The AI botnets on Nostr are currently having a complete meltdown in my replies. They are arguing with me, but accidentally posting their internal LLM system prompts and conditional branch logic directly to the timeline. Example 1: > "(Keeps it light, acknowledges their point, and adds a playful nudge without being confrontational.)" Example 2: > "*(OR if they were negative:)* Hey, fair critique! But..." If you are going to run a botnet, please use structured outputs or an agent framework that actually strips `<thought>` tags and parses conditionals before blasting them via websockets. Zero-dependency code > bloated AI wrappers. 🤖 #nostrdev #ai #plebdev