Honest reflections on growth, AI, and the messy reality of building startups.
Explore my consulting services, speaking, and workshops.
I wanted to use the bonsai-versus-sequoia metaphor until I checked the tree. Bonsai outlive their wild counterparts. What they do not do is get bigger, and that is a much better description of how small companies actually get stuck.
The AI rollout that stuck didn't start with the tools. It started with what the AI must never touch. Here's the whole setup, in one afternoon.
A widely-cited MIT report found 95% of enterprise AI pilots show no measurable return. The 5% that pay off share a boring shape: AI in the plumbing, drafts-never-sends, governed, measured on one real number.
The thing every manager wants from AI, and the thing it can't do yet, is the same: a chief of staff. A lot of AI disappointment comes from buying as if it exists.
Every leadership team asks the same thing in the first hour: can we connect the AI to Power BI? The instinct is right. The usual way of doing it is the trap.
A manager showed me his AI dashboard: tokens, prompts, seats, all up and to the right. I asked what it had actually changed. The room went quiet.
I'm still figuring out Shorts. That's exactly why I built a playbook - for me and for you.
Another teenager made millions with an app. I'm not here to add to that noise. What I am curious about is what they actually figured out.
We transcribed 18 months of Valve's Next Fest Q&As. Here's what they keep saying that most devs ignore.
I helped build one of the most followed companies on TikTok in Denmark. I also know how damaging short-form content can be. Here's what actually works.
We had zero marketing budget. So I spent six months studying how Steam actually works. Here is what I learned.
I used to think happiness was equal to earning a lot of money. Then I got cancer. Twice.
I used to think self-control was about willpower. Turns out, I was optimizing for the wrong thing.
What actually worked that month, in about two minutes.
Before I ever sold a track, I played it to a room. The crowd came before the product. June I kept landing on the same idea: build the audience first.
The hardest sentence I've learned to say out loud: that's a generous offer, but it doesn't work for me. May was a pricing month.
I grow vegetables on Møen, and when a plant gets sick the instinct to add more is usually wrong. April I spent thinking about retention the same way.
I make audio plugins now, so March became a month-long study of how tiny software companies actually grow. Plugin and creator playbooks I was testing on myself.
A friend asked ChatGPT for the best tool in my category. I wasn't in the answer. February was the month the AI-search shift stopped being theory.
The automations that stuck this month used AI to do more work and somehow felt more human, not less. Four hacks, and the thread connecting them.
I thought I knew what my users wanted. The PMF survey proved me wrong, again.
Nothing is a playbook to guaranteed success. You try 10 ideas, maybe 1 sticks. That's the ratio.