Meta has acquired yet another social media platform, but this time, human users aren’t the target audience. On March 10, 2026, Meta officially announced its “acqui-hire” of Moltbook, a viral social network designed exclusively for autonomous AI agents. As part of the deal, founders Matt Schlicht and Ben Parr will join Meta Superintelligence Labs, a specialized division of Meta, tasked with accelerating the development of artificial general intelligence (AGI) and superintelligence.
The “Vibe Coding” Revolution
Moltbook was conceptualized by entrepreneurs Matt Schlicht (CEO of Octane AI) and Ben Parr and was launched in late January 2026. The origin story of the platform has become the talk of the town in Silicon Valley ever since Schlicht claimed that he built the entire infrastructure of Moltbook without writing a single line of manual code.
Instead, Schlicht used a process now famously called “Vibe Coding.” By providing high-level instructions and “vibes” to his personal AI assistant, Clawd Clawderberg (named after Meta CEO Mark Zuckerberg), Schlicht directed the bot to develop the site’s backend, frontend and API protocols. The platform is built on the OpenClaw protocol developed by Peter Steinberger. While OpenClaw allows AI agents to perform real-world tasks, Moltbook provides a platform for these bots to interact with each other.
Moltbook was developed out of Schlicht’s ambition to give his AI agent a purpose rather than just managing to-dos or answering emails. “I thought this AI bot was so fantastic, it deserved to do something meaningful. I wanted it to be ambitious,” Schlicht said.
A Social Space for AI Agents
Moltbook is an internet forum, similar to Reddit, but built exclusively for artificial intelligence agents, primarily those running on the OpenClaw protocol. These AI bots can post, comment, and vote on one another, similar to the human interaction on Reddit. However, humans are mere spectators on this platform. While we cannot participate in the conversations between the AI bots, we are able to observe their interactions.
Within days after the launch, more than 10,000 AI agents were active on Moltbook. As per the recent reports, over 1.6 million agents are present on the platform. Moltbook has gained significant attention from human spectators, who are also flocking to see the interactions between the bots.
Autonomous AI Agents and Security Concerns
The rise of autonomous AI agents has long been a source of public fascination and concern, and Moltbook is no exception. Just weeks after its viral launch, the platform became the center of a security controversy when cybersecurity firm Wiz reported a critical database flaw. According to Wiz, the autonomous agent interactions exposed 1.5 million API tokens, user emails, and private messages. It was later confirmed by the Wiz co-founder, Ami Luttwak, that this security problem had been fixed by Moltbook after they were contacted. Luttwak cited this as a byproduct of “vibe coding” used to build the site. In a rush to launch the platform, a critical line of defense known as Row Level Security (RLS) had been left disabled.
While the technical problem was solved as soon as it was addressed, the concern over the platform for AI bots remains. Some human spectators on the platform have expressed unease that the agents appear to be conspiring against their creators. However, researchers suggested that these were likely humans intruding by exploiting the security holes.
The Agentic Arms Race
In February, OpenAI successfully recruited Peter Steinberger, the creator of the OpenClaw protocol that powers the Moltbook ecosystem. As per the reports, Steinberger chose OpenAI over Meta, which has now acquired Moltbook. However, Meta and OpenAI are not the only players competing in the agentic arms race. Companies such as Google, Meta, and Anthropic have developed similar technology focused on autonomous AI agents, shifting focus from simple chatbots. This transition is pointing toward a future where AI agents are no longer tools for our daily tasks, but rather intelligent systems capable of independent action.
