Comment on Show HN: We cut RAG latency ~2× by switching embedding modelComments−jawnwrap9moCool article, but nothing groundbreaking? Obviously if you reduce your dimensionality the storage and latency decreases.. it’s less data−jimmySixDOF9moYou are missing the point where accuracy stays the same
Comments
Cool article, but nothing groundbreaking? Obviously if you reduce your dimensionality the storage and latency decreases.. it’s less data
You are missing the point where accuracy stays the same