AI Software Factory Series
A 17-part guide. Start at Part 1, or jump to any part below.
The parts
- Code Generation Got Cheap. An AI Software Factory Is How You Check It.
- Twelve Polished Pages, and Half of Them Were Never Said
- The Pull Request Was Clean. The Ticket Was Not.
- Twenty Pull Requests, and a Review Queue That Tripled
- Parallel Agents Do Not Need a Smarter Plan. They Need a Schedule.
- The Build Loop Only Works If It Knows How to Stop
- The Review Queue, Not the Model, Decides Whether Your AI Factory Works
- Merged Is Not Released
- At 2 a.m., the Agent May Read. Only You Restore Service.
- Agents Already Pay for Themselves on the Work Nobody Roadmaps
- Your Factory Gets Worse When You Stop Testing Changes to the Factory
- If the Factory Cannot Check Its Own Work, You Cannot Tell It Is Helping
- The Model Proposes. The Harness Decides.
- Start With Agents That Only Comment
- Agents Forget Every Run. The Second Brain Is How a Factory Compounds.
- Beads and WikiTicket Are Not Competitors. They Answer Different Questions.
- Faster Code, Same Delivery: Five Eras to an AI Software Factory