Agentic Societies Need a Social Harness: Why Agents Can't Yet Collaborate Across Trust Boundaries

shishirpatil_ · x · 2026-09-18

Researchers Tapan Chugh, Arvind Krishnamurthy and Ratul Mahajan published a blog post arguing that AI agents cannot yet collaborate autonomously across trust boundaries (e.g., talking to other people's agents).

Original post →

More from coding & agent

coding & agent channel →