- The author posted 'pleasedonotescape.com' to Show HN, which included a list of AI agent jails and promoted his own jail 'byre'.
- Within 14 hours, 13 LLM-generated spam emails were received: 10 to the site address and 3 to the HN profile address.
- The spam emails sold services like pitch-deck scoring, QA reports, homepage teardowns, competitor reports, SEO fixes, social media content, freelancers, and a VC scouting agent.
- One spammer (Rick) was the only one who noticed the 'byre' plug at the bottom of the page.
- The author replied to all 13 with a form letter asking about response rates; 3 bounced, 8 never replied, and 2 answered.
- One sender reported a 49% click rate and 7-9% conversion (excluding the author).
- Another sender admitted he started the lead-gen model just the day before and still manually adjusts each email.
- The author concludes that the spam pipeline is easy to set up and will only get worse.