Armchair Architects: Large Language Models (LLMs) & Vector Databases

Just to clarify, because a lot of people misunderstand how RAG works, when you say:

PostgreSQL teams have already added a plug-in into PostgreSQL where you can take any text field, turn it into a vector and then you can take that vector and embed it into an LLM.

You don't embed vectors into a LLM. You use them to retrieve relevant context from a database, then the associated text is added to the LLM prompt. Not the vectors themselves.