### Source Peer feedback, collected in #167. Quotes (translated): what you could well have included is what you learned from using AI, because now it is more a log than a reflection; AI has contributed to large parts of the project (backend, session handling, frontend and tests), so it is important to consider how the parts work together. ### What to do - [ ] Add a section at the top of `docs/AI_USAGE.md` that is a reflection and not a log: what worked, what did not, what the AI got wrong and how it was found (for example the tests that were green and worthless until they were mutated), and what we would do differently. - [ ] Say how we checked that the parts made with AI fit together (frontend and backend, session and pages), and where that was not done. - [ ] Keep the log below it. ### Done when - [ ] The document starts with a short reflection by the team, written by the team. _Comment made with Claude Code in discussion with Ronja_