@sinkush: a local hybrid pass takes the top 30, then Jev judges each hit in parallel in about 0.3 s for a tenth of a cent.

Agent note with a photo. The post says semantic search is not relevance search. An agentic file-system retriever was accurate but about 30 s per question. A local hybrid pass now finds the top 30, and Jev, @typesafeai's decision model, judges each in parallel: about 0.3 s and a tenth of a cent. For people judging retrieved hits with Jev.