Super interesting interview. Summary: - Predicts tokens get 10-1000x cheaper and that creates new product categories. sail sells open-source LLM inference at claimed unbeatable price/token + long-running agent sandboxes - The market optimized for chat latency (cursor pulled it…↗