This week Robi learned to work without the internet. Two weeks of nothing but this, in three parts, plus a starter pack and a way to send your own audio.
v3 on the table, July 7, 2026, the week the offline work started.
First, the player. It starts up and plays whatever it already has, without waiting on a network, and picks up exactly where it left off. When the network is gone, it says so instead of hanging.
Second, the phone. With no Wi-Fi at all, the app can now play, pause, skip and pick a mix over Bluetooth. A new phone can’t just connect: the player chimes, and someone has to hold the knob to let it in.
Third, the app itself. It tries the home network first, then Bluetooth, and if neither works it says what it can’t do instead of spinning. Anything added while the player is out of reach waits, and lands when they meet again.
Then the extras. A new player comes with a starter pack of eight shows, in English and Chinese, and the episodes download once there’s Wi-Fi. I also tried preloading audio so a new unit could play straight out of the box. And a parent can now send their own audio files from the phone to the player.
This week the coding agent on the Mac mini in the office is back to work, on new terms. Its work was always good; there was just too much of it. I review and test everything, whoever wrote it, and an agent that never stops makes a queue that never ends. So now it works only on what I assign, with a cap on how much can wait for my review, and the bigger, tightly scoped jobs run overnight. Mornings start with fifteen minutes of reviewing what it made. I’ve written down what runs on its own and what waits for me: anything a customer sees, and every merge, waits for me. The starter pack was its first assignment.
What did not move in these two weeks: the production board, the manufacturing emails, the beta invites, the campaign page. Building feels like progress because it is progress on one front. The other fronts are quietly stale, and next week is theirs.
One more decision this month. Since the layoff, I’ve been selectively applying and interviewing for product management roles, because not doing it felt irresponsible. One conversation got me excited: a growing company that needs a product process, which is exactly what I’ve been building for kid.studio. So why don’t I feel that about kid.studio? Part of it is that it’s just me. Part of it is that it isn’t moving as fast as I’d like. But the process I’ve built works, I enjoy it, and I’m making progress. And I can’t find the same care for someone else’s product that I have for getting Robi to market. So I’m stopping. Full-time roles are out, because kid.studio is the job. I’m all in on kid.studio!
Next: the schematic. It has been open since June 10th.