Just hinking about the way LLMs can analyze huge amounts of data to cross references and even use your writing style to dox you… well… seems like we gonna need some text spoofer tools that change each of your messages to a different grammar set and random typo generation so they all seem like sent by different users…