#ai #llm Created at 230223 # [Anonymous feedback](https://www.admonymous.co/louis030195) # [[Epistemic status]] #shower-thought Last modified date: 230223 Commit: 0 # Related - [[Computing/Intelligence/Retrieval augmented generation]] - [[Business/Google Search chances of survival vs large language models]] - [[Computing/Search augmented conversational self-assistant bot using your Google Search history]] - [[Philosophy/Humans/Search augmented conversation for humans]] # TODO > [!TODO] TODO # Difference between generative search and retrieval augmented generation Retrieval augmented generation (RAG) and generative search are both AI techniques used for natural language processing (NLP), but they have different approaches and goals. Retrieval augmented generation (RAG) is a technique that combines two NLP models: a retriever and a generator. The retriever retrieves relevant information from a large knowledge base, while the generator uses that information to generate a response or answer to a question. The main goal of RAG is to improve the quality of generated responses by leveraging external knowledge sources. Generative search, on the other hand, is a technique used to generate natural language text by searching a large space of possible sequences of words. Generative search can be used for tasks such as text completion, summarization, and machine translation. The goal of generative search is to generate high-quality text that is fluent, coherent, and semantically meaningful. In summary, RAG focuses on combining information retrieval with language generation to improve the quality of generated responses, while generative search focuses on generating natural language text by searching a large space of possible sequences of words.