Scattered documentation across platforms like Confluence and Slack becomes stale, slowing down onboarding and creating repetitive work because it lives separately from the code.
Despite advancements from manual sync to AI agents in the IDE, the fundamental issue persists because developers still act as the manual bridge between code and separate documentation systems.
AI agents treat outdated documentation as a source of truth, leading them to confidently generate incorrect code and making stale docs a source of bugs, not just an inconvenience.
Choose between a `docs` folder, which enforces code review via pull requests for every change, and a GitHub Wiki submodule, which lowers the friction for updates but requires a sync process.
When documentation and code share the same context, AI agents can automatically generate, update, and validate code against documented rules and standards.
Convert long, easily ignored architectural documents into custom lint rules using an AI agent to enforce standards automatically within the IDE.
Transition from being a manual bottleneck to an editor of a system where AI agents draft, update, and enforce documentation as an integrated part of the development cycle.
Jobs that call for the skills explored in this talk.