of course, any compression or cropping would make it pointless, but it's the thought that counts
I've been thinking about a timeline feed that uses an llm for this sort of thing — at current llama pricing you could process 10k+ posts for under a dollar