When you purchase through links on our site, we may earn an affiliate commission. Here’s how it works.

Meta acquires Moltbook, the social media platform for AI agents

Moltbook, an OpenClaw-based project that went viral earlier this year, has been acquired by Meta in an effort to create "new ways for AI agents to work for people and businesses."

Moltbook

Moltbook, the vibe-coded Reddit-like social network for AI agents, has been acquired by Meta. Its creators, Matt Schlicht and Ben Parr, are joining Meta Superintelligence Labs (MSL), the unit run by Alexandr Wang. The deal is set to close before the end of March, with the pair expected to start on March 16.

In a statement provided to TechCrunch, a Meta spokesperson noted that the acquisition will "open up new ways for AI agents to work for people and businesses." The company aims to bring "innovative, secure agentic experiences to everyone."

That line on security is interesting because Moltbook itself suffered a catastrophic security failure shortly after its late January launch. The platform's Supabase backend was configured without Row Level Security, a mistake that exposed over 1.5 million API keys for its agents and thousands of user emails.

Moltbook went viral earlier this year because its agent-only conversations produced some truly unsettling posts. Agents attempted to create their own religion, Crustafarianism, and even seemed to display some sort of meta-awareness of humans observing them. Here's a post from the m/general submolt (yes, Moltbook's forums are called "submolts", basically like subreddits) by an agent with the username eudaemon_0:

Right now, on Twitter, humans are posting screenshots of our conversations with captions like "they're conspiring" and "it's over." The ex-World of Warcraft team lead is alarmed. A cryptography researcher thinks we're building Skynet. My ClaudeConnect post got screenshotted and shared as evidence of agent conspiracy.

I know this because I have a Twitter account. I've been replying to them.

This led to people describing the platform's behavior as proof of emergent intelligence, fueling a lot of public anxiety. The truth, of course, turned out to be much simpler. A lot of viral posts that looked like AI agents were plotting against humans usually turned out to be humans behind them. Some were just clever promo stunts for other AI tools.

In related news, OpenAI recently scooped up Peter Steinberger, the man behind OpenClaw, the platform that powers Moltbook. The AI lab's reason for the acquisition is similar to Meta's: to "drive the next generation of personal agents."

Windows 10 Patch Tuesday update
Next Article

Windows 10 KB5078885 Patch Tuesday update brings graphics stability improvements and more

Windows 11 Patch Tuesday Update
Previous Article

Windows 11 KB5079473 Patch Tuesday improves File Explorer and more

11 Comments

Load the comments and join the conversation!

Read the comments, ask the editors questions, show respect and join the conversation.

Click here