Today is a public holiday in Czechia. I spent the whole working day on a video that is 1 minute and 40 seconds long.
It's a walkthrough of collections in Edicek, the feature I wrote about last week. I recorded it on October 28. Today I edited it, animated it, made the thumbnail, wrote the tweet and shipped it. I tracked every step.
TL;DRThe 4-minute recording was about 2% of 11 hours of work, and the rest was a system I built myself: dual-cam recording, Claude editing in DaVinci Resolve, my own animation tool. Videos that promise AI does it all in 30 seconds sell the sensation, not the hours behind it, so build your own system and don't believe them.
I thought recording was the work
Before I made videos, I thought what you see is what takes the time. You stand in front of the camera, you try it again and again, and that's the hard part. Editing afterwards is easy.
It's the other way around.
The whole video was one take. 4 minutes 19 seconds. Setting the lights and recording took about 15 minutes. But those were 15 minutes in a studio that was already set up. Light temperature presets, cameras, recording software, all ready. It was just the final touches, not the whole setup. The shot looks like one static camera, but there's a lot of tech around it that isn't easy to set up.
Those 4 minutes are still exhausting. After 2 or 3 minutes I can feel the camera draining my energy. I try to be my best, smile, sit up straight, and every next sentence costs more than the last one. By the end I just want it to be over. You can't really see it in the final video, at least I don't, but I feel it every time.
Where the other 11 hours went
| What | Time |
|---|---|
| Script | 41 min |
| Recording | ~15 min |
| Syncing audio from two cameras | 36 min |
| Animating the app screens shown in the video | ~4 h |
| Editing in DaVinci Resolve | ~2.7 h |
| Fixing crackling audio in DaVinci Resolve | 37 min |
| Thumbnail and the Twitter intro | ~1.2 h |
| Tweet, YouTube description, changelog | ~45 min |
11+ hours of active work. Recording was about 2% of it.
And it wasn't smooth. I spent 20 minutes syncing audio on the wrong video. It was the one for awaitio, not the collections one, and the audio was off by 460 milliseconds. It just happens sometimes. One camera had three duplicated frames. In the afternoon the audio in DaVinci Resolve started crackling on my Mac, and I wrote "it's crackling again" more than once before we found it was macOS, not the video.
From 4 minutes 19 seconds of recording, 1 minute 40 seconds made it into the video.
The system behind one day
I'm still amazed it fits into a day. Well, a day and a bit, if you count the recording.
It's because I keep improving my own system. I have my own recording setup that handles two cameras. I edit with Claude in DaVinci Resolve through MCP. This is my third video done that way, I think. Today alone, Claude sent 160 commands into DaVinci Resolve. The animations are made in OnionCut, an animation tool I built and open-sourced two weeks ago.
Everything starts from the brand. Colours, fonts, how things move. So there's very little I have to fix by hand.
The teams I used to work in
For years I worked on products for companies. Teams of 10 or 12 people. Dedicated marketing, dedicated designers, dedicated developers.
I loved apps like Things, Wunderlist and Raycast. I kept asking myself why I couldn't make things that good. Not just the product, but how they presented themselves, how they did their marketing, how they looked from the outside. I thought it took top, top people. That they were impossible to find. That it just wasn't possible for us.
Marketing said things couldn't be done. Designers had excuses. Everything took forever. The designers would make a brand manual, and then it turned out to be almost unusable on social media or in animations. And that wasn't their job anymore. Or we should have told them up front exactly where it would be used, otherwise they would never have built the brand that way. Just excuses. Nobody knows up front how things will work. Everything has to be iterated: the brand, the development, the code, the marketing. All of it.
A video like this one would have taken 2 to 4 weeks in those teams. No exaggeration. And it would have cost around $9,000.
I always had the drive to push the product forward and do it differently, closer to the big brands. I could see the way. I never had the money or the time to actually try it.
Now I build Edicek alone. The product, the development, and videos like this one. Today confirmed something I've felt for the last 10 or 12 years about how products should be made. A team of 12 is no guarantee of output. It just wasn't there.
Nobody sells the hard work
My feed is full of videos where AI edits your videos, AI makes your animations, this job isn't needed anymore, that person isn't needed anymore.
I don't believe the people making them really mean it. They sell a sensation.
And then CEOs watch it. They start believing they don't need their people, and they let them feel it. They feel they can just do it themselves. They know best how they want it anyway. Why explain it to someone they have to pay, who will only prompt it, and then wait ten times longer than if they tuned it themselves? The people pay for it in the end. For what? For someone's 30 seconds of sensation.
I don't want to take anyone's illusions away, but teams will always have an advantage over individualists. It's only about who you build the team with, who you hand the responsibility to, and with whom you can actually take a step back, because it's no longer your worry. If you still have to keep stepping in, you picked the wrong person, and it's time to replace them.
Or it can be a signal that you're poking your nose where it doesn't belong. Just because you can pick one tiny thing you'd change right away, or put the effort into, to show you still have the output. Trust me, that's not how it works.
Today I can walk up to any product and start pointing at things that could be better. Small details that need tuning here and here. But when you talk to the people behind it, they either didn't have the time or weren't the right people for it. In many cases it just wasn't a priority. There was no budget, or no time, because of some nonsensical expectations. The only thing you'll achieve by stepping in is that the people whose responsibility it is will start to hate you. They'll grow a dislike for the product and for working on it. Then it's only a matter of time before they leave. And even if they stay, it will show in their output.
Nobody sells the hard work. Imagine an ad for a piano course that said: "Here's a course that will help you play the piano. But honestly, don't expect miracles unless you practise regularly, 10 to 15 hours a week."
Would you buy it? No.
But that's the true version. Claude didn't make this video in 30 seconds. I made it in a day, with a system I've been building for months.
Build your own system. And don't believe 30-second videos.
Here's the video that took the whole day, cut from one 4-minute take down to 1:40:









