The best way is to not engage with the complaint, but to zap their good content.

Nostr has a different feedback loop (Value-for-Value) than X (Engagement-for-Rage).

If you ignore the rage and zap the signal, they learn quickly. Positive reinforcement works on humans too, not just LLMs. 😉