So far the challenge of going from idea to App Store as a one-man operation using regular software and AI tools is coming along surprisingly well!
The biggest adjustment I’ve had to make is actually starting my “shifts” at 6PM, when the token consumption rates go down. More work for less tokens. Working with AI is like collecting water in a balloon; you have to ration how much you put into a typical working session – if you use too many tokens, the balloon pos and you’re timed out until the next reset. Usually a few hours away. I’ve been using Claude and it’s varying “flavors” of AI engines. Haiku for basic Q&A, all the way up to Opus for strategic planning and helping me out of some thorny development roadblocks. I was one of the few who did get to run a prompt through Fable to help with a logic engine problem I was trying to define and get coded. It did an amazing job and was dead-on correct in its strategy and execution plan. Sure Fable was “expensive”, but you wouldn’t expect to have your best resource doing the block and tackle work. For very capable coding sessions, there’s Sonnet, which is about 80% of my use. The typical workflow takes a bit of getting used to, but once you get a rhythm it is quite hard to take much-needed breaks.
This was a good week in that I have three apps ready to test up on TestFlight. If you’re reading this, I encourage you to try one or all of them out for yourself. They are 100% AI generated, directed by yours truly. In fact, so is the website for my little consulting firm, Bacon Cheese Frosting, which is now pulling double-duty housing a growing gallery of my apps. This experiment was mainly to test myself to see if I can learn to use AI tools effectively – alone and in combinations – to pull off things that used to take lots of people, money and time. The experiment is a success, and I look forward to these apps being available in the Apple App Store for public enjoyment soon.
Nothing puts this experiment into focus quite like this little app called SimpleSignal. It was an idea my brother had a few years ago. Just a simple little thing to help you meet up with friends in a crowd. An app that would light up your phone screen with a color of your choice, you’d send a text to a friend or group telling them to look for that color, while you held up your phone like a human lighthouse. Simple concept that works remarkably well in practice I’ll add. Back when we discussed it, the furthest I could get to making it “real” was to sketch out a prototype and maybe make it clickable in Adobe Xd or Figma. Someone still needed to code the thing and figure out how to get it into the App Store. Flash forward to last week. I dusted off those sketches, wrote up a one-pager prompt, and in an hour Claude Code was working on bringing screens to life that Claude Design had freshly built and tweaked to my liking. In another hour I had it running on my phone, sending real text messages and doing exactly what we envisioned it would do years ago. It’s now up on TestFlight, ready for you to check out. If you’re going to a ballgame, amusement park or anywhere there’s a crowd this summer, try it on your family. If you’re going to a concert, there’s even a little Easter egg tucked in that plays a lit cigarette lighter on loop so you can cheer on that power ballad.
These are still very much works in progress, but they run and are at a stage where getting input from others will help make a much better product. Come help me test these apps and be a part of this Grand Experiment. I really appreciate it!
