Building knowledge bases from YouTube data using LLMs -- my workflow after 52 guides
About this article
I've been building a system that turns YouTube channels into structured knowledge bases. Thought I'd share the workflow since Karpathy's recent post about LLM knowledge bases The problem I was solving: I wanted to extract structured, cross-referenced knowledge from YouTube channels. Not just transcripts, but the full picture: what the creator teaches, what the audience actually cares about, what questions remain unanswered, and what patterns emerge across multiple channels, playbooks, etc. Da...
You've been blocked by network security.To continue, log in to your Reddit account or use your developer tokenIf you think you've been blocked by mistake, file a ticket below and we'll look into it.Log in File a ticket