Realtime Learning of Thousands of Simulated Agents

Using SIMD and ECS to run thousands of micro recurrent networks learning their environment with reinforcement learning.