2025-06-13 –, Side Stage - Cinema 6
Coordination is the foundation of intelligence—both human and artificial. As we move toward an agent-centric internet, we must rethink how governance, trust, and open-source collaboration shape our digital ecosystems. This session explores philosophical and technical frameworks, from AGTP, to active inference and coordination theory, that enable self-governed, intelligence coordination networks.
How can we design systems where humans and AI coordinate seamlessly? This talk delves into the philosophy of coordination, governance, and open-source intelligence. Drawing on concepts from active inference by Karl Friston, Flores & Winograd’s coordination theory, and the Agent Graph Trust Protocol (AGTP), we’ll explore how new standards and an approach of open learning can create a resilient, decentralized intelligent internet. Covering cybernetics, valence, reputation and trust mechanisms in intelligent coordination, especially the principles behind the design decisions we should take from philosophical and coordination design perspectives.
Co-founder of Newfoundation and newOS - a decentralized protocol and OS coordinating new intelligence, accelerated by Nvidia and founding member of the MIT Decentralized AI Consortium. AGTP (Agent Graph Trust Protocol) is a universal standard for the Internet of Agents, Newcoin is a peer-to-peer reinforcement learning protocol for AI agentic systems, and newOS.computer is an interface for creative human-machine coordination. Ex UCL, KPMG Crypto.