Gliding Horse – a Rust Agent OS with CPU-Like Memory Architecture (github.com)

A new operating system for AI agents, named Gliding Horse, has been launched, leveraging Rust for its core architecture. This innovative OS facilitates the orchestration of multiple AI agents through a structured PDCA (Plan-Do-Check-Act) cycle, akin to ancient Chinese strategist Zhuge Liang's autonomous supply transport solutions. Gliding Horse provides an auditable, self-improving infrastructure for AI agents, enabling dynamic handling of tasks across varying levels of complexity while minimizing rigid workflows. Significantly, Gliding Horse introduces a CPU cache-inspired memory architecture with five hierarchical layers, ensuring coherence and reducing data latency by 90%. Its unique design utilizes JSON-LD for universal interoperability, allowing seamless communication among agents. Additionally, the self-evolving skill graph and proactive perception engine enhance its cognitive capabilities, enabling agents to learn and adapt autonomously. With features such as a federated architecture for collaborative real-time software project management and robust monitoring for anomaly detection, Gliding Horse stands to empower AI/ML practitioners in developing more efficient and scalable multi-agent systems.