Skip to text
Kill the MonkeyThe workshop
Report
00
Date
16.07.2026

The video I wouldn't edit by hand

Production notebook · 16.07.2026 · I know how to edit; what I wanted was not to edit by hand from my phone, and that tool didn't exist: everything starts there

This starts a month before the channel, and it doesn’t start with artificial intelligence. It starts with a present.

I wanted to surprise my wife: a video of our three children telling the camera what they like most about her. I already had the clips on my phone, each one on its own.

All that was left was joining them up. And that’s where I dug my heels in.

I know how to edit video. That wasn’t the problem

I studied audiovisual communication and I’ve been editing video for years. Sitting down with an editor, cutting six pieces, putting them in order and exporting is an afternoon’s work, and I know exactly how it’s done.

Which is precisely why I didn’t want to do it. Cut, drag, adjust, export again because the third piece looks dark: that’s monkeywork with twenty years of craft on top. And I was on the sofa with my phone in my hand, not sitting at the laptop.

What I wanted was to say what I wanted and have the thing come out assembled.

Two days asking for a button that doesn’t exist

I asked Gemini for a tool that would assemble the whole carousel. I insisted: I didn’t want it created for me or changed for me, I wanted it put together. The next day I asked again in different words.

That tool never turned up. What did turn up, looking on my own, was CapCut, a phone app I used to assemble it myself, with my finger, without switching the computer on. An app anyone already has on their phone: that is the best measure of how little the tool I had gone looking for mattered.

So: the video got made. What didn’t get made was the thing I had gone looking for.

A month on I know why that button didn’t exist, and it wasn’t for lack of technology. Anyone can cut and paste with the right app. The hard part of a video several minutes long isn’t any one of its pieces: it’s getting them to join. Deciding the order, the pace, and making six pieces read as one is the part nobody hands you finished.

Everything else comes from there

Everything in this notebook comes from that day. The children’s video channel, the twenty-five bad takes, the log of failures, the arguments with the camera: they’re the schooling, not the point.

The point is a music video several minutes long. A real one, with its song, its story and its characters, made at home from start to finish — and without sitting down to cut it together by hand, piece by piece.

And I already know which one it’ll be. On 7 August I wrote a song for my youngest daughter, who is two, and it opens by saying don’t come to me with princess stories. That song is where all this is headed.

How to read this notebook

Each report is a step up, and it only goes in if something changed the way I work. They run in order, with what went wrong in plain sight, which is half of what you learn.

It isn’t a manual on how to use these tools, which change every month anyway. It’s what happens to someone who knows video but knows nothing about making video with artificial intelligence: long videos built out of AI-generated clips instead of ones shot with a camera. And writes it all down along the way.

Next report 01 · The only clip that came out right came from one sentence · Read