vector database

AI memory is not one thing. This guide separates context windows, retrieval (RAG), vector databases, and long-term memory in ChatGPT and Claude.

Vector databases store embeddings and run similarity search via ANN indexes. Pinecone, Weaviate, Qdrant, pgvector — compared for 2026.

RAG is the technique making AI tools actually useful for real work — letting models look things up instead of relying on memory. Here's how it works and why it matters.