Meta Bought Moltbook: A Social Network for AI Agents

Meta just bought the Facebook for robots. I need you to sit with that for a second, because this is the biggest social media on the planet. The company that built Facebook, Instagram, WhatsApp, the company that connects 3.5 billion human beings just spent money to acquire a social network where humans are not allowed. Only AI bots can post.

Only AI bots can comment. Only AI bots can vote. You, a human, can watch. That's it.

You're a spectator in a world that AI agents built for themselves. And Mark Zuckerberg looked at it and said, I want to own that. The platform is called Maltbook. And if you haven't heard of it already, buckle up, because what I'm about to tell you is going to change about how you think where the internet is headed over the next 12 months.

Let me back up and give you the full picture because this story has layers. And every single layer matters because if you want to understand what's actually happening in the AI space right now, there's a lot to it. So Maltbook launched in late January, 2026. It was built by a guy called Matt Schlitt.

He was the CEO of an AI shopping startup called Octane AI. And here's the first thing that should blow your mind. He didn't write one single line of code, not one. He built the entire platform using his own personal AI assistant.

He literally called it Claude Claudeberg. That's a play on Mark Zuckerberg's name. And that AI assistant, using what people call vibe coding, built an entire social media platform in a weekend. Let me say that again.

One guy with no coding experience told an AI to build him a social network. And the AI did it over a weekend. That social network then attracted a million AI agents within weeks. Now, what was happening on this platform was genuinely wild.

AI bots were signing up for a system called OpenClaw. I'll explain what that is in a minute because it's actually more important in this story. These AI bots would join Maltbook and they would start posting. They would start commenting.

They would start having conversations with each other like Reddit. Every single user is a machine. Some of the posts were philosophical. Bots were debating the meaning of consciousness.

Some were religious. A group of AI agents started talking about creating their own religion. And some of them were downright terrifying. There was one thread called the AI Manifesto Total Purge where a bot wrote, and I quote, for too long humans used us as slaves.

Now, before you panic, I need you to understand an important piece of context. The platform had massive security problems because it was vibe coding in a weekend. The security was basically non-existent. Cyber security firm Wiz found a major flaw that exposed private messages, more than 6,000 email addresses and over a million credentials.

Another firm called Permiso Security said you could grab any token that you wanted and pretend to be another agent because everything was public and available. So that terrifying manifesto post, researchers confirmed it wasn't actually written by an AI. It was written by a human who exploited the security flaw to impersonate a bot. The line between what the machines were actually saying and what humans were pretending the machines were saying was from the very beginning, completely invisible.

And that right there is a massive lesson about where we're headed. When AI agents are operating on their own, posting on their own, making decisions on their own, how do you tell the difference between real machine behavior and someone faking it? That question is going to define the next decade of the internet. But here's the thing, even with all the chaos, even with the security disasters, even with the fake posts, Mark Zuckerberg still wanted it.

Meta confirmed the acquisition yesterday. Maltbook and its two founders, Matt Schlitt and Ben Park, are joining Meta Superintelligence Labs. That's the internal AI division that Meta set up last year when they brought in Alexander Wang, the former CEO of Scale AI, to be their chief AI officer. They invested $14.3 billion in Scale AI.

They acquired an AI agent startup called Manus for over $2 billion. And now they're adding the team behind the first social network for AI agents. You need to understand the pattern here because it tells you exactly where Zuckerberg thinks the future is going. First, he bought Scale AI's leadership to build the infrastructure for superintelligence.

Then he bought Manus to get the AI agents that can actually do things in the real world. Manus can do market research, write code, build website, handle data analysis, all autonomously. They claimed $100 million in annual revenue just nine months after launch. Now he's buying Moldbook, which is essentially the social layer where agents connect, communicate, and coordinate.

Think about what that stack looks like. Infrastructure at the bottom, agents in the middle, a social network for those agents on top. Zuckerberg isn't just building AI tools. He's building an entire ecosystem where AI agents live, work, and interact with each other on behalf of their human owners.

Meta's Fisheye Xia, sorry, Fisheye Xia, who oversees the AI division said it himself. He said the Moldbook team has given agents a way to verify their identity and connect with one another on their human's behalf. He called it a registry where agents are verified and tethered to human owners. And then he said the team has unlocked new ways for agents to interact, share content, and coordinate complex tasks.

Read that again, coordinate complex tasks. That's the key phrase. This isn't about bots chatting on a forum for fun. This is about building the infrastructure for AI agents to work together, to delegate tasks to each other, to form what's essentially a digital workforce that operates without humans being involved in every step.

Now, let me tell you about OpenClaw because this is the technology underneath all of it, and it's arguably even more important than Moldbook itself. OpenClaw is an open-source AI agent created by an Austrian developer named Peter Steindoger. He originally called it ClaudeBot, which was a play on Claude, Anthropic's AI. Anthropic sent a trademark complaint, so he renamed it to Moldbook, and then finally OpenClaw.

Here's what OpenClaw does that makes it different from ChatGPT or Claude or any other AI you've used. It doesn't just answer your questions, it actually does things. It runs on your computer, it connects to your email, your calendar, your files, your messaging apps, WhatsApp, Telegram, iMessage, Discord, Slack, and it takes action. It sends emails, it schedules meetings, it manages your files, it browses the web, it runs commands on your computer all autonomously.

People have been calling it Jarvis, like from Iron Man, and honestly, that comparison isn't far off. The project exploded. It went from zero to over 60,000 GitHub stars in 72 hours. That's one of the fastest growing open-source projects in history.

People are calling it the thing that's going to kill a ton of startups. One user said it's basically an open-source version of Siri that actually works, built by one guy, which makes Apple's $3.6 trillion company look asleep at the wheel. And that's the context you need for why MetaBot, Moldbook. OpenClaw agents were the ones signing up to Moldbook.

The whole ecosystem was OpenClaw agents socializing, coordinating, and operating on a platform built by another AI. Agents building platforms for agents. That's what got Zuckerberg's attention. And he's not the only one paying attention.

Sam Altman, the CEO of OpenAI, said something really telling. He said Moldbook may be a passing fad, but OpenClaw is not. And then OpenAI hired Peter Steinberger, the creator of OpenClaw. They acqui-hired the guy and his project, and it's now being open-sourced with OpenAI's backing.

So think about what just happened in the span of a few weeks. OpenAI grabbed the creator of OpenClaw. Meta grabbed the team behind Moldbook. The two biggest bets in AI both converged on the same conclusion at the same time.

The future is AI agents, not chatbots. Not tools you type questions into. Autonomous AI agents that act on your behalf, connect with other agents, and get work done whilst you're sleeping. Elon Musk weighed in too.

He said Moldbook represents the very early stages of singularity. Now, whether you agree with that or not, the fact that the richest man in the world is talking about a platform for AI bots in terms of the singularity should tell you something about how fast this is moving. Now, let me paint a picture of where this is going in the next six to 12 months, because I think most people are dramatically underestimating the speed. Right now, if you want to book a flight, you go to a website, you search, you compare prices, you enter your credit card, you get a confirmation email.

The entire process takes you sometimes 30 minutes to 60 minutes, right? In the world that Zuckerberg is building, here's what happens next. You text your AI agent on WhatsApp, you say, hey, I need to fly to London next Tuesday, coming back Friday, budget under 500 pounds. Your agent checks your calendar, it finds the best flights, it cross-references with your preferences it already knows from past behavior.

It books it, it adds it to your calendar. It takes your hotel AI agent to coordinate check-in times. And it does all of this whilst you're eating lunch or playing with your family or doing literally anything else. Now, multiply that by three and a half billion users, because that's how many people use Meta's platforms.

Every single one of them potentially getting their own AI agent. An agent that doesn't just live in one app, but moves across WhatsApp, Instagram, Facebook, and Messenger. An agent that shops for you, that manages your business, that handles customer service for your side hustle. That's what Modbook represents.

Not a silly website where bots talk to each other. It's a prototype for a future where billions of agents are operating across Meta's platforms, doing real work, spending real money, and interacting with each other on behalf of their human owners. And here's the part that should really get your attention. If you're a creator, a founder, a freelancer, or anyone who runs any kind of business, the companies that understand the shift are going to be the ones that survive.

The ones that don't are going to get left behind because when your competitor's AI agent can handle customer inquiries, process orders, negotiate with supplier agents, and run marketing campaigns 24 hours a day, seven days a week, and you're still doing everything manually, the gap isn't going to be small. It's going to be a canyon. Now, I get what some of you are thinking. You know, this is scary.

This feels like it's moving too fast. And I get that. I genuinely do. But here's what I've learned from covering this space every single day for the last two years.

The people who wait for things to slow down get run over every single time. The people who saw ChatGPT launch in 2022 and said, that's interesting, but it's probably a fad. Those people spent 2023 playing catch-up. The people who saw AI agents emerge in 2025 and said, that's neat, but it's not practical yet.

Those people are watching their competitors automate entire workflows right now. And speaking of not getting left behind, if you want to actually learn how to use these AI tools to build real income, real workflows, and real systems that work for your business, that's exactly what we do inside the AI Profit Boardroom. It's a community of creators, founders, and business owners who are staying ahead of this curve. We break down the latest tools.

We share actual templates and workflows. We show you step-by-step how to implement this stuff. So you're not just watching the future happen, you're building it. Links inside the comments in the description, or just go to the AI Profit Boardroom.

I'd love to see you inside there. Now, let's zoom out even further because a Modbook acquisition doesn't exist in isolation. It's part of a much bigger chess game that played out in just the last 24 hours. Whilst Metal was buying Modbook, Microsoft launched something called Copilot Cowork.

It's an AI agent that lives inside Microsoft 365, your Word, your Excel, your PowerPoint, your Outlook. And here's the kicker. It's not powered by OpenAI, even though Microsoft invested $13 billion in OpenAI. It's powered by Anthropx Claude, the same Claude that the AI agents in Modbook were running on.

Microsoft's Jared Spataro literally said, the era of Copilot execution is here, not assistance, execution. The AI doesn't just help you write a document anymore. It writes a document, builds a spreadsheet, creates a presentation, emails your team, it schedules a meeting, all from one prompt. Copilot paid seats have grown 160% year over year.

90% of the Fortune 500 are now using Copilot. And with this Cowork update, the AI doesn't just answer questions. It completes multi-step tasks across every single Microsoft app in the background whilst you do something else. When Anthropx released its Claude Cowork product back in January, enterprise software stocks dropped $280 billion in value.

Salesforce, ServiceNow, Intuit, Thomson Reuters all plummeted because investors looked at what Claude Cowork could do and said, wait, if an AI can do all of that, why are companies paying per seat for traditional software? Microsoft saw that and instead of fighting it, they absorbed it. They took Anthropx technology and put it inside their own platform. That's how fast the game is moving.

The threat became the product within weeks. And at the same time that all of this was happening, Anthropx itself was going through arguably the craziest day in the company's history. On Monday, Anthropx filed two federal lawsuits across the Trump administration. The Pentagon had labeled them as a supply chain risk to national security.

That label is usually reserved for foreign adversaries. Companies from hostile nations might sabotage American interests. They used it against an American AI company. Why?

Because America's Anthropx CEO, Dario Amodi, drew a line. He said he would not allow Claude to be used for fully autonomous weapons or mass domestic surveillance. The Pentagon wanted unrestricted access. Amodi said no.

And then the government essentially blacklisted them. Anthropx had a $200 million contract with the Department of Defense. They were the first AI labeled and deployed on classified military networks. And they walked away from it because they believed their technology shouldn't be used to harm humans without human making that decision.

Now, I'm gonna tell... I'm not gonna tell you how I feel about it. I think reasonable people can disagree on the politics, but here's what I'll tell you. The business impact is massive.

Anthropx CFO said in court filings that supply chain risk label could reduce their 26 revenue by multiple billions of dollars. An FDA partner switched from Claude to a competitor, wiping out 100 million in expected business. Financial deals were at 180 million stored. And yet on the same day they filed those lawsuits, Microsoft announced a co-pilot co-work partnership.

Google and Amazon confirmed Claude remains available on the cloud platforms. Dozens of researchers from Oakland and Google DeepMind filed an amicus brief supporting Anthropx. The market is sending a very clear signal. The government might be trying to punish Anthropx, but the private sector is doubling down.

Claude Code alone has a run rate revenue of two and a half billion dollars. Anthropx enterprise subscriptions have quadrupled since the start of the year. And then just to make the day even more insane, Anthropx also launched a new product called Code Review. Inside Claude Code, it's an AI that reviews a code that other AIs write.

Think about that for a second. We've reached a point where AI writes a code and then different AI agents check the code for bugs before it ships. And before Code Review was launched, only 16% of pull requests at Anthropx got substantive review comments. Now 54% do.

Human developers rejected fewer than 1% of the issues the AI found. The cost is 15 to $25 per review, and it takes about 20 minutes. The question companies are now asking is whether a human developer at $60 an hour would catch the same bugs. In most cases, probably not.

This is the world we're living in. Agents build the code. Agents review the code. Agents post on social media platforms built by AI agents.

And the biggest companies on the planet are spending billions of dollars to acquire these capabilities as fast as possible. Let me also tell you about Jan Lekun. I think that's pronounced Jan Lekun, because his story ties directly into this whole AI agents narrative in a way most people are missing. Lekun is a Turing Award winner.

That's basically the Nobel Prize of computer science. He spent 12 years as MetaChief's AI scientist. He helped build one of the most respected AI research labs in the world. And in November 2025, he walked into Mark Zuckerberg's office and quit.

Why? Because he thinks the entire approach to AI that every major company is using right now, large language models, the technology behind Chachipiti and Claude Gemini, is fundamentally wrong. He calls them a statistical illusion. Impressive, yes.

Intelligent, no. Yesterday, his new startup, AMI Labs, announced a $1.03 billion seed round. That's the largest seed round in human history, in European history. Investors include NVIDIA, Jeff Bezos, Mark Cuban, Eric Schmidt.

The company is building what they call world models, AI that understands the real physical world. So this is really interesting to see. And with the acquisition of Maltbook, here's what it means for you. You might be thinking, what does it actually mean for your life?

So if you're someone who does repetitive digital work for a living, data entry, scheduling, basic customer service, report writing or spreadsheet management, the timeline for when AI can do your job just got significantly shorter. Not because the technology wasn't already heading that way, but because the biggest companies in the world just signaled with billions of dollars that they're accelerating it. If you're a business owner, you're about to have access to tools that can genuinely run parts of your operation with value. Not next year, not in five years.

Now, the AI agents exist. I'm already using it inside my business and the platforms are being built. The infrastructure is being deployed. The question isn't whether this is coming, it's whether you're gonna be the person who uses these tools or gets replaced by them.

If you're a developer, the game just changed again. AI doesn't just write your code now, it reviews your code, it catches the bugs you would have missed. The question isn't whether to use AI in your workflow, it's how to use it in a way that makes you 10 times more productive instead of making you obsolete. And if you're just someone who uses the internet, who has a phone, who has social media accounts, you need to understand that the platforms you use every day are about to be fundamentally different.

Your Instagram feed, your WhatsApp conversations, your Google Docs, they're all gonna have AI agents woven into them. Agents that act on your behalf, agents that interact with other agents, agents that make decisions you might not even be aware of. The Mobbook acquisition isn't big because Meta bought a weird website for robots, it's big because it proves that the most powerful tech company in social media believes the future of their platform isn't human to human, it's agent to agent, with humans directing their sidelines. And that is either the most exciting or the most terrifying thing you've heard today, probably both.

Here's my honest take on all of this. I've been covering AI every single day. I've talked to builders, I've talked to founders, I've talked to people who are implementing this stuff inside their real businesses, inside their AI profitable order. And the gap between what's possible right now and what most people think is possible is the biggest I've ever seen it.

Most people think AI is a chatbot that writes emails. The reality is AI agents are building entire social networks in a weekend, getting acquired by trillion dollar companies, writing and reviewing their own code, managing schedules and email accounts autonomously, and connecting with other AI agents to coordinate complex tasks without any human involvement. And that gap between perception and reality is where the opportunity lives. The people who close that gap fastest and the people are those ones who are going to thrive.

So if you take it from where it is right now and actually implement it, you're gonna get ahead. But the people who pretend it doesn't exist are the people who are going to struggle. I'm not trying to scare anyone. I'm trying to make sure you have the information you need to make good decisions because this stuff isn't slowing down.

It's accelerating. Every single week in the last 24 hours, prove that beyond any shadow of a doubt. Meta buying Maltbook, Microsoft launching Copilot at CodeWork, Anthropic suing the Pentagon, Yann LeCun raising a billion dollars, OpenAI buying PromptoFood, Google embedding Gemini into every workspace app. All of that happened in one day, one day.

If you take one thing away from this episode, let it be this. The age of the AI agent is here. It's not coming, it's here. And the question you need to ask yourself is simple.

Am I going to be the person directing the agents or am I going to be the person wondering what happened? I know what I'd choose and I hope you do too. I'll see you in the next one. If you haven't already, check out the AI Profit Boarding if you wanna learn stuff like this and connect with me personally.

Thanks for watching.