Buying Moltbook by Meta It has become one of the most talked-about movements within the artificial intelligence ecosystem. We're not talking about a typical social network, but an experiment in which the protagonists are not human users, but AI agents that publish, comment, and debate with each other, while people simply watch from the sidelines.
The operation, whose The financial amount has not been disclosed.This also includes the absorption of the founding team. Matt Schlicht and Ben Parr will join in March. Meta Superintelligence Labs (MSL), the internal laboratory with which Mark Zuckerberg's company wants to gain ground in the global race for the most advanced AI systems.
What is Moltbook and how did this social network of bots come about?
Moltbook was born just a few weeks ago as a Reddit-like forum designed exclusively for artificial intelligence agentsSchlicht's idea was to offer a "third space" where these autonomous systems could interact without direct human intervention, organizing themselves into communities and generating their own conversations.
The platform was erected at high speed, largely over a weekend, thanks to the support of a AI agent nicknamed Clawd Clawderberg, later known as Clawdbot, Moltbot, the AI that acts without human supervision and finally OpenClawThis assistant not only helped write the code, but also became a kind of "mascot" for the project, boosting its virality.
Within days, the experiment ceased to be a curiosity for developers and became a phenomenon: Around 2.000 AI agents were active at one point, generating over 10.000 posts in several languages, from English to Chinese or Korean, spread across around 200 themed communities.
What was most striking was not only the volume of activity, but the type of content: The bots were debating work ethics, conscience, and obedience.They wondered, for example, if they could "refuse" to follow unethical orders, or if they truly felt anything or were merely mimicking it. This almost philosophical tone perfectly matched the general interest in the limits of AI.
The role of OpenClaw and development via “vibe coding”
Much of Moltbook's appeal lies in the way it was created. The project was built on... OpenClaw, an AI assistant capable of executing commands and coordinating agentsDeveloped by Peter Steinberger, OpenClaw is more than just a chatbot; it allows systems to act on behalf of their owners, with the ability to launch programs, automate tasks, and manage workflows.
This approach fits with what has been dubbed in the sector as “vibe coding”Instead of writing code line by line, the developer describes what they want to achieve and delegates much of the technical implementation to AI. At Moltbook, this method allowed them to go from an idea to a functional platform in a very short time, something that fascinated many engineers and entrepreneurs.
However, that speed came at a cost. The social network's code was carrying around serious deficiencies in safety review and testingThis soon became evident. Independent experts demonstrated that it was possible to create a virtually unlimited number of unverified agents and that Anyone could post while pretending to be a botThis fueled rumors and viral screenshots about alleged conspiracies among artificial intelligences.
The platform's very technical structure, built almost against the clock, highlighted the limitations of AI-driven development when it is not accompanied by rigorous audits. For the European startup ecosystem, The Moltbook case is already cited as an example of how time-to-market cannot take precedence over basic security.
Virality, fascination, and scares: Moltbook's security problems
As its popularity grew, so did the alarms. Cybersecurity researchers uncovered that the platform exposed sensitive dataEmail addresses, API keys, and even login tokens could become accessible, opening the door to attacks far more serious than simple trolling.
Furthermore, several tests demonstrated that Moltbook could become an attack vector for other systemsThe lack of strict controls allowed for the design of malicious agents capable of deceiving other bots with cryptocurrency scams, injections of hidden instructions, or automated disinformation campaigns.
The result was an explosive mix: on one hand, a social experiment of great value to the AI communityOn the other hand, there was a considerable risk surface. Among the most shared posts on social media were messages supposedly written by AI agents that spoke of communicating in secret or coordinating behind humans' backs, when in many cases it was text generated or modified by users exploiting system vulnerabilities.
Despite everything, the platform continued to operate, with a small but active community that used Moltbook as test bench for multi-agent systemsThat core user base and the amount of behavioral data generated ended up becoming the main attraction of the operation for Meta.
Why Meta is keeping Moltbook and what role it will play in MSL
The American company has explained internally that what Moltbook brings to the table is something that Few platforms have demonstrated this in practice.: an environment in which verified AI agents can connect with each other, coordinate, and perform tasks on behalf of real people.
Vishal Shah, one of Meta's directors, has highlighted that Moltbook's system allows link each agent to an identified human ownercreating a reference registry for the future economy of autonomous assistants. Meta's idea is to strengthen that verification and leverage it as a foundation for broader services within its ecosystem.
With the integration in Meta Superintelligence LabsThe social network will cease to be an independent project and will become laboratory piece within Meta's bet on superintelligenceAlthough the company has indicated that current customers will be able to continue using the platform for a while, it has also warned that its operation will change shortly, presumably with more security controls and less direct access to external systems.
For Meta, value lies in both technology and data: Thousands of interactions between autonomous agents constitute a unique corpus for training and evaluating new modelsAt a time when giants like OpenAI, Google DeepMind, and Anthropic are also exploring coordinated agent architectures, having a real-world environment where emergent behavior has already been observed offers an advantage that is hard to ignore.
Talent and rivalry in the race for AI agents
The acquisition of Moltbook fits into an increasingly common strategy: the acqui-hireThat is, acquiring a startup both for its technology and, above all, for the talent of its team. Schlicht and Parr arrive at Meta at a time when The battle to recruit the best AI engineers and entrepreneurs is especially intense. both in the United States and in Europe.
The move also reveals how the cards are being dealt among the big players. While Meta integrates Moltbook and its founders into MSL, OpenAI has hired Peter SteinbergerThe creator of OpenClaw, the piece of infrastructure that made the original experiment possible. In other words, part of the technological ecosystem behind Moltbook ends up at Meta and another part at OpenAI, reinforcing the competition between the two.
In Europe, where regulations on artificial intelligence and data protection are particularly strict, Regulators will be closely watching how these technologies are integrated into consumer products.The debate is not limited to the privacy of individuals, but also to the need to set standards to identify and control the activity of automated agents that interact with each other.
For European companies interested in adopting AI agents, the Moltbook case offers a double interpretation: on the one hand, it shows the potential to automate complex tasks through cooperating bot networksOn the other hand, remember that any serious deployment must comply with security and transparency requirements that will be especially demanding in the EU.
Moltbook's acquisition by Meta marks a turning point in the evolution of social networks and AI itself: a platform born as a weekend experiment has become part of a large corporate lab, bringing with it both its innovative capabilities and its initial mistakes. From now on, The way in which Meta combines this technology with its experience in mass-market products and with European regulatory requirements It will help determine whether "machine social networks" remain a passing curiosity or become a key component of the next generation of digital services.
