Automating Responses to Real Estate Spam with LLMs

The author built a system using LLMs to automatically respond to spam text messages from real estate brokers. The system involves modifying the Android-SMS-Gateway-MQTT app for bidirectional MQTT communication. A Python script listens for incoming texts via MQTT, uses an LLM to generate responses based on pre-defined personalities, and stores conversation context for coherence. Ollama is used for convenient experimentation and personality adjustments. The author shares screenshots of amusing interactions but also notes legal and security considerations.