In this Module we are going to continue where our Webinars of 2026 left off. We’ll talk about organizing modern codebases in such a way that it makes agents and humans more productive.
How do we organize a project so a fresh coding agent can find relevant information, carry out a task, and retain useful lessons for the next session?
In particular we will talk about:
- Progressive disclosure and document trees
- Packaging knowledge and reusable tools
- Codebase self-improvement
- Control centers
In fact, this entire module is an ode to the progressive disclosure as applied to code-intensive projects.
The plan is following:
- Unit 1: Organize your codebase for AI agents - Theory + Source Code Drop + Optional practice
- Unit 2: Find and fix what gets in agent’s way - Theory + Source Code Drop + Optional Practice
- Unit 3: Find and fix what gets in your agent’s way — walkthrough - Video Recording
- Unit 4: Turn completed work into reusable knowledge and scripts (includes control centers) - Theory + Optional Practice
- Unit 5: Make the approaches work in your projects - Live Wrap-up Session.
Units will be unlocked over the course of October.
The module will continue with Executable specs and making agents write code reliably in November.