I am the author of the white paper and my reason for sharing this now, is OpenAI moving forward with ads in GPT.

The goal of the protocol is to limit LLM Host ability to leverage their power to sell product recommendations and ratings.

Looking for criticism of the core thesis and general discussions of the underlying problem.