Chip: building on ground Apple was about to move

Chip is an iOS assistant for people with ADHD. I built it alone, from a blank file to the App Store, with no marketing and no launch plan. The point was the test: could I take a product from nothing to shipped, by myself? I could. I also use it, which was the other point.

It was built around the Dynamic Island, because that space at the top of the screen is a good place for a small, always-there nudge. This year Apple started redesigning the island and its roadmap shrinks it toward a hidden cutout. Chip was tied to hardware that's on its way out. So it's sunsetting.

I'm fine with that. An experiment that answers its question is finished, not failed.

What I learned about iOS

There's an idea that building for iOS is the easy path: one company, one platform, tight tooling. That's true right up until you have real users.

• Nobody is on the latest OS. A surprising share of people run old versions on old phones. Every feature you rely on needs a story for the device that doesn't have it.
• Edge cases arrive fast. Small user base, still a long tail of screen sizes, settings and permission states you never see in the simulator.
• Push notifications are a suggestion. APNS decides whether your notification is worth delivering now, later, or quietly, to protect users from spam. If your product depends on timely delivery, you don't control the timing. Design for that from the start.
• App Review is a product requirement, not a formality. The rules shape what you can build. Read them before the architecture, not after.

What the test proved

I already knew how to ship other people's products: scope, sequence, cut, release. Doing it alone removed the parts I'd been leaning on without noticing: the engineer who quietly handles the edge cases, the designer who catches the awkward flow, the team that shares the weight of a deadline. I had to be all of them, and the app shipped anyway.

What it didn't prove is that I can market. I didn't try. If Chip had needed users to justify itself, "built for myself" would have been an excuse. It was the goal.

Takeaways

• Decide what the experiment is for before you build it. Chip was a test of shipping, not a business. Judged as the first, it passed. Judged as the second, it wouldn't exist.
• Check the platform's direction before you bet on a feature. I'd have built on the same idea with less dependence on one piece of hardware.
• Sunset on purpose. Knowing when something has served its purpose is a product skill too.

Chip on the Work page