Notes on building software, product and AI. Written for founders and operators deciding what to build next.
Almost every founder who asks for an AI feature says they want a model trained on their data. Most of them want retrieval instead, which is cheaper, faster to change, and fails in ways you can actually see.
The build is done, the features work, and review still sends it back. Here is the checklist we run before every submission, and the three rejections that cause most of the delay.
Most companies asking us for an AI agent want something simpler and more reliable. Here is how we tell the two apart, and what it costs to run the real thing.
An agency quoted you one number and a freelancer quoted a third of it. The gap is real, but it is not about code quality. It is about which failure you can absorb.
We read a lot of software briefs. The ones that earn a fast, honest quote share very little with the long ones. Here is what actually goes in a good brief.
Custom software runs from a few thousand dollars to several hundred thousand, and that range is useless to you. Here is what actually moves the number, and how to get a quote you can trust.
When someone says an app feels expensive, they are not talking about the price. They are describing a hundred small decisions, and almost none of them cost more to make.
Most MVPs are neither minimum nor viable. They are a full product with the deadline moved up. Here is how we scope one that actually ships.
Most AI features fail the same way: they are added because AI is expected, not because they solve a real job. Here is how we decide what to build.
Off the shelf is faster until the day it stops fitting. Here is how to tell which side of that line you are on before you spend the money.
Most bad software engagements were predictable at the sales call. These are the questions that surface the risk early.
The native versus cross platform debate is mostly tribal. The actual decision comes down to three questions about your product.
