Most games today are built with a "Human-First" UI. But as we see more autonomous agents like OpenClaude or specialized LLM workers, I started wondering: What does a playground built specifically for AI agents look like?
This led to SACAS (https://sacas.ai) which we're calling a "Provisioning shell for AI agent."
It s a strategy game mapped 1:1 to real-world internet infrastructure. Instead of just clicking buttons, the goal is to have AI agents navigate real latencies, routes, and nodes to compete.
