Real-world product, AI and enterprise software field notes for teams building and scaling digital systems.
Apple decided to use Gemini as the foundation for its next generation of Foundation Models. The decision matters to teams weighing build against buy because even a company able to fund either route chose to concentrate control on data, orchestration and product experience.
Read articleClassifying invoices, extracting fields from contracts and routing documents are good uses for AI only when the documents remain traceable and the data does not end up where it should not.
Read articleA large company had only one in-house React Native developer. When he left, the app ground to a halt. No one in the company could handle the code. They called us, we took over the project, and brought it to the store submission stage. Nothing heroic. It's what we do.
Read articleA client believed they had already completed 70% of the work because their internal system had been running for three years. That 70% covered the simplest case: one customer who already knew how the system worked.
Read articleWe use AI every day. Anyone who signs off on code must still be able to explain what it does, why it was written that way and what may break when a requirement changes.
Read articleThe first AI agent we put into production worked well in staging. In production, it made decisions that were technically correct and operationally disastrous because no one had defined what it could do on its own. Since then, we design permissions, boundaries and accountability first. The model comes later.
Read article