How about people who want to spam LLM output just provide their input alongside the output? I'd be happy to read their input.