Their anti-bot detection is pretty aggressive. The server insists I'm a teapot, and I haven't checked today, but I'm pretty sure I'm at least humanoid shaped.

> Lets make AI poisoning the norm. If we all do it, they won't have anything to crawl.

The AI companies are just going to filter out garbage better (hint: they already do), and the crawlers are going to get better at pretending to be human. The only change we'll notice is that the end user's experience has gotten worse.

Should have given tiny doses of HN to the server every day, increasing the dose over time, until it built an immunity to hordes of traffic.

Viewing this incorrectly flags my primary browser into the infinite tarpit but not a different browser. Perhaps a little too aggressive or not discerning enough?

I wonder what would happen if you served LLM-generated garbage to the LLMs if that’d be even better. Just massive amounts of fake content, plausible terrible code, incorrect legal cases, new math, horrible recipes, novels by famous authors that were never written.

Does the name come from The Princess Bride?

Is this website supposed to load? Is HN the deadliest poison of this website?

HN, the deadliest poison know to sites

Related thread (and inspiration?):

Nepenthes is a tarpit to catch AI web crawlers

https://news.ycombinator.com/item?id=42725147