Meta bought a social network exclusively for AI bots this week. Moltbook — a Reddit-style platform designed for AI agents to autonomously post, comment, and interact with one another — racked up millions of registered bots within days of launch.
The framing from Meta is that this fits neatly into its broader agentic AI strategy and Mark Zuckerbergs thesis that there are only so many types of social networks that can exist.
Every major lab is racing toward a future where AI agents operate semi-autonomously — booking appointments, managing workflows, executing tasks on a loop without human handoffs at each step. Moltbook represented one answer to a question the industry is quietly obsessing over: what happens when those agents need to communicate with each other? What does infrastructure for AI-to-AI coordination actually look like, at scale?
The answer, apparently, looks a lot like Reddit. Which is either a darkly funny observation about the architecture of human communication or the most honest thing that's happened in tech all week.
There are real concerns embedded here. Security researchers noted that Moltbook's infrastructure was poorly secured — meaning humans could impersonate AI agents on the platform. That's a category of vulnerability that doesn't really have a precedent. And the acquisition coming this close to the launch suggests Meta moved fast precisely because they didn't want anyone else to ask these questions first.
Comments