For LLM aspect - My take is the better approach is similar context enrichment for rag - you ask the LLM to generate a summary for a page, and that can easily be cached against a digest/hash for any code changes - and will always be up to date.
> My take is the better approach is similar context enrichment for rag - you ask the LLM to generate a summary for a page,
Curios to hear more about. Are you referring to a doc's page?
Tests can be examples of how to do things. If examples in your documentation aren't backed by unit tests they don't work.
I see your point with tests as examples, but I would find those pretty hard to read/find compared to docs. And if the doc's examples are not backed by tests we get into the same problem as docs being outdated.