Prompt used
The brief, constraints, model/tool, and important follow-up prompts.
Testing how well AI can reposition and restructure the benchmark platform itself.
Linked case
The IdeaGrains website itself is a benchmark case for testing whether AI can reposition, simplify, and structure a product-testing platform.
Full write-ups use this structure as tests are completed, with prompts, screenshots, model outputs, code notes, caveats, and final scorecards where relevant.
The brief, constraints, model/tool, and important follow-up prompts.
The baseline product state, known issues, screenshots, and repo constraints.
Screenshots, code diffs, UI notes, and behavioural changes from the model output.
The parts worth reusing: structure, implementation choices, UX improvements, or analysis.
Bugs, weak assumptions, overreach, regressions, or outputs that needed correction.
The manual judgement, debugging, edits, and product decisions needed to make it usable.
A practical call on whether this model/tool should be used again for similar work.