Default RAG to hybrid (dense + sparse) retrieval with re-ranking for 25–40% precision

Starmorph·medium signal

In 2026 the recommended default is hybrid retrieval — dense embeddings for meaning plus sparse/BM25 for exact terms, part numbers, and rare entities — followed by a re-ranking pass, which lifts precision 25–40% over naive RAG at modest latency cost. Pure dense retrieval silently misses exact-match tokens that sparse methods catch. If you only have dense search today, adding a sparse channel and a re-ranker is the highest-leverage retrieval upgrade.