Gemini

This is another tab in my daughter’s media player. The last two posts were about an English song tab. This one is about English stories. The Need: The Stories Only Live on the Device We have a set of eight English picture books about a cartoon pig family. It comes with a small talking reader. The device has eight book icons down the left and right edges. You press a book icon, then press one of ten buttons in the middle, and it reads that page out loud. ...

This is part two. In the first post I turned 19 free English board books into a tab in my daughter’s media player. The tab worked. Then I used it with her for one evening and found three things that bothered me. The Need: Three Small Annoyances Every video started with a roughly 10-second channel intro and ended with a promo card for other videos. In a song tab, that is a jolt every time. My wife and I had agreed she shouldn’t watch video yet, but the tab played video by default. The lyrics and translation were below the picture. To read them I had to scroll. And my daughter doesn’t like it when “her” picture shrinks or disappears off the top of her phone. None of these is a big problem. All three would have stopped us using the tab. ...

A transcript is a record of what a person said. If an AI “fixes” it into what the person should have said, it is no longer a record. I learned this on my own archive of Korean sermon recordings. The Need: Fix the Mistakes, Keep the Speaker Speech-to-text gets Bible words wrong. It writes a hardship word for a different hardship word, a town name as nonsense, a book of the Bible as another book. For a searchable archive, that matters, because people search for those names. ...

I like the idea of running my own AI. No bill, no data leaving my house. So when I needed to clean up a long transcript, I tried the local way first. The result changed my mind about when “free and local” is worth it. The Need: Speech-to-Text Gets Names Wrong Whisper turns a Korean sermon into text very well, but it makes a certain kind of mistake. It hears Bible words as ordinary words. A town name becomes a nonsense word. A book of the Bible becomes a different book. And one example I like: the Korean word for Psalms, sipyeon (시편), sounds almost like “ten pyeon” (10편). Whisper wrote “10” where the speaker said “Psalms”. ...

Most of the 36 pictures in part one got one AI generation and, at most, one round of revisions — generate, glance, done. One took five. It’s 오빠 생각, “Thinking of My Brother,” an old Korean folk song, the kind of thing grandparents still hum — and one of the few songs in that whole library I’ve actually sung to my daughter myself, at bedtime, not just something we play from a folder. ...

I have a set of Korean children’s Bible verse audios. It’s a set of 10 short verses. Each verse is set to a simple song. This helps with memorization. This set was in a folder structure, just like the nursery rhymes I talked about in Part 1. It had the actual audio files and a list of titles. A three-year-old child cannot read these titles. The idea of one picture per item already worked once. So, this became the next clear candidate. Honestly, this was the part I cared about most. ...

My daughter is three. She can’t read yet, which means for months there was a whole folder of songs in her Kids Player that only I could operate. She’d point at the tab, I’d scroll the list of titles, and she’d wait for me to land on the right one. She could listen to the songs. She just couldn’t choose one herself. That’s a small thing, but it bothered me more than I expected. Everything else in her player — her audiobook sets, her story tab — she already drives on her own, by tapping a cover image. This one folder of family and folk songs was the exception, and the only reason was that it had titles instead of pictures. ...

I asked for a hunter’s pistol to become a rifle. Everything else in the picture stayed exactly the same — same animals, same background, same pose. Clean, one-item edit. A few days later I asked for two things in the same request: add a conch shell that belongs in the scene, and remove a child who shouldn’t be in it. Half the animals the song’s lyrics called for vanished. Nothing I’d asked to touch. ...

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. ...

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. ...