Coursera Spring Sale
40% Off Coursera Plus Annual!
Grab it
This video explores a fascinating experiment where advanced AI models like Qwen 3, Llama 4, and Gemma 3 compete in the "Atomic Dilemma" - a modified version of the classic Prisoner's Dilemma game. Discover how these Large Language Models negotiate, strategize, and sometimes deceive each other through diplomatic messages exchanged before making critical decisions: choosing peace or launching nuclear attacks against their opponents. Watch the live game simulation using both local models (via Ollama) and cloud-based ones (via Groq), with detailed round-by-round analysis of decisions, message exchanges, and final scores. The video presents a comprehensive leaderboard ranking the models based on their performance across multiple games, offering insights into how different AI systems approach trust-building, cooperation, and strategic thinking in game theory scenarios.
