• Login
  • Register

Work for a Member company and need a Member Portal account? Register here with your company email address.

Project

Large Population Models: Orchestrating a Billion Agents

Our world stands at the cusp of an unprecedented transformation. As billions of AI agents begin to augment human society, we face a fundamental challenge: coordinating collective behavior at scales far beyond human cognitive limits. While current AI approaches like large language models excel at enhancing individual agent capabilities (Personal AI), they cannot inherently address the massive-scale coordination challenges that define our interconnected world (Population AI). Enter Large Population Models (LPMs): a new AI paradigm simulating millions of interacting agents simultaneously, capturing collective behaviors at population scale. It's like scaling up AI agents exponentially to understand the ripple effects of countless decisions.

Copyright

Cite Ayush Chopra and the AgentTorch project before using this picture

Through Large Population Models (LPMs), we introduce a paradigm shift - from maximizing individual intelligence to discovering protocols that enable beneficial collective behavior at population scale. Using commodity hardware, LPMs can execute 300,000 interactions per second and scale to 60 million agents per GPU, achieving dramatic improvements over conventional approaches.

Copyright

MIT Media Lab and AgentToprch