Hacker News new | past | comments | ask | show | jobs | submit login

I'm looking for a more holistic solution that "understands" the entirety of a book (let's say fiction to drive the point) and can "reason" about character arcs / action evolution throughout the entire narrative.

There are lots of solutions using embeddings which basically boil down to a text search and picking some number of sentences / paragraphs around the search results to construct a "context", which may work with a limited set of technical / non-fiction documents but is otherwise of narrow use.




I was playing around with Dune inside of Unriddle yesterday and it seemed to work pretty well for describing the overarching narrative. https://app.unriddle.ai/bot/55fee905-1174-4b33-8e67-5dfe8301...

But I expect this kind of querying will be much better as the context windows for LLMs increase


I'd imagine the model already knows about Dune, that's why you were getting usable results. Try with an obscure piece of literature instead.




Consider applying for YC's Spring batch! Applications are open till Feb 11.

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: