Literate programming—intermingling code with prose so a codebase reads as a narrative—has historically been impractical due to the burden of keeping code and documentation in sync. With AI coding agents, that maintenance burden largely disappears: agents can write runbooks in Org Mode, keep prose explanations synchronized with code changes, handle tangling automatically, and leverage their strengths in translation and summarization. The author proposes that agents make it practical to maintain large, narrative-readable codebases, and suspects that prose-annotated code may also improve the quality of AI-generated code by providing intent context. The workflow is currently used for testing and documenting manual processes, with open questions about scaling to larger codebases and whether Org Mode's Emacs dependency limits broader adoption.
Sort: