Notes on building products, and the companies around them. What we have learned shipping our own software, and running the operations behind it.
The hardest part of pricing a first product is not the number. It is that charging money is the only test that tells you whether anyone actually wants the thing. Free postpones that test, and the delay is the expense.
Read itThere is exactly one part of a first product that has to be yours: the thing you are here to prove. Everything else is plumbing, and plumbing is cheaper to rent than to build. The hard part is drawing the line precisely.
Read itThe stack feels like an engineering choice, so it gets made on engineering grounds. It is really a decision about how fast you can change the product a year from now. So we pick the boring one on purpose.
Read itThe risk is not that a studio builds the wrong thing. It is that the day the studio leaves, the team inheriting the code cannot move at the speed it was built at. A clean handoff is a design decision, not a closing formality.
Read itThe fastest way to waste a first build is to build it. Half the founders who come to us for an MVP need a week of validation more than they need a codebase, and we would rather say so.
Read itWe say we build companies, not just apps. A federal determination letter is the kind of proof you cannot fake: BeeReady, the nonprofit we stood up from zero, is now a recognized 501(c)(3).
Read itThe launch is not the day you ship. It is the six weeks of positioning, surfaces, and measurement around the day you ship. Most technical founders only do the middle.
Read itThe company does not slow down because the work is hard. It slows down because every decision still routes through one person, and that person is out of hours.
Read itWe use AI on every project we ship. We also spend real time undoing what it does. Pretending otherwise makes teams trust the output more than they should.
Read itThe first month of a fractional CTO engagement is mostly listening, measuring, and fixing the two things that are quietly costing the team a day a week.
Read itYou cannot evaluate the code, so you have to evaluate everything around it. The first engineer is the one hire where getting the process right matters more than getting the resume right.
Read itMost MVP timelines are not lost to writing features. They are lost to deciding things that did not need deciding, and to redoing setup that should have been standard.
Read itWhen generating code is cheap, the scarce thing is doing the unglamorous parts the same way every time. That is an operations problem, not an engineering one.
Read itA prompt you retype is a conversation. A prompt you version, review, and reuse is infrastructure. Only one of those compounds.
Read itEvery AI-generated site we have inherited was missing the same eighteen things. None of them are visible in a screenshot. All of them are visible to Google.
Read itThe React components are the easy part now. Prerendering, share images, CSP headers, and sitemaps are what still take days. So we stopped writing prompts and started writing infrastructure.
Read it