Nuclear Codes: A ChatGPT Multi-Agent Simulation

AI agents hold parts of "nuclear codes" and must decide whether to collaborate or deceive each other. The idea is to explore the complexities of AI-driven social interactions in real-time scenarios.

Technical Specs: Built with Python and visualized with a Node.js server, the simulation leverages the OpenAI API to enable study multi-agent AI decision-making. Outputs a streaming Mermaid diagram that can be visualized.

Goals: To study real-time AI decision-making in high-stakes scenarios. To offer a framework for further exploration in AI behavior. How to Contribute: Hosted on GitHub and open for contributions. Fork, submit issues, or propose enhancements to dive deeper into this compelling world of AI, trust, and deception.