Post #4402328
2026-06-10 18:21 UTC
How this actually works. When an LLM "searches the web," it is not crawling the whole goddamn internet itself. ChatGPT rents Bing's index and bolts on its own live crawler, OAI-SearchBot.
Translation: it inherits every shitty garbage-in problem Bing's index already has. Including 20 years of SEO spam nobody ever bothered to clean up.
Replies (1)
-
@k3ym0@infosec.exchange 2026-06-10 18:22
Ask Silver found cloned Russell & Bromley and Dunelm sites surfacing as sources in ChatGPT shopping answers. The quoted take was "the LLM got poisoned," implying training data. That's bullshit, and the timeline proves it. R&B went into administration in Jan 2026 and lost its standalone site. The fake domains are fresh, stood up after that.