A Scalable Communication Protocol for Networks of Large Language Models
This work addresses the problem of scalable and efficient communication for networks of AI agents, enabling unprecedented scalability and automation in collaborative AI systems.
The paper tackles the challenge of enabling efficient and scalable communication in networks of large language model (LLM)-powered agents, known as the Agent Communication Trilemma, by introducing Agora, a meta protocol that combines standardized routines, natural language, and LLM-written routines to achieve full decentralization and minimal human involvement, resulting in the emergence of self-organizing, fully automated protocols that solve complex problems without human intervention.
Communication is a prerequisite for collaboration. When scaling networks of AI-powered agents, communication must be versatile, efficient, and portable. These requisites, which we refer to as the Agent Communication Trilemma, are hard to achieve in large networks of agents. We introduce Agora, a meta protocol that leverages existing communication standards to make LLM-powered agents solve complex problems efficiently. In Agora, agents typically use standardised routines for frequent communications, natural language for rare communications, and LLM-written routines for everything in between. Agora sidesteps the Agent Communication Trilemma and robustly handles changes in interfaces and members, allowing unprecedented scalability with full decentralisation and minimal involvement of human beings. On large Agora networks, we observe the emergence of self-organising, fully automated protocols that achieve complex goals without human intervention.