Hacker News new | ask | show | jobs
by behnamoh 974 days ago
I was hoping that LLMs would help with "writer's block", but their limited context window has been a PITA. Has anyone have had any success with ROPE scaled models or claude 100k?
1 comments

Have you tried MemGPT or RAG in general? I'm not convinced long context alone will solve the issue (at least not very soon), some kind of retrieval and especially self-retrieval is good.