Amusingly, you can use LLMs to analyze a user's posting history to try to determine if they're a shill.