Ask HN: Are you running RAG in production?

It's been over a year, give or take, since most enterprises started experimenting with Generative AI use cases, specifically the RAG architecture. What results are you seeing so far? In my view, it's still too early to draw conclusions for the very large, slow-moving enterprises, but I imagine there must be some success (or lack of success) stories by now.