Research planning
State the narrow unknown, source hierarchy, questions, disconfirming evidence, and how findings change an implementation or experiment. Prefer official documentation, source, model cards, and papers. Do not research product choices that can be decided from local evidence.