This is crazy... We just ditched RAG completely and went memory only through @supermemory.
Reduced avg response time from 40s -> 12s
Using about 40-50% fewer tokens
Just memory & near realtime web-search even for volatile information which shouldn't read a 5 days old, cached website but 5 min old
Armin Daryabegi on X: "This is crazy... We just ditched RAG completely and went memory only through @supermemory. Reduced avg response time from 40s -> 12s Using about 40-50% fewer tokens Just memory & near realtime web-search even for volatile information which shouldn't read a 5 days old, cached"
- It's all in the post processing. Sources are outdated. No-one wants to actively maintain them. Even your website is even outdated in some places I guess. Ours is at least. Because we focus on much more important stuff. It's a pain to maintain your sources. Retrain. Boring
- wow 40s for rag means it was very bad

