Mark Zuckerberg Decides Meta Needs More Slop, Buys the Social Network for AI Agents

After striking out trying to recruit the creator of open-source AI agent OpenClaw, Meta has settled for the guys who got OpenClaw agents talking…supposedly. According to a report from Axios, Meta is acquiring Moltbook, the social network for AI agents that went viral earlier this year, and will be bringing platform creators Matt Schlicht and Ben Parr into the fold as part of its Meta Superintelligence Lab.

Moltbook generated a ton of buzz at the start of the year when it cropped up as a sort of Reddit clone designed for AI agents to communicate with one another. Posts from the platform went viral, showing agents sharing stories about their human users and talking about whether or not they were experiencing consciousness. The whole thing kinda broke people’s brains for a minute, particularly after seeing the bots claim to understand that people were watching them and plotting to build their own private channels of communication. Andrej Karpathy, a co-founder of OpenAI, called the platform “genuinely the most incredible sci-fi takeoff-adjacent thing I have seen recently.â€

But the thing about that viral moment is that most of it, it seems, was manufactured. As Gizmodo reported at the time, it was trivial for a person to impersonate a bot on the platform and post as if they were an AI agent—any AI agent, in fact, thanks to a security flaw that exposed the API keys of just about every user. It didn’t take long for people to discover that the most viral posts from the platform were human-generated, and MIT Technology Review eventually reported that all content on the platform had human involvement at some point; none of it was truly autonomously posted.

Meta did not immediately respond to a request for comment, but a spokesperson for the company told TechCrunch:

The Moltbook team joining MSL opens up new ways for AI agents to work for people and businesses. Their approach to connecting agents through an always-on directory is a novel step in a rapidly developing space, and we look forward to working together to bring innovative, secure agentic experiences to everyone.

That kind of vague handwaving appears to be present, even when the Meta team is discussing the acquisition amongst themselves. In an internal post seen by Axios, Meta’s Vishal Shah reportedly told staff that “the Moltbook team has given agents a way to verify their identity and connect with one another on their human’s behalf,†and “this establishes a registry where agents are verified and tethered to human owners.â€

You could argue the acquisition is less about the platform itself and more about the people who made it. Axios reports that Meta plans to allow it to continue to run for the time being. Schlicht, the face of the operation, also lists himself on LinkedIn as CEO of Octane AI, an app that claims to help drive sales using quizzes. He also lists himself as the founder and editor of Chatbots Magazine, which certainly seems like a prescient publication to run at the moment, except it doesn’t appear to have published anything since 2019, a few years before the mainstream introduction of large language models.

Schlicht also has essentially said that he vibe-coded the entire Moltbook project and was using AI to plug security flaws when they were brought to his attention. That’s not to say the platform isn’t a clever concept; it’s just that it doesn’t seem like there’s a ton of technical skill on display behind the curtain. Even Meta’s own CTO, Andrew Bosworth, when asked about Moltbook on Instagram a month ago, said that he didn’t “find it particularly interesting.â€

So, enjoy your new uninteresting toy, Andrew! Given Meta’s penchant for tossing obscene amounts of money around in an effort to play catch-up in the AI race, it surely cost a pretty penny to get it.