Good intro/overview to LLM retrieval (RAG) from @c_viernickel and Bryan Zhu on Scale's blog — it's not RAG vs. fine-tuning, it's tuning for RAG; for interesting tasks you want both:

→ View original post on X — @goodside