Boxed Up

I'd been playing FIFA Ultimate Team for years. The loop — open packs, build a roster, trade in the marketplace — was one of the most engaging mechanics I'd ever encountered, and EA had been printing money with it for a decade. I kept asking why nobody had transplanted it outside of football.

I wasn't a sneakerhead. But the market was a self-evident fit. Sneakers had everything the FUT loop needed — scarcity built in, social currency, an existing collector culture, and a passionate online community. So I built Boxed Up to test the hypothesis: would the trading-card-game mechanic work in any vertical with collector energy, not just football?

For months, that was the answer. I ran paid ads, they brought in five to ten new players a day. Enough to prove the loop worked — retention was solid, people loved opening packs — but the unit economics didn't pencil at that pace. I needed a different growth mechanism.

So I built a Discord and pushed every active player to join it with an in-app prompt. About twenty members joined.

A few weeks in, one of those twenty members messaged me. He said he was going to get his friends to play.

What he actually did was post Boxed Up in the Discord of another game — a “Pokémon GO but for sneakers” product in the same niche. Players there had been waiting for a real trading and pack-opening loop. They swarmed.

I went to sleep with twenty members in the Discord. I woke up with a hundred and twenty.

From that night, the growth never stopped. Peak days were 2,500 new players, and that pace held for months.

Paid ads had bought users. The Discord kept them and then recruited their own. Most of the next hundred thousand players came through trade requests with friends, in-app prompts, and chatter in the server.

The Discord became more than a growth channel. It was where the in-game economy lived — players negotiating trades, swapping rare cards, building reputations. It's where I heard which features were broken before any bug report came in. It's where the meta of the game emerged: which cards were undervalued, which boxes had the best EV, who the whales were.

Fifteen thousand members and twenty thousand App Store ratings later, it's still the most important asset on the product.

Boxed Up runs a real in-game market. Players trade cards with each other for in-game currency, and prices float on supply, demand, and scarcity. Any product with a live economy attracts adversarial players, and Boxed Up was no exception.

When an exploit hit, I had hours, not days, to identify the vulnerability, ship a patch, and ban the offending accounts before the distortion broke trust with legitimate players. [Add one specific incident worth recounting — what kind of exploit, how you spotted it, how fast you shipped the fix.]

Solo founders don't usually run live ops at that pace. I learned to.

In 2026, I rebuilt the entire app from the ground up. Not because the existing code was broken — because I wanted to design something different: a codebase shaped around what an AI coding agent does well, not just around what I do well.

I migrated the backend to a setup that was clean, idiomatic, and structured the way an agent expects to find things. I rewrote the frontend on a tech stack the AI was meaningfully better at. [Add: which stacks — what did you move from and to? Even one or two names lands the story harder.]

The result wasn't “AI helps me code.” It was “AI does most of the coding; I do the product and the judgment calls.” The rebuild took weeks instead of months. Features that had been stuck in my backlog for a year started shipping in days. That was when I stopped treating AI tooling as a force multiplier and started treating it as a teammate — one I'd designed the codebase to make effective.

Boxed Up is four years old, profitable, and still growing. The bigger thing it taught me: the engine I'd built — the economy, the marketplace, the pack mechanics, the trivia layer — could be re-skinned for any cultural vertical with collector energy. The fixed cost was the infrastructure. The variable cost was the assets and the theme.

That insight became Filmed Up (movies, 2025) and Tuned Up (music, 2026). A family of three games on a shared core. Combined: $375K+ revenue, 550K+ downloads.