Algorithm, Search & Distribution
Retrieval-augmented generation
Also called: RAG
Retrieval-augmented generation puts a search step in front of a language model. The question is used to pull relevant passages from an index, those passages go into the model's context, and the answer is generated from them. This keeps answers current without retraining the model, and it makes citation possible. Creators meet it from the other side: AI search products and chat assistants that summarise the web are usually doing this, and whether your page or transcript is retrievable decides whether you get quoted. Retrieval reduces invention without eliminating it, since a model can still misread what it retrieved.