Tell HN: We need an accounting system for cognitive debt
AI-generated code creates "cognitive debt"—code produced faster than humans can understand it. The author proposes a ledger system with microcertifications tracking who understands what code, verified by domain experts. Every line of code would be cognitive debt until certified as part of a living mental model.
Background
- "Cognitive debt" (term from professor Margaret Storey) describes the gap between code created by AI coding assistants and the human understanding needed to maintain it. Code compiles and tests pass, but no one truly grasps what the system does.
- The author (HN user "mikael") proposes an "accounting system": an append-only ledger that tracks who understands each part of the codebase, at what depth, and who verified that understanding. Every line of code is assumed "cognitive debt" until certified by a human demonstration to a domain expert.
- Background context: "agentic coding" tools (Claude Code, Devin, Copilot agents) now autonomously write and revise large codebases. Teams increasingly feel they can't safely change code they didn't write — they've become "tourists in their own codebase."
- The proposal builds on existing concepts (code ownership, bus factor, software comprehension) but reframes them for an era where AI, not humans, produces most of the code.