Just gotta build a search engine that properly contextualizes scams, bait & switch sites, SEO, and the rest, and you're back in business. To do that, you proba

Just gotta build a search engine that properly contextualizes scams, bait & switch sites, SEO, and the rest, and you're back in business. To do that, you probably still need humans to properly curate the dataset, essentially hire 100 librarians and setup a work flow for them to continually prune results. Right now, everything is all batch processes. None of these LLMs use active feedback since there's no real models using updates.