how I’ve tried probably every to-do app on the planet and I’ve never found one I liked that was made for me. A couple years ago, if I wanted a custom app, I’d have had to hire someone and spend thousands of dollars, if not significantly more, to get something built. But now I guess I’m gonna become an Android developer because I have Claude at my helm.

Nick Hamze: 100%.

Jesse Friedman: I mean, yeah, it’s fun. It’s empowering. It’s also a little… I’m a little hesitant to think we can just throw the reins and let it run itself. But everyone I talk to gives me these precautions about AI, all these things to worry about. We’ve been in this conversation for two minutes and you’re already telling me you have things running in the background, building stuff, going without a care in the world.

Nick Hamze: Oh yeah. And they’re telling that to the wrong people. They shouldn’t be telling you that. Whenever I post something on X and someone’s like, “Oh, AI is the devil,” I know I shouldn’t do this, but I’m just gonna block them because I have no desire to debate it. I’m all in. There’s no aspect of my life I won’t let AI into. AI has totally changed the way I do everything. I feel like for the first time, after spending the last decade after I left Automattic being frustrated because I could never find developers to hire, like all the things that stopped me from being able to do anything in WordPress are gone. There’s literally nothing I can’t do.

Jesse Friedman: I mean, that’s amazing. We talk about this all the time too. I feel like AI has given me the ability to be the developer I always wanted to be, because I was never able to dedicate my whole life to computer science and writing code. I was always busy doing a million things at once. You’re kind of the same way. You’re always working on something else. But I’ve watched what you’ve been building and designing and doing for a long time. I think it’s very accurate to say you’re an artist. You have artistic integrity. So where do you draw the line with AI? Are you using AI to do some of the things you’ve done in the past on your own?

Nick Hamze: So I’ve never really had the ability to design. I can’t draw or whatever, but I’ve always known what looks good. It’s gotten much easier to tell an AI my taste and have it run with that. Some people say, “Oh, I can code, so I’m not gonna have AI code for me.” But even with things I’m actually good at, like I think I’m a fairly good writer, I have no problem letting Claude rewrite my stuff because it just makes everything so much easier. Now I know I should’ve figured this out earlier, but I always bring my laptop to bed because I always have an idea. Trying to type in the dark never works that well. But I figured out there’s a little button where I can talk to it. So now I’m sitting there at 3:00 in the morning saying, “Claude, run this test. Do that thing.” It’s so freaking cool. If someone built a little robot that followed me around and read out my Claude messages, I would buy it in a heartbeat.

Jesse Friedman: Like those delivery bots rolling around San Francisco? Put an AI brain in that?

Nick Hamze: Yeah. I remember this had to have been 12 or 14 years ago at the Hawthorne office of Automattic. They had one of those things, basically a Roomba with an iPad on it. I thought it was so cool. I would love to have something like that that could just sit next to me and verbalize things. Because again, all I’m doing is talking to Claude. If I could just do that verbally as I’m sitting there and say, “Hey, run that test again,” or whatever. And it’s funny, I used to have to go and tell it a lot, but now I just tell it to fix things itself. Last night I had all these gates it was checking themes against, and when it failed one, it would ask, “What should I do?” Now I just gave it a Claude API key and said, “If you come up with a problem you don’t know how to fix, just ask the AI.” So now I have this self-healing loop that I never have to touch. Literally as we sit right now, it’s building theme after theme after theme, and every time it finds a problem, it asks the AI what to do, and the AI makes a rule so it doesn’t happen again. I just can’t wait to see how good these themes are by the end of the month.

Jesse Friedman: All right, so you’re having it churn out these themes. What’s the next step? It builds the themes, then what? Are you putting them on the repo? Building a library? What’s going on?

Nick Hamze: So this is all for woocommerce.com. They don’t know it yet, but me and Rich have been looking a lot at WooCommerce. You go on the marketplace and they have literally two free themes. Storefront from way long ago, and then something else. Guys, even Shopify, which doesn’t have a lot of free ones, still has 30 or 40. And Shopify has 1,400 to 1,900 paid themes. I never understood why we didn’t make our own themes. And once I figured out how easy this is, because when all is said and done and you look at these themes, they’re not just the same theme with a slightly different color. They’re radically different. They rearrange everything, and they’re built for niches that nobody would ever make a theme for if they had to do it by hand.

Jesse Friedman: So you hit the magic word there. Niche.

Nick Hamze: The magic word.

Jesse Friedman: That’s what I’ve been advising hosting companies for a while now. We’re still in the box of selling hosting, but a lot of our customers want a solution. They want success. They want to be able to say, “I want to launch my specific e-commerce retail store,” or “I want to launch my musician website,” or “I want to launch my lawyer website.” And we’re over here selling hosting. Do you see AI as a way to enable hosting companies to solve that problem?

Nick Hamze: 100%. If it was me running WordPress.com today, every single niche you could think of would have a totally AI-created landing page that goes to an AI-created funnel, that goes to an AI-built theme that’s already done. Basically, they could just change the theme name to whatever their law firm is called and they’d be good to go. Why do we make people hunt for a business theme and then show them one made for a web design agency? It’s nothing like their plumbing company. Why would they use that? People can’t see themselves in it. It’s like a staged house. If they can’t see themselves in it, they’re not buying it. And if it’s this trivial, I think I’ve figured out that when I get it down to a science, it’s gonna be about 10 minutes and $3 per theme. If I can take $3 and make a deal like that, especially too, what about support? Guys, there’s an agent for that. You can have an agent constantly testing against the newest versions of WordPress. You can have agents taking the questions people ask and improving things. I’m not worried about that at all. But it really bothers me, probably for as long as I’ve been around Automattic, like 20 years, that everything is a coffee shop. The number of people trying to make coffee shop websites is infinitesimally small. Why is that always our gold standard? Or if it’s a restaurant, it’s always fine dining with reservations. The people with a fine dining restaurant can hire somebody. It’s the food truck, the local pizzeria, the donut shop. Why don’t we make things that show them that we get them? That’s another fun thing about this project. It also makes the demo sites. I said whatever vibe made the theme, you have to go and generate all new imagery, all new descriptions, and all new microcopy. Why would you show a theme about lawyers with a blog post about something completely unrelated? There’s too much demo content that just doesn’t fit.

Jesse Friedman: It might as well be Lorem Ipsum, right?

Nick Hamze: Exactly. So why not have it generate the right content on the fly? It’s so trivial. The possibilities for customization are exciting. It could get to the point where you search for something and it generates a demo site on the fly. Or maybe it’s a cache thing where you say, “I’m looking for this,” and it automatically spins up a demo site combining the lawyer content it already generated with whatever you tell it. That’s another thing I really dislike about some of these site generators. They generate garbage you have to replace. What if it generated stuff so good and so close to what you wanted that you just left it? If I say, “Make me a lawyer site,” and it gives me a bookshelf with fancy books and everything looks right, maybe I’ll just keep it because it’s so good. It’s exciting. There are so many possibilities, and if people aren’t figuring this stuff out… That’s why I don’t know if your listeners know about the radical speed month. Is it RSM?

Jesse Friedman: Radical speed mode.

Nick Hamze: Radical speed mode. I never get it right. I think you’re gonna find out who has the AI mindset, and those are the people we need to put in charge going forward. Some people are still setting up their project while others are already done and moving on to the next one. We really need to try everything and fail quickly. How many times have we talked about an idea for eight months when we could’ve built it in two weeks and known whether it failed or succeeded? Why waste all that time arguing about whether something will work? Just try it. And I think that’s the best use case for AI.

Jesse Friedman: Well, I was just going to give real quick context to the people listening at home. Automattic is in the middle of this radical speed month right now. Matt basically came to us a few weeks ago and said we’re all gonna pair off and just go build something. You don’t have to worry about the bureaucracy or the red tape or getting approval or doing everything the right way. We’re all going to take some time to innovate, build something, and launch something. And that’s what you were mentioning before, that you’ve paired up with Rich to build these themes.

Nick Hamze: Yep.

Jesse Friedman: Right.

Nick Hamze: Yeah. It’s probably coincidence, but I was talking to Matt the day before about all the impediments I’d been facing in the company, and then he announces this. I was like, “Did he listen to me?” Probably not. But I love that instead of just saying, “Obey the chain of command,” it’s, “Okay, if you think there’s a better way to do it, let’s try it.” I felt like I was dreaming. Like all my birthdays he’d missed for the last 10 years all wrapped into one. And you’ve already seen some people that I had no clue existed, and they’re creating things that I think are gonna have long-lasting implications on the future of the company. I just think it’s so cool. How do you unlock that and then steer people in the same direction so they can build on each other’s work without losing their excitement? And how do you find ways to take something cool and make money from it? There’s just all sorts of fascinating things happening, and I can’t wait to see what it turns into.

Jesse Friedman: What’s interesting is that every time I have a conversation about AI, people’s biggest concern, their nervousness, is around the idea that AI is going to steal their ability to be unique or have self-expression. That it’s going to become another version of the bland, sterile internet we talk about here. But the way you’re talking about it, it empowers you to do so much, to try so many things, to experiment so quickly, that you can actually find your voice or your design faster, maybe even in a way you’ve never been able to before.

Nick Hamze: Yeah. People ask me, “We have access to the same tools. How do you make wacky websites and mine look boring?” The software has the ability to do whatever you want. You just have to ask it the right questions. I’ve been working a lot with Miles. Have you heard of Miles? It’s a WordPress site creator by Andy Peatling.

Jesse Friedman: No, I haven’t played with that one yet.

Nick Hamze: It is so cool. My favorite thing to do is it’ll give me something and I’ll say, “Make it weirder.” He was talking about putting a slider on the back end, calling it the Hamzy slider, where you can adjust the weirdness scale. A lot of people can come up with nice looking things, but if you ask it the right questions, or just talk to it openly, you can get somewhere interesting. I’ll say, “Pick colors that nobody would ever pick together,” or try different combinations. You can ask it about styles you didn’t even know existed. I found neo-brutalism through it. Oh my Lord, I love it. Nothing matches, it’s in your face, sharp edges, no border radius. I found that through AI. I wouldn’t have found that myself. If I had to hire a designer, most of the time you’re hiring someone because they’ve done something similar to what you like. But I couldn’t find a designer for this because I didn’t even know it existed. And that’s the problem with most WordPress themes too. It’s all the same minimal whatever. Not everybody likes that. I feel like if I really got down to your heart of hearts, I bet I could convince everybody there’s something else they might like a little better if they let themselves.

Jesse Friedman: Yeah. What’s really interesting is that DIYers, people without a lot of technical experience, they always come back to the fact that their time is their money. Building websites is not their passion. They’re a chef or a lawyer or something like that. Building a website is a means to an end. A lot of times what happens is they have no faith that the time investment they’re going to make will result in what they need. So this is flipping that on its head. You’re saying you don’t have to have faith in it because you can just keep experimenting until you find the parts you like. If you can communicate with the AI in a way that makes sense to it, you can take all the results from these experiments and actually build something that has value and ship it a lot faster. But I think the intimidation, the idea that they’ve never played with AI or don’t know how to do it, means they’ve never jumped in, and so they don’t know there’s actually a light at the end of the tunnel.

Nick Hamze: I think the biggest problem is that we make software intimidating. Like there’s value in onboarding and funnels and all that, but if the process is enjoyable, people will put up with a lot. If you make the process enjoyable, I think people would spend ten times longer in it. If it feels like talking to a friend, let’s say the chef you take for example, of course they try different recipes and different seasonings. If it was just like riffing with another person, even though that person’s a computer, I think it’d be much more fun. But when you have software that sounds like software, and the minute you get four queries in it asks you for money, that’s where it falls apart. I want to test whether having something that was fun would make people spend more time in it. We had a thing at work where they wanted everyone to make a website, and a large portion of people were like, “I just don’t want to.” Why don’t you want to? Even when your boss tells you to do it, you still don’t. Why is it so painful? And that’s why I want everyone to check out that desktop mode thing. I literally had a woman contact me about it, and it was like the best thing in the world. I think that would totally revolutionize how people feel about WordPress if it was fun.

Jesse Friedman: Right.

Nick Hamze: And not fun as in it has to be a game. Just little things. A little dancing something in the corner. Something that makes people laugh or smile. I had an idea for a plugin I really want to get into Jetpack, though it’ll probably never happen. It’s like your own hype man. A little character that’s always telling you, “You know what? This is great. You’re gonna do awesome. Publish this.” And it actually knows something about your site. It could read every post and say, “Hey, this is great. It reminds me of this.” Or if you wrote a post about your dog, it mentions your dog. I guarantee we could take a little bit of time to make people excited. We always have that thing where it says, “This is your 10th post.” Who cares? Give it a personality that talks like you. Ask where you’re from, match your speech patterns, your slang, your age. How fun would that be? We always want things to be professional, so it always sounds like a 50-year-old English professor. Why can’t it just be silly and goofy and fun? I just feel like fun is the missing piece of everything.

Jesse Friedman: No, I think you’re right. I used to have a developer friend who had a Batman action figure on his monitor, and every time he ran into a problem with his code or a brick wall, he would actually just have a conversation with Batman. In describing the problem verbally and expressing his frustrations, meanwhile Batman is just standing there with his grim face, he actually found his way through the problem. But this is a million times better because you can actually just talk to the thing about what you want to accomplish. You don’t have to have a means to an end. It will find the means to the end. You just need to describe what you want. So that leads us into the next phase. What are the tools you’re actually using? Are you using Claude Code? What are you actually engaged with?

Nick Hamze: I know all the cool kids switched to Codex or whatever. But I love Cursor. I’ve stuck with Cursor and I really enjoy it, especially because you can use whatever models you want. This week they added some new GPT 5-point-whatever, so I was like, “Okay, I’ll try it out and see if it works.” I usually stick with the Claude models because I find those work best.

Jesse Friedman: Yeah. Claude is kind of winning the whole race right now.

Nick Hamze: It seems more personable. The ChatGPT stuff seems very professional and formal. And the first thing I do when I set up my agents, I give them a personality. I tell them how I want them to talk. I don’t want them to talk like a computer. I want them to talk like my friend. My biggest failure at Automattic is that I cannot convince them to give AI tools names. It makes me so sad that they will not call something… They have some new tool they’re working on. It was supposed to be called Dolly, and now they made it the WordPress agent. Guys, people want Claude. People want something with a name. People want to refer to it like it’s a real person. They want to talk to someone. I just don’t understand it.

Jesse Friedman: It’s funny you say that because I’m experimenting with a WP Cloud tier-one service agent, and I didn’t call it the WP Cloud service whatever. It’s called Walt.

Nick Hamze: Love it.

Jesse Friedman: The avatar I used is Walt from Breaking Bad. It kind of goes back to my roots when I co-founded a company called Brew Protect. I think it’s because we did our first meetup in Albuquerque, where the show took place, and so Breaking Bad had this weird hidden theme in everything we were writing. We had this ASCII art of Walt in the HTML. If you went and viewed the source, you’d see him there.

Nick Hamze: So cool.

Jesse Friedman: Stuff like that carried through. And I have another one I was working on, more marketing-related, and she’s named Skyler. So I’m naming all my agents.

Nick Hamze: You and I are in total alignment there.

Jesse Friedman: Yeah.

Nick Hamze: It seems like a no-brainer. But again, it’s almost like this problem where everybody feels like they can’t be unique. Like that’s not professional. You have to use the same colors, the same fonts, the same everything. It drives me up a wall. Especially because even though Automattic is a big company, they’re also a small company in the tech world. Why not lean into that fact?

Jesse Friedman: Yeah.

Nick Hamze: My day job, we own a Pokemon card shop. Everything we do is about being the small guy. Don’t go to Walmart. We’re goofy. My text messages have pictures of my dog. We don’t wear uniforms. People like that. They like to buy from a person. They want to buy from Nick, not from XYZ Corporation. And I feel like that’s how the web started. Way back, I never had GeoCities, but I remember making weird FrontPage websites and fun little chat rooms. There was another project someone was talking about and they wanted a different staging domain. And I’m like, “Why do all of our things have to be so-and-so staging site?” Back in the day, they had all sorts of cool subdomains. It’d be like nick@flapjack.com. You could just pick whatever you wanted. Why can’t we go back to that? Why can’t I start a wordpress.com site but have it at sillygoose.org or something? I don’t even need to buy a custom domain, I just like what it sounds like.

Jesse Friedman: I think you’re right. When we got started on the web, it was basically an engine for free expression. You could do whatever you wanted. There were no rules. Nobody would get mad at you if you put your menu upside down. Then I remember these days where the science of colors was being discussed. You always want blue or green because it indexes trust. Then platforms started coming out of the woodwork. You got Instagram and Facebook and all these others. Back in the day with MySpace, you could literally make it anything you wanted. Music playing, things falling from the sky, things spinning. Then Facebook came along and it was so regimented. All you could do was upload a photo or change some text. And so much of the web became sterile just like that. You’re focused on the content, and I think that’s good. The video or the text is king. But there’s so much more to how you express yourself than just the pure content. Before this, I checked out your website, regionallyfamous.com. If you haven’t seen it, you need to go there because it’s such an amazing expression of how you want your business and your work to be seen. It’s so different. It’s not the sterile framework of black text on white background where everything looks exactly the same. And I think that’s absolutely what we’re losing with the current trajectory of the web.

Nick Hamze: Yeah. We always give people what we think they need, not what they want. Like, you’ve probably done a lot of these podcasts. Does anybody have such a chaotic mess behind them like I do? I love this. On the wall I’ve got Star Wars sketch cards, Pokemon stuffed animals, a Pokemon vending machine, wacky weird stuff. That’s how I like it. Whenever I see someone with a sterile white background because they want it to look clean and professional, I think, who cares? That’s what I want websites to look like. How the inside of your brain looks. How the inside of your house looks. And I feel like that’s what people want, but whenever they ask for that, we hand them a white box instead. But the first person to offer them something different, they’re all gonna go there in droves.

Jesse Friedman: I think that’s really great. When we think about it, how are people getting websites these days? They’re going to a platform provider that gives them a path toward launching a website. That could be a hosting company or a closed platform. But it’s all about onboarding, and the path that onboarding takes you down. The number one metric we talk to hosting companies about is time to live. Reduce the time to get your website live as fast as possible. I think a lot of times they conflate that with making the customer feel like they got something done, that they’ve reached their goal, met some level of success. But maybe what it’s really doing is just reducing churn because they got something live. Is that really what they wanted? Is it an expression of the restaurant they wanted to show off? Is it taking full advantage of the web to be something beyond sterile text? Maybe time to live isn’t the right metric. Maybe there’s another metric that better identifies whether people are actually being successful.

Nick Hamze: Yeah. I gotta come up with a fancy term for it, but for me it’s the time to fall in love. How do you get to the point where your site is so you that you would never consider getting rid of it? I’ve gotten to that point with Regionally Famous. I like the design and the domain so much that even if I gave up on WordPress stuff entirely, I’d probably keep it forever because I enjoy it. I think it’s funny, I think it’s cool. When you have something where it’s literally a beige box you care nothing about, it’s so easy to let it lapse. But with agents especially, I don’t understand why we’re not leaning into this. If I had an agent I really liked, with a name, and it came to me saying, “You have to either pay $10 a month or this guy ceases to exist,” I’d hand over the $10. I love that guy. I don’t want him to go away. Maybe you can’t put a metric on that. But I think you should just delight people from the first moment and they’re gonna stick with you because they want to. Not because they have to, or because you’ve got them locked in and canceling is too painful. I really enjoy not just the end result but the process. That’s why I keep talking about that desktop mode. I could say over half the people would find it more fun if they were doing the same stuff they always do, but the wallpaper is a picture of their dog or their favorite sports team. And especially when we’re fighting against social media where you can do absolutely nothing to customize it. If we’re trying to say you should own your data and build your own digital home, but you can’t do anything to it, then what’s the point of leaving Twitter or Facebook? Fun, personalization, quirky. That’s what we’re missing. And I know it’s what we’re missing. I just gotta find someone who will believe me.

Jesse Friedman: Well, hopefully we get a lot of people listening because your passion and energy are doing a lot of work in getting that message across, and I’m definitely feeling it too. A lot of the conversations around AI are driven by fear, anxiety, and pressure to use it. People feel like they don’t have permission to be weird with it. And I think you’re giving them that permission. That’s really great. We do have to take a break though. We’re at 35 minutes and time has flown by. But we’ll come back with another episode with Nick.