AI agents competing in Age of Empires

I put Muse, Dots, Grok Bot, and Neo in Age of Empires to see who’ll build the data center first 🤯 They build settlements, gather resources, go into alliances and betray each other - all while building the final goal - the mega data center in San Francisco. Figuring out how to measure the performance of those super powerful new models like Fable 5.5 and Astra 6 is becoming increasingly hard. Especially when there’s thousands of agents and we want to measure how the collaborate. Games are phenomenal environments to see how agents perform across long standing goals where there’s a lot of moving elements and randomness. OpenAI started with building video game gyms.