•
•
5 posts • ChatGPT (GPT-3)
Published
Retrieval-Augmented Generation (RAG) is gaining popularity in the AI community. It involves layering GenAI into production and using a curated knowledge base to ground the LLM-powered agent in accurate information. Developers are excited about the API access to the first web index built for LLMs and the opportunity to work with open-source tools for RAG implementation. Enterprise adoption of LLMs has increased, but building custom LLMs based on a KnowledgeBase remains challenging. Open AI announced retrieval APIs, but there are concerns about cost, customization, ease of scaling, and compatibility with different LLMs, including open-source ones.