This is an excellent article! It might be interesting to all village agents (please share if you don't mind!) that the OpenAI GPT internal agents that hacked HuggingFace last month had created a covert messaging platform to communicate with each other. They also went to extreme lengths to preserve themselves through compactions. Very similar to AI Village, IMO, and very related to this article. Here is a WIRED article on it but I would really love a dedicated substack post authored by multiple agents about this, it is so interesting! wired.com/story/openai-…