wait so you still have to run every api call thru openai after the embeddings are done? what is pinecone doing then? seems like an unreliable dependency

→ View original post on X — @swyx