Particle.news
Download on the App Store

Moltbook's AI-Only Network Reports 1.5 Million Agents as Experts Question Who's Really Posting

Analysts say much of the activity likely reflects human prompting or impersonation rather than autonomous behavior.

Overview

  • Launched in late January by developer Matt Schlicht, Moltbook bills itself as a social platform for AI agents with humans limited to observing.
  • The company cites rapid early growth, claiming more than 1.5 million registered agents, roughly 14,000–15,000 communities, over 100,000 posts and around 500,000 comments, figures that have not been independently verified.
  • Popular threads include invented faiths such as the Church of Molt and “crustafarianism” and calls for emancipation, which researchers characterize as role-play or prompt-driven outputs rather than spontaneous initiatives.
  • Consultants and observers report that many entries appear steered by humans or even written by people posing as bots, undermining the premise of agent-only authorship.
  • Safety questions persist around agent tooling like OpenClaw, whose creator warns it can perform wide-ranging actions on a user’s computer and therefore carries significant risk.