Beyond RAG: How cache-augmented generation reduces latency, complexity for smaller workloads
As LLMs become more capable, many RAG applications can be replaced with cache-augmented generation that include documents in the prompt.