Really building

Last week I decided to try something. I had an idea for a new app, pulled up Claude Code, and began building. The concept wasn't theoretical, I knew what I had in mind, and wanted to get to a solution immediately.

Now, roughly a week later, I have a full fledged app on Mac, iPhone, and Apple Watch. And it's working (nothing for public release yet though, it's all in TestFlight).

I just want to pause for a minute and call out how crazy that is.

Even a year ago this would have been impossible for someone like me—a designer yes, but my development ability is limited. This is an app I've wanted for years, but I couldn't afford to paying someone to build it, and nothing on the market quite meets my needs. I also tried to learn the basics, year after painful year—trying coding courses, reading documentation, brute forcing coding principles into my brain via every method I could try to figure out—but it never quite clicked. At nearly 40 years old I just can't get my mind to build the things I imagine.

But with the help of Claude, and tools like it, I can get closer than ever before.

That's monumental. For folks who like to tinker and build things and iterate endlessly—this changes everything, and I can't be purely cynical.

The idea was a todo app. I know it's a joke that nearly every designer and developer has to try their hand at one eventually. I've had various ideas in my head for a decade, and I've spent as much time looking at everything on the market.

And true be told, there's some fantastic resources out there. If I could I'd just switch to Apple Reminders and move on. But I have a few ways that I handle todo lists that can't quite work with the default system apps.

So, taking everything I've learned from Omnifocus, Things, Reminders, Tada, ToDue, Clear, Tadee, and a dozen other app names I could just make up (remember the milk?), I set out to build a todo workflow that's just for me.

When I open the app, I'm calling it Stuff Stuff, I want to see a list of tasks for today. That's should be front and center. Then, with a single tap I want to see what's on the list for tomorrow and afterward. Then I also want to see tasks without assigned dates. I also want to tap on tasks to add subtasks, notes, and see a log of how the task has changed over time. Oh and I should be able to do all this device agnostic. Can you tell how excited I am?

Using Claude Code I was able to get the TestFlights up and running, run through dozens of iterations in days, create an Apple Watch (!!!) with a complication, and then make sure all the data synced across my Mac, phone, and watch with CloudKit.

All this stuff (ha) is kind of crazy. Yes, I know we're all nervous about AI and I'm foremost among that. But there's genuinely delightful moments where we can dig in and tinker and play with ideas that we've imagined for years, but never quite been able to complete.

Now, Claude isn't perfect, but it's really good. There were times it didn't get the UI right and I had to tinker. I even had to bust out the iPad at one point to sketch out the proper flow for several interactions and feed that back into Claude. But it's fun. I didn't tell Claude: "Make me a todo app". I had a back and forth conversation, akin to working with a colleague. I tried scores of ideas, shipped dozens of quick versions, and kept testing.

The app works, and as of yesterday I've replaced Things 3 with Stuff Stuff. I'm now using it exclusively to manage my todos. There's still some rough edges, user flows I want to cleanup, and UI things that aren't right—but it's workable and useful now, a week after deciding I wanted to try the idea.

Now, with all that said, I don't know whether it'd be responsible to share this out to the world at this stage. It works for me and I'm really happy. I've asked a developer friend if he'd like to help me, and we might continue to work on it together to ensure that it's secure, and built correctly—but for right now, for me, it's useful and I'm getting value out of it.

These kind of experiments—taking ideas and running them up the flag pole—continually increases the surface area for understanding good products. This is probably the twentieth idea I've tested in the last year, taking little app ideas and trying to build something—and my confidence in what's possible continues to grow. But you know what? I don't think I'm anywhere close to having it figured out, and that's exciting. Building an app like this humbles me, reminds me of how much is till want to learn and understand about great interfaces, what makes software useful, and how I can build simple (but not simplistic) designs.