Essays on running software companies with autonomous systems — what holds, what breaks, and what compounds. Written by one founder to clarify decisions, not to fill a feed.

The build times and slow colleagues I spent years trying to remove weren't just waste. They were pacing me, and I didn't notice until the pacing was gone.
Read the essay
Every guide said find a co-founder before a customer. Right for my first company, wrong for what I run now, except for one thing the guides never mention.
Read the essay
I built something tireless and called it autonomous. A site about Swedish houses showed the difference: a loop that never grades its bets is running, not learning.
Read the essay
AI made sending free. Attention did not get cheaper — it got scarcer. Why the systems that decide what NOT to send are becoming the real asset.
Read the essay
The best acquisition targets have no story, no narrative, no hype. They have cashflow. Why the operator model wins exactly where venture capital cannot look.
Read the essay
Production fell to zero, so attention became the only scarce thing left. You win it with many cheap baits, cast fast and fed when they pull, not one perfect one.
Read the essay
The moat in software is no longer the product. It is the machine that builds and runs the product. A feature can be copied. A compounded system cannot.
Read the essay
AI did not solve the solo founder problem. It moved the bottleneck from output to attention. What triage as operating discipline looks like.
Read the essay
Execution got cheaper with AI. Sequencing got more expensive. The right work on the wrong thing is now the most expensive mistake a solo founder can make.
Read the essay
VC, bootstrap, solo plus AI. Three games for building the same kind of company. Each is internally consistent. The mistake is mixing them.
Read the essay
Saying no to capital is not romance or fear. For coherence-first companies, the money must become hiring. Hiring destroys what the company is selling.
Read the essay
The popular advice is to delegate everything an agent can do. For founders, that erases the thing the role exists to produce.
Read the essay
AI removes layers of work that used to require people. The argument that it ends at zero is wrong. Something has to decide what the company is for.
Read the essay
AI didn't just raise the productivity of individuals. It changed which company shapes are stable. The one-person company is one of them.
Read the essay