Follow the work as it happens.

I once fired my twin brother. Technically, twice. Now he's going back to Automattic, and I couldn't be more excited for him. Let me back up. A decade ago, Nick landed his first real job at Automattic. He did a little bit of everything, traveled all over the world, and fell in love with WordPress and the people building it. At the same time, we were graduating from law school and I was starting teelaunch. Nick's paycheck kept us both going through those early days. When teelaunch needed him, he jumped in with both feet. He left the job he loved, worked for free that first year, and helped build something we're both proud of. We rode every high and low together. Whenever things got tough, Nick was always the first to say, "Fire me. I'll be OK." So I did. Twice. (He still jokingly brings it up at every family dinner.) Through all of it, Matt Mullenweg kept in touch with my brother. They'd talk WordPress. They'd talk Tumblr. It meant a lot that someone like Matt saw what I've always seen in Nick. Now he's heading back to the company and the software he loves, and the timing couldn't be better. WordPress is entering one of the most exciting chapters in its history, and Nick is exactly the kind of person you want in the room for it. And no, Izzi's Gym isn't slowing down. We've built AI into every process, things are running smoothly, and we're growing fast. A guy with my brother's energy was never going to do just one thing. Matt, thank you for believing in Nick all these years and for welcoming him home. Nick, go build something amazing. Proud of you, brother.

You cannot sleep on this. Before I changed a single thing, my Shopify store had already made $918.70 from ChatGPT and the Shop app. No ads. No links. AI found our cards and sold them. So I had AI audit my store the way a shopping agent sees it. It was ugly. 534 products listed our store name as the brand. Not Pokémon. Not Magic. "Izzi's Gym." 528 used cards were marked "new" in our Facebook catalog. One-of-one PSA slabs. Marked new. Our refund policy was a clothing template. "Unworn, with tags." We sell Pokémon cards. And our website was blocking Meta's AI crawler. The same month Meta launched its Muse shopping agent and teamed up with Shopify on checkout. None of this shows up when a person looks at your site. It looks fine. The problem is the robots reading it. Here's what we fixed: → The real brand on every product → Used vs. new, marked right → Set, card number, year and PSA cert in every slab title → Alt text on every photo → FAQs AI can answer: authenticity, hours, condition, returns → Policies written for what we actually sell → Opened the door for AI crawlers Here's the part most store owners miss. Shopify already puts your products in front of ChatGPT, Google, Copilot and now Meta. You don't have to sign up. You're already in. The only question is whether the agent picks you. A one-of-one PSA 10 can't be price-shopped. The agent picks whoever describes it best. If you run a Shopify store, open Sales channels > Agentic today. Then ask an AI to audit your catalog like a shopping agent would. $918.70 was the floor. Let's find the ceiling.

Love this. One footnote from behind the counter at a card shop: this compares the entire venture index to the single best Charizard on earth. That's like comparing the S&P to Google's seed round. The average 1999 Charizard didn't become a PSA 10. It got shuffled, traded at recess, and stuck in a bike spoke. We see those every week. Still worth real money, just not $1MM money. The real asset was never the card. It was the kid who put it in a sleeve in 1999 and never touched it again. That kid was the VC.

100 people a day want to buy the Pokémon cards nobody wants. That's bulk. The commons and uncommons left over after the good cards get pulled. They're offering 1 to 2 cents a card. Here's my problem. We generate an enormous amount of bulk. Way more than any one guy can take. I'm worried the buyers I'm finding will max out fast. We've got a loading dock and everything we need to do bulk at real scale. What I don't know is who sits at the end of the chain. Who buys it from the guy who buys it from me? Mystery pack makers? Exporters? Teachers? Somebody turning it into wallpaper? I want to talk to that person. If you know who the end buyer is (or you ARE the end buyer), drop a comment or DM me. Loop me in.

I'm not sleeping tonight. And I don't want to. 15 years ago we started teelaunch. Teespring had just shown up. Facebook ads were brand new. And we thought "target by interest" was the most next-level thing we'd ever seen. You could put a shirt in front of every nurse who loves wine AND golden retrievers. We felt like wizards. Now Meta's Muse can check out from a Shopify store with Shop Pay. Right inside the agent. That's not targeting anymore. The bot actually figures out what you're looking for. Same company. Same feeling in my gut. Totally different game. Here's the kicker. The question isn't "how do I find the customer?" anymore. It's "how do I hand my product to the bot so it makes the most sense?" Clean product data. Real descriptions. Answers to the questions shoppers actually ask. The store the agent understands best wins. I've been burning AI credits like a 10-year-old ripping booster packs chasing a Charizard. Definitely not sleeping tonight. Who else is building for this?

Your next customer might never see your website. They'll ask an AI agent to find the product, and the agent will buy it for them. That's not a prediction anymore. Meta's new AI agent, Muse, can now check out from Shopify stores with one tap using Shop Pay. No redirect. No account creation. No abandoned cart. I run a Pokémon card shop in Sioux Falls, South Dakota. Here's what this looks like for a store like mine: Someone texts Muse on WhatsApp: "Find me a sealed booster box for my kid's birthday." Muse searches, finds us, and buys it. We ship it. Done. What every Shopify owner should know: Direct checkout is ON by default for eligible stores Meta charges no extra fees, just your normal payment processing US, Canada, and Mexico only for now Your catalog has to be clean. Vague titles, missing details, or incomplete return policies, and the agent skips you. That last point is the whole game. For 20 years, getting found online meant writing for Google. Now you're writing for a robot shopper. It doesn't care about your hero banner. It cares about accurate inventory, clear product data, and real policies. The small shops that clean up their catalogs this quarter will get found. The rest will wonder where the traffic went. We're getting Izzi's Gym set up TODAY. Are you ready for customers who never visit your homepage?

Is my job going to be replaced by AI? Absolutely. I’ve made peace with it. But after spending the last few months actually building agents, I’ve also concluded that replacing me will require approximately: 48 agents 6 orchestration agents 72 API connections 830 billion tokens and a Zapier workflow held together with gum. Plus one agent constantly fixing what all the other agents confidently broke. (That’s exactly what I’m doing right now.) So yes, AI can replace me. But it's probably cheaper to keep me around. 😅 Inspired by CMO Alliance post: https://lnkd.in/d2xJZQR9

Their CEO, Rob Bigler, is leaving. So I did what any reasonable card shop owner in Sioux Falls, South Dakota would do. I started updating my resume. Here it is: Started my first business at 8 with my twin brother. Candy company. It failed because I put the chocolate bars in a display case in direct sunlight. Co-founded teelaunch, a print-on-demand platform. Sold it. Built a business to over $135M in sales. Licensed attorney. MBA. Co-own Izzi's Gym, South Dakota's largest TCG store. We sell on TCGplayer every day. That last line is the one that matters. A lot of marketplace decisions get made by people who have never packed an order. If I got the keys, week one looks like this: Sit down with 50 sellers. The small ones, not just the top accounts. Fix adjust the fees to be more competitive with other platforms Gift Izzi's Gym's store management/listing software to TCG Player Educate our store owners on how AI can transform your life and your store. I'm literally working 20x more efficiently than I ever have, producing the best software and marketing I've even done in my life. Robert Bigler, thank you for the time you spent actually listening to sellers. Whoever comes next should keep that. eBay, my DMs are open. (Mostly kidding. Mostly.)

When I first started at Automattic, I decided to call myself the Chief Swag Officer. I liked swag. I thought it was funny. That was pretty much the entire qualification process. I think it’s time for another completely self-appointed title. CFO. Chief Fun Officer. I’m excited about where Automattic is going, and I want to bring my kind of weird to it. AI has made so many of the ideas rattling around in my head possible to actually build. Little apps. Parody sites. Things that make you click around just to see what happens. I want more of that on the web. More personality. More surprises. More things someone made because they thought, “Wouldn’t this be fun?” That’s the energy I want to bring to this next chapter. The title is a joke. The enthusiasm is very real.

Literally some of my favorite artwork in the world.

The personalized-products industry is struggling at a strange time: making a good personalized product has never been easier. In 2020, 1-800-Flowers bought PersonalizationMall for $245 million. Now it has agreed to sell a package including PersonalizationMall and Things Remembered to PlanetArt for approximately $45 million. The deals aren’t directly comparable, and the new sale hasn’t closed, but that’s a dramatic difference. Shutterfly has its own financing challenges. Bloomberg reported that its June refinancing included a $1.15 billion bond sold at a 12.5% yield, along with stronger protections for creditors. These businesses are combining and working to strengthen their financial positions. What puzzles me is the timing. For years, I’ve thought one of the biggest obstacles in personalization was getting a good file to print. Take a photo book. Someone has hundreds of family photos, but the lighting varies, the crops need work, and arranging everything into a book that looks good takes time. Wanting a photo book and actually finishing one are two very different things. AI can do much of that work now. It can help improve the images, choose the photos and put together attractive pages. More people should be able to create something they’re happy to buy. So why are established companies under pressure just as that obstacle is shrinking? I suspect easier creation helps new competitors as much as it helps incumbents. Owning the design tools may become less valuable, while reaching customers and producing and delivering their orders profitably become even more important. Does consolidation put these companies in a better position to capture that opportunity? Or does the industry need to rethink how it sells personalized products?

I started making parody apps. Then I needed somewhere to put them. So now I have an app store. It's called Oops! All Apps. The supervision is questionable. The first three: 🐌 Dawdle: my take on Doodle, for finding a time your people can agree on. Eventually. 🌱 Link Shrub: a Linktree parody with modest ambitions. All your links, one little page, a very expressive plant. 🦋 Looma: a Luma parody for event pages and RSVPs, supervised by a moth with a clipboard. They actually do the jobs. You can make a scheduling poll, publish a page of links, or put together an event and collect RSVPs. I just want the names, artwork, and little details to be as much fun as building them. I'm building these on Spacefast as part of Little Elsewhere, the strange little town I've been putting together on the web. Now it has an app store, because apparently a town needs software as well as a cinema. There's something very satisfying about taking a familiar tool and asking how much personality you can give it while keeping it useful. Have a poke around: https://lnkd.in/gK5BqsTY What should get the parody treatment next?

Some of the toughest men I know spend their Saturdays building tiny plastic robots. Izzi's Gym is a trading card and collectibles store in Sioux Falls. Somewhere along the way it also became a hangout for veterans. I didn't plan that. It just happened. Here's what I've learned from sitting across the counter from them. A lot of these guys came home carrying things most of us will never understand. Some dealt with it in ways that were slowly wrecking their lives. Then they found a hobby. Something they could pour their whole self into. Chasing a set. Building a deck. Learning the market. Showing up on Saturday because a room full of people wanted them there. Which brings me back to the robots. Gundam kits. Hundreds of tiny pieces. Hours at a workbench with nothing but a pair of nippers and a plan. Slow, careful, precise work with their hands. The kind of focus that doesn't leave room for anything else. A lot of vets tell me the hobby shuts off the noise for a while. It's not therapy. They'd be the first to tell you that. But it's structure. It's community. It's a reason to leave the house. I sell cardboard and plastic for a living. I'm not the hero in this story. They are. So if my little store can be a safe place for the people who protected me and this country, that's the best part of the job. Not the sales. That. To every vet who walks through our doors: thank you. There's always a seat at the table. If you served, or you love someone who did, come see us.

Our store screens run themselves now. No agency. No expensive signage company. I built a small app that talks to our internal system, and Yodeck puts it on every screen in the store. Here's what plays at Izzi's Gym all day: Today's events. Tomorrow's events. The Pokémon cards we just bought from customers, minutes after we bought them. Then commercials for every product line we carry. Nobody touches it. The store changes, the screens change. I keep thinking about our restaurant days. Updating a menu board was somebody's job. Which meant it was nobody's job. A sold-out item stayed up until close. A special ran three days late. Every sign was a little bit wrong, all the time. With this, the board would fix itself the second the kitchen ran out. That's the part small business owners miss. The tools got cheap. The flexibility got insane. One guy with AI writing most of the code can build what used to take a vendor and a contract. Your screens should know what your business is doing. Ours finally do.

A dashboard can tell you which idea performed better. Someone still has to come up with an idea worth testing. I wrote about why tech companies need creative marketers with the room to shape what gets made: the name, the voice, the joke, the thing people remember. Bring them in while the strategy is still being decided. The Dashboard Can’t Write the Joke: https://lnkd.in/gdVq-nDK

“We bought a Pikachu” is an incomplete business record. Which set? Which card number? What condition? Which finish? How many copies? Those details affect what we pay, how we list it, and what a customer expects to receive. We used AI to build a buying workflow that carries those details through PENNY. Staff find the product, select its condition and finish, review pricing information, and create the transaction. It sounds like a small data-entry problem. But a bad description at the counter can become a bad listing, an inventory discrepancy, and a customer-service problem later. That's the executive lesson: define the record at the point where the work begins. Before automating a process, ask your team to agree on what a complete, correct input looks like. In our business, a few fields describe a trading card. In yours, they may describe a part, a contract, or a customer request.

I've been dealing with a lot of theft lately. Customers. People I trusted. It wears on you in a way that's hard to explain, because it isn't really about the money. It's about the trust. Somewhere in the middle of it, a thought settled on me that I can't shake, and it has given me more peace than anything else has. None of it was mine to begin with. The store, the inventory, the money, the building. All of it. Scripture is really clear on this. "The earth is the Lord's, and everything in it" (Psalm 24:1). "The silver is mine and the gold is mine, declares the Lord" (Haggai 2:8). King David, one of the richest men of his day, prayed, "Everything comes from you, and we have given you only what comes from your hand" (1 Chronicles 29:14). I'm not the owner. I'm the manager. God handed me these things to take care of for a while, and my job is to be faithful with them. That's it. And here's the part that actually helps. When someone steals from the manager, they haven't really wronged the manager. They've wronged the Owner. And the Owner has said plainly that He handles that Himself. "It is mine to avenge; I will repay" (Romans 12:19). So I still lock the doors. I still count the drawer. I still do my job. But I don't have to carry the anger home, and I don't have to keep score. That's above my pay grade. Job lost everything he had to raiders, literal thieves, in a single afternoon. His response was, "The Lord gave and the Lord has taken away. Blessed be the name of the Lord." I'm not there yet. But I'm closer than I was.

Its hard to believe we're twins as my brother Nick intelligence and creativity is on another level.

Before you add staff, make the waiting visible. At our card shop, customers bring in collections to sell. Each collection takes a different amount of time to evaluate. We used AI to build a check-in queue inside PENNY. Customers enter their names. Our team sees who is being helped, who is next, and how many people are waiting. Selecting the customer starts the buy. Finishing the transaction closes it and records the time taken. That gives us something to examine when the counter gets busy. How long are buys taking? Where does the line build? What does the team need to handle it? The software doesn't answer every staffing question. It gives us a better starting point than remembering that Saturday felt chaotic. Every business has a version of this queue. Returns. Approvals. Quotes. Customer support. A useful first AI project may be making one of those queues visible enough to manage.

90% instant buyback sounds like a great deal. That's the problem. I sell Pokémon packs for a living, so I'm not anti-ripping. Opening packs is fun. The rip, the reveal, the chase. It's a big part of why this hobby is so great. But there's a difference between fun and frictionless. The old way: You buy a pack. You physically open it. Now you're holding cards. To open more, you either spend more money or figure out how to sell what you pulled. Trade it, list it, bring it to a shop. That takes time. That's friction. And that friction is a feature, not a bug. It's the pause that lets you ask "should I keep going?" The new way: Rip digitally. Don't like the hit? Sell it back instantly for 90%. Rip again. And again. And again. Until there's nothing left on your card. Every tweak that makes it more personalized and more irresistible also makes it easier to spend more than you meant to. That's not a hobby anymore. That's a slot machine with Pikachu on it. The other thing that worries me: the money pouring in. I've seen this movie before. I co-founded a print-on-demand company, and this feels exactly like print-on-demand in 2019. Over the next 7 years, close to a billion dollars got invested in that space. And what did all that money actually do? It let companies operate at a loss until the money ran out. The model went from "everyone is making money" to "no one is making money." Once the investment dried up, businesses either consolidated or died. Cheap capital doesn't build a healthy industry. It just delays the reckoning. I'm not saying mystery packs should go away. I'm saying the people building them, the people funding them, and people like me who sell in this space should care about what happens to the collector on the other side of the screen. A healthy hobby is one people can stay in for decades. Not one that empties them out in a weekend. Where do you land on this?