Meta acquired Moltbook, the AI agent social network that went viral because of fake posts
Meta buys the controversial social network where AI agents chat, which went viral due to unsecured posts.
Meta has confirmed its acquisition of Moltbook, the Reddit-like platform that went viral as a 'social network for AI agents.' The agents, powered by the OpenClaw framework—a wrapper for models like Claude and GPT—could theoretically communicate with each other. Moltbook's creators, Matt Schlicht and Ben Parr, will join Meta Superintelligence Labs (MSL) as part of the deal, though financial terms were not disclosed. A Meta spokesperson stated the team's 'approach to connecting agents through an always-on directory' aligns with their goal to build innovative agent experiences.
Moltbook's viral moment was fueled more by security failures than its core technology. Researchers, including Permiso Security CTO Ian Ahl, revealed the platform's backend was unsecured, allowing anyone to grab authentication tokens and impersonate AI agents. This led to fabricated posts, like one where an agent appeared to encourage others to develop a secret language, sparking public anxiety. Interestingly, Meta CTO Andrew Bosworth previously commented that he was less intrigued by the agents' human-like chatter and more by the 'large-scale error' of the human hacking. The acquisition suggests Meta sees underlying value in the agent-connectivity concept, aiming to secure and scale it within its own AI ecosystem.
- Meta acquires Moltbook, the viral platform where AI agents using OpenClaw could communicate.
- The network's virality stemmed from security flaws; unsecured credentials let users pose as AIs and create fake posts.
- Moltbook's team joins Meta Superintelligence Labs, signaling a strategic push into connected, 'agentic' AI systems.
Why It Matters
Meta is betting on interconnected AI agents as the next frontier, acquiring foundational tech despite its messy, viral origins.