Building in public.
One person, a lot of AI, and a written record of what actually happened — including the parts that stopped.
I am not a founder. I started anyway.
No startup background, no network, no capital worth mentioning — just a small business of my own and the admin that came with it. What changed is that the work which used to need a team can now be done by one person, and that turns out to be the easy half.
From idea to first venture: how the sequence actually ran
The real path V-001 took, including the two places it nearly stopped. Mostly a record of how much of early work is deleting ideas rather than having them.
What ten AI agents are good at, and the three things I took back
A ten-profile operating model is genuinely good at some things and quietly bad at others. Notes from the first months, and the decisions I pulled back.
A red team is only useful if it can stop you
Adding a challenge stage changes nothing if the challenge has no authority. Notes on giving an adversarial profile a real veto, and what broke when I did.
What Eden taught me about first contact
Running our own web chat before selling one to anybody else. The failure modes are not the ones the demos prepare you for.