Solving weirdly specific problems

I have approximately zero programming experience, but I wanted to experiment with Claude Code, so I decided to try to build a hyperlocal surf-forecasting app for myself.

Surfline is the incumbent surf-forecasting monopoly, but their predictions are infamously unreliable for San Francisco's powerful, shifty beach breaks. So, like many other Bay Area surfers, I check niche websites that post raw oceanographic data to get a sense of what to expect.

I spun up Claude Code for the first time, identified the best public data sources, defined exactly what I wanted, created a detailed plan, and pressed go, conjuring a ghostly legion of tireless minds to weave this idea into reality.

Claude built the app in a single shot.

I was delighted! So I iterated the design, added a bunch of features, and started using it daily.

Friends asked to use it, so I needed to get it in the App Store. And if I was going to do that, I might as well add a modest paid subscription to support improving it over time. Oh, and obviously I'd need a landing page, so I worked with Claude to build and deploy a site.

Honestly, I can't quite believe it myself, but Dialed just launched in the App Store!

The app is extremely simple. All it does is bring data from different sources into one place to help you decide when and where to surf along this beautiful stretch of coastline.

With its global scale, it would never make sense for Surfline to do something this niche, nor would the tiny TAM for Dialed justify the investment required to build it without AI tools. But with those tools, I was able to create a piece of software to solve my own weirdly specific problem, and offer it to others who might find it useful.

What other weirdly specific problems have abruptly become feasible to solve?