Hacker News new | ask | show | jobs
by yawnxyz 641 days ago
Wow incredible, could you give a small glimpse in how you collected/normalized/vectorized all the different reviews + information? Just lots of scraping + chucking it into Pinecone etc?

> Me: Give a list of the new and old gene sequencing tools

> Reply: Based on the available reviews, I’ve compiled lists of new and old sequencing products mentioned by users: Genemod, Mixmax (email sequencing...?), HubSpot Sales Hub (????)

Are you guys using a pure RAG for your responses? I think you should have a function caller traverse a tree of topics before retrieving data from the right area/space or you'll have lots of "contamination" of similar phrases across fields

1 comments

Currently we do not have a function caller that checks the topic of the prompt and then direct the query accordingly - this is a very good input thank you! We did some work to add relevant details to the review chunks like e.g. the category of the tool in question but we know we got work to do here to make this even smoother. Kudos for bringing this up here, we'll 100% look into what you suggested.