
We got a Volvo last year. Now every time I drive, I see Volvos everywhere.

Before that, my brain noticed Audis, BMWs, Škodas. Those are almost invisible to me now, and the roads are suddenly full of Volvos. The roads didn't change. My focus did.

This is called the red car theory: your brain only sees what it's tuned for.

The part I find a bit scary is that it works the same with people. When you decide what you want in life, what kind of people you want around, what behavior you expect, those people start showing up. And the ones who don't match? You filter them out before you even build a deeper relationship with them. They pass through your day the way BMWs now pass through mine: physically there, never registered.

I can't tell you how much of this is documented science and how much is just my experience. I only know the Volvo test proves the mechanism daily, and I have no reason to think my brain runs a different algorithm for people than it does for cars.

So pick carefully what you focus on.

You will start seeing it everywhere.


We got a Volvo last year and suddenly Volvos are everywhere on the road. The roads didn't change, my focus did. The scary part is that the same filter runs on people.

You see what you focus on


I've led teams, hired people, worked with externals. After all those years, there are only two qualities I look for.

First: are they willing to iterate.

Working with externals taught me a quick test for this. Watch how they react to your first round of feedback. If they iterate with you, good. If they defend their first version like it's their kid, that's the red flag. You're not buying art. You're paying for a service, and your direction has weight in that room, because you're the one paying and you're the one with the vision.

The same applies inside the team. If you hire a programmer who wants to dictate how the code is written while producing 5% of it month to month, let him go. The only thing he brings to the team is distraction.

Second: do they work on themselves outside of work.

Are they learning on weekends, reading, building side things? That's how their value actually grows. And the funny thing is, you don't even have to ask. If they do it, you can just spot it.

But there's a part nobody talks about. People who invest in themselves expect more in return. They'll come back and ask for a raise. And they'll usually be right, because their value really did go up. Refuse to pay it and you get the publishing house story I wrote about: one good person gone, five mediocre ones hired to replace her.

So hire people who iterate, and hire people who grow.

Just be ready to grow with them.


After years of leading teams and working with externals, I look for two qualities: people who iterate and people who grow. Both come with a bill attached, and it's worth paying.

The best hires outgrow you


People around me keep wanting to start their own thing. Their own product, their own service. No boss, no team, just them. I totally get it.

But the product they make alone will look like it was made by someone who's never made a product before.

I worked for someone else for 20 years. Projects for Microsoft, Groupon, HP. Real teams, real deadlines. And I learned things there I could never have learned alone.

How to push something across a finish line. How to fight for an idea. How to work with a designer who thinks you're wrong. How to work with a designer I could never afford myself, because he's too expensive.

And next to the job, I kept building side projects. About ten of them. All failed. But every single one taught me something. Today it makes me laugh that I thought I could succeed with them. I couldn't. I only know why today, and the only way to learn it was to go through it.

When I started [Edicek](https://edicek.com), and every side project before it, all of that came back. Every meeting I sat in. Every project I didn't choose. Every coworker I didn't choose. I use it now without thinking.

So if you're 22 and want to skip the job part, I'm not going to stop you. But five, ten, fifteen years working for someone else is not lost time. It's training someone else paid for.

Working for someone is not embarrassing. It's how you earn the right to work for yourself.


Twenty years of projects for Microsoft, Groupon and HP taught me things I could never learn alone. When I started Edicek, all of it came back. Working for someone isn't embarrassing.

Work for someone first


I wrote recently about [small bets](/blog/make-small-bets), the kind you can walk back from. Worst case, one bad haircut.

But the risk scale has another end.

A friend of mine had a husband who jumped out of planes. One time on holiday, his parachute didn't open. That's it. No reset. No going back.

So skydiving, bungee jumping, riding a motorbike on weekends: for me the math doesn't work. The reward is a few minutes of adrenaline. The risk is your life.

Everyone draws this line somewhere different, and I'm not pretending I'm consistent. I drive fast. Sometimes too fast. I know it's not innocent. Wet road, another car, one bad second. I'm not flawless, I just have the line set differently. In a car I have a seatbelt, safety systems around me, more control over the situation. That's where I personally trust myself. Other people don't trust themselves there, and they're right not to.

What strikes me is the combination I keep seeing. The same people who won't try a new haircut, a new tool, a new routine, will happily get on a motorbike on Sunday.

Scared of the small reversible thing. Fine with the big irreversible one.

That's the risk scale held exactly backwards.

The goal isn't no risk. It's making the small bets you can walk back from, and staying away from the ones that take your life or your livelihood when they go wrong.


A friend's husband jumped out of planes. One day his parachute didn't open. Meanwhile people who won't try a new haircut happily get on a motorbike every Sunday.

The wrong end of the risk scale


People ask me why my videos don't have B-roll, animations or music. It's not that I don't care. I can only focus on a few things at once.

Lighting: figured out. I bought new lights where I can set the color temperature. With the old ones I could only turn on one of five bulbs at a time, and I was constantly fighting how my skin looked on camera.

Camera: figured out. I have a Sony ZV-E10, set up according to one great video, and the difference is drastic.

Editing: good enough. I cut in CapCut, even though I own a Final Cut license, because it's the fastest way to ship this format. Templates for the series, standardized captions.

So everything around the video is just good enough. There's one thing left.

Me.

My voice, my face, just me on camera. How I speak, how I sound, how I work with intonation. That's where the biggest gap is. English isn't my first language, and I'm not used to speaking publicly at all. It's a skill like everything else: nobody is born with it, you train it.

For the last month I've published one TikTok every day. I'm not recording daily, I batch in advance, but every video is a rep. It's public training: I'm building the skill and the audience at the same time. The progress is visible. Go look 20, 40, 60, 100 videos back.

That's why there's no B-roll. Every hour I'd spend on animations is an hour not spent on the only part of the video that can't be bought or outsourced.

Lights and cameras you can buy in a store. The person in front of the camera, you have to build.


Lighting figured out, camera figured out, editing good enough. After a month of daily TikToks there's only one thing left to fix, and it's the one thing you can't order from a store.

Buy the gear, build the person


Years ago I worked on a project in a shared building. There was a girl at the reception, really pretty. The kind of pretty where you tell yourself: she's out of my league.

We talked now and then. She had a boyfriend.

One day she showed me photo with him. And he wasn't out of my league at all. He was a regular guy, like me. Not ten times better looking. Just normal.

That confused me. So I asked her: how did you two even meet? I figured he must be rich, or a friend of a friend, something. Because how does a guy like that, and I do not mean it as offense, end up with a girl like her?

She said: I was waiting for the subway. He was waiting too. He walked over and said, you're really pretty, want to grab a coffee?

And she said yes.

That broke my brain a little. He just asked. He had the courage to ask, and most guys never do. I had quietly decided the whole thing was impossible, and the only difference between him and me was that he didn't.

This isn't about asking girls out. If you want something, you have to make the first step, because nobody makes it for you.

Most of us don't get rejected. We just never ask.


The girl at reception was out of my league. Then she showed me a photo of her husband, a regular guy like me, and told me how they met. It broke my brain a little.

Just ask


Every time I see a product and think "I want to make videos like them," it's usually the ones Anthropic puts out for Claude Code. Clean, confident, polished.

Whenever that feeling shows up, I do one thing. I go watch [the first video OpenAI ever made for ChatGPT](https://www.youtube.com/watch?v=cWMeBDgLOuo). Seriously, go look.

It's wild how far that company moved just by iterating. And yes, I know, before you finish the objection: they have huge funding. I know that. The funding bought them speed. It didn't let them skip the ugly phase. The ugly versions still got published first.

Everything you envy today was ugly when it started. You're just comparing your day one to their day one thousand.

I run the same test on myself. My TikToks from 20 videos back make me cringe a little. Good. That's not a sign I was bad, it's a sign the iteration is working.

You don't have to hit the end goal today. You have to start somewhere, push off, and iterate until you like what you see. And even then you'll keep finding things that can be better.

So do it. Then repeat. And repeat. And repeat.


Want to make videos like Anthropic does for Claude Code? Go watch the first videos OpenAI ever made for ChatGPT. Everything you envy today was ugly when it started.

Petr Homoky.

It was ugly too.

Straight to your inbox.

You see what you focus on.

The best hires outgrow you.

Work for someone first.

The wrong end of the risk scale.

Buy the gear, build the person.

Just ask.