Hacker News new | ask | show | jobs
by ammario 649 days ago
To scale such an approach you could have the LLM generate JS to walk the DOM and extract content, caching the JS for each page.