Adjusting the complexity and scope of AI inference and planning based on available computational resources.