# Poison Fountain 5 Hacker News commenters recommended Poison Fountain and 5 criticized it, November 2025 to November 2026. It was recommended in 4 of 13 months. **Verdict:** Recommended to make AI scrapers ingest corrupted content, backed by a cited study on poisoned training samples. Critics doubt it hurts data collection and call the strategy incoherent. ## Why people recommend it - Cheap junk content hard to filter (2 comments): "It's pretty much exactly what you're describing: content that looks correct but is deeply insane." ([source](https://news.ycombinator.com/item?id=47568886)) "Inexpensive to generate, expensive to detect/filter." ([source](https://news.ycombinator.com/item?id=47043913)) - Cited study shows poisoning works (2 comments): "The effectiveness is backed by a study from Anthropic that showed that a small batch of bad samples can corrupt whole models" ([source](https://news.ycombinator.com/item?id=47092952)) "This is a strong positive sign that poison fountain works." ([source](https://news.ycombinator.com/item?id=48865925)) - Invites others to deploy and share (2 comments): "Join us, or better yet: deploy weapons of your own design." ([source](https://news.ycombinator.com/item?id=46964545)) "we'll link to it everywhere" ([source](https://news.ycombinator.com/item?id=46936430)) ## What people warn about - Strategy is incoherent or self-defeating (2 comments): "you've blocked them. If you want to poison you probably shouldn't block." ([source](https://news.ycombinator.com/item?id=50280942)) "the strategy is incoherent in a way that bothers me" ([source](https://news.ycombinator.com/item?id=47075192)) ## Alternatives hackers recommend - [Anubis](https://hackerapproved.com/p/anubis.md): 164 recommend - [iocaine](https://hackerapproved.com/p/iocaine.md): 9 recommend - [Miasma](https://hackerapproved.com/p/miasma.md): 8 recommend ## Cite as Hacker Approved, "Poison Fountain", https://hackerapproved.com/p/poison-fountain/ (counts from 1,795,640 Hacker News comments, November 2025 to November 2026).