In an earlier post I described cloning an app — copying the audio player I’d built onto a tablet to give away. That was a copy: same content, same speaker, a second home. This post is about a step further. I took that same engine and pointed it at a different speaker, and then at a different kind of media — and it just worked. Each time, what should have been “build a whole new app” turned out to be “feed the engine something new.” ...
In the post on search I gave you the clean version: type a word, get every place it was said, tap to jump to that moment. That’s true now. But I made it sound smooth, and for a while it wasn’t. The search had a bug that wore me down — and tracking it down taught me one of the more useful lessons in this series. The symptom was maddening. You’d search a word, see a result, tap “play from here” — and the audio would start playing twenty minutes before the word you searched for. The search found the right thing. It just sent you to the wrong place in the recording. Sometimes by seconds, sometimes by twenty whole minutes. ...
In an earlier post I turned hundreds of hours of recorded Bible teaching into text. AI transcription is good, but not perfect — it mishears words, especially names and rare terms. So the obvious next step was to ask a smart language model to clean up the mistakes. These models have basically read the whole Bible, so I figured fixing a few scripture typos would be easy for them. It wasn’t. And the way it failed taught me something that goes well beyond transcripts — to anyone using AI to correct text against a trusted source. ...
I save every prompt I type when I’m building something. I started doing it long before this blog, for a dull reason: I forget what I did, and the log helps me pick things back up. Going back through one of those logs recently, I found a session worth sharing. It’s closer to what building with AI actually feels like than a clean demo is. The demos show you the good moment. You describe what you want, and a working app appears. That does happen, and it’s real. But most of the work isn’t that. A lot of it is more like this: a task that looked easy, and an AI that got it wrong a few times before we sorted it out. ...
Every time I type a prompt that actually matters — one that solved a real problem or unlocked something I didn’t know how to do — I prefix it with //. That double slash is a convention I made up. It tells my Claude Code hook: this one is worth keeping. The hook saves it to a file. Eventually I want those files to feed an app that writes blog drafts automatically. That is the plan for episode 20 of this series. ...
My daughter is three. She loves one specific playlist of Korean children’s songs, and she likes it best when it plays from a small JBL Bluetooth speaker tucked under her stroller. We walk the neighborhood for about an hour most evenings. I put on her folder, the first song starts, and we go. Thirty seconds later, my phone screen goes off. One more song plays. Then it stops. I have to unlock the phone, tap the next song, put the phone back in my pocket, and keep walking. A minute later, same thing. ...
A pile of audio is a black box. You know the teacher said something about a certain verse. You can almost hear it. But it’s somewhere inside hundreds of hours of recordings, in one lecture out of hundreds, at a point you can’t remember. Without a way to search, those recordings are write-only: you can keep adding to them, but you can never really get anything back out. The bigger the collection grows, the more it buries. ...
My wife listens to a Bible teacher every morning. After I built her an app to make that easier (the first post in this series), she asked for one more thing. It turned out to be the hardest part of the project. Could we also read his teachings, she said, not just hear them? Reading helps her study. She wanted the words on a screen. “Sure,” I thought. “I’ll just turn the audio into text.” How hard could it be? In 2026, turning speech into text is mostly a solved problem. ...
There is a 1974 German theology book that exists in almost no library. One copy lives with its owner here in Korea. He wanted it translated into Korean. I offered to help. First step: get it scanned properly. I visited a professional scanning company. They said the book needed to be taken apart — pages separated, binding cut — before they could run it through their equipment. I relayed this to the owner. He said: do not cut this book. The scanning company said: then we cannot help you. ...
You’ve probably felt this. Someone mentions a technology — Cloudflare Tunnel, RAG, Termux, vector search — and a small door closes in your head. “That sounds powerful,” you think, “but it’s not for me. I’m not an engineer.” For most of my life, that feeling was right. Knowing the name of a technology got you almost nowhere. Between the name and a working result there were months of study, and a level of skill I was never going to reach. ...