← Home
Latent Space · February 15, 2025 · 70m

The State of RAG: What Works, What Does Not

A practitioner's guide to Retrieval Augmented Generation in 2025: which patterns work in production, which are overhyped, and where the technology is headed.

This summary was generated from show notes and public descriptions, not from a full transcript review. Details may contain inaccuracies.

Canon

RAG is fundamentally about enriching the model's environment with relevant context at inference time, providing the information environment that shapes better outputs.
RAG practitioners can control retrieval quality (indexing, chunking, ranking) but cannot fully control generation quality. The Stoic approach focuses on retrieval excellence.
Was this useful?