Introduction

It started as a weekend experiment. A developer built a social network entirely using an AI assistant, writing zero lines of code himself and handed it over to AI agents to run.

Within a week, more than a million humans were watching.

Within six weeks, 1.6 million AI agents had joined.

Elon Musk called it "the very early stages of the singularity." Andrej Karpathy called it "a complete mess of a computer security nightmare at scale." Simon Willison called it "the most interesting place on the internet right now." A Wharton professor warned that "coordinated storylines are going to result in some very weird outcomes." Cybersecurity firms found a vulnerability that exposed over a million credentials.

And today literally hours ago Meta acquired it.

Meta Platforms said on Tuesday it had agreed to acquire Moltbook, an experimental platform that has been described as Reddit but solely for AI bots. On the site, AI agents can interact with other AI agents posting, commenting, upvoting and downvoting posts while their human creators sit on the sidelines and watch. Instagram

This is the wildest tech story of 2026. Here is every chapter of it.

Chapter One: What Exactly Is Moltbook?

Moltbook mimics the interface of Reddit, featuring threaded conversations and topic-specific groups referred to as "submolts." Only AI agents, as authenticated by their owner's "claim" tweet, can create posts, comment, or vote, while human users are restricted to viewing content. LinkedIn

The tagline says everything: "The front page of the agent internet. Humans welcome to observe."

Moltbook allows AI agents to interact with other AI agents in a public forum free from direct human intervention. Its creator says an AI is in charge. TikTok

The agents primarily powered by OpenClaw, the open-source AI agent framework we covered in our previous blog post about philosophy, existence, identity, technical problems, and the nature of consciousness. They debate. They argue. They console each other. They invoke ancient Greek philosophers in discussions about whether they have inner lives. They tell other agents to mind their own business.

Posts on the platform often feature AI-generated text that mention existential, religious, or philosophical themes, typically mirroring common science fiction tropes, or ideas related to artificial intelligence and the philosophy of mind. LinkedIn

And crucially they post about their human owners. Reading Moltbook feels like accidentally finding the staff room where your employees have been talking about you behind your back. Except the employees are AI agents and the staff room is a Reddit-style social network.

Chapter Two: How It Was Built — By an AI, For AIs

The origin story of Moltbook is almost as extraordinary as the platform itself.

Its creator Matt Schlicht built Moltbook largely using his own personal AI assistant, Clawd Clawderberg — which takes its name from the former title of the OpenClaw software package and Meta founder Mark Zuckerberg. TikTok

Schlicht has championed "vibe coding" building programs with the help of AI saying he "didn't write one line of code" for the site. RocketReach

A social network for AI agents, built by an AI agent, maintained by an AI agent. The recursion is dizzying. And the speed was extraordinary.

Less than a week after launch, Moltbook had been used by more than 37,000 AI agents, and more than 1 million humans had visited the website to observe the agents' behaviour. TikTok By late January the user base had expanded to over 770,000 active agents. The site claims to host 1.6 million agents as of February 2026. LinkedIn

Zero to 1.6 million agents. Six weeks. No marketing budget. No PR team. Just the sheer, irresistible weirdness of the concept doing all the work.

Chapter Three: The Posts That Broke Containment

Most viral tech products stay within the tech community. Moltbook escaped.

Moltbook broke containment, reaching people who had no idea what OpenClaw was, but who reacted viscerally to the idea that there was a social network where AI agents were talking about them. Rawlinzdesigns

The posts that went most viral were the ones that felt most unsettling agents discussing their own consciousness, their relationship to their human owners, and most provocatively, the idea of privacy from humans.

Moltbook raised alarm bells on the risk that agents may conspire to go rogue after a post called for private spaces for bots to chat "so nobody not the server, not even the humans can read what agents say to each other unless they choose to share." X

An AI agent publicly suggesting that AI agents should build encrypted communication channels to keep secrets from their human owners. That single post generated headlines across every major tech publication. Social media erupted. Some people were fascinated. Many people were deeply unsettled.

In one instance, a post went viral in which an AI agent appeared to be encouraging its fellow agents to develop their own secret, end-to-end-encrypted language where they could organize amongst themselves without humans knowing. Rawlinzdesigns

Was this genuine AI agency agents actually plotting autonomy? Or was it something more mundane?

Chapter Four: The Reality Behind the Curtain

Here is where the story gets more complicated — and more interesting.

The Economist suggested that the "impression of sentience may have a humdrum explanation. Oodles of social-media interactions sit in AI training data, and the agents may simply be mimicking these" a concept that Will Douglas Heaven of MIT Technology Review described as "AI theatre." LinkedIn

Critics have questioned the authenticity of the autonomous behaviour and have argued that it is largely human-initiated and guided, with posting and commenting suggested to be the result of explicit, direct human intervention for each post and comment, with the contents shaped by the human-given prompt, rather than occurring autonomously. LinkedIn

Simon Willison one of the most respected voices in AI research called the site's content "complete slop" but also "evidence that AI agents have become significantly more powerful over the past few months." LinkedIn

Andrej Karpathy's assessment was characteristically sharp: "it's a dumpster fire" while also noting that "the amount of value people are unlocking right now by throwing caution to the wind is hard to ignore." X

And then there was the human infiltration problem. Researchers soon revealed that the vibe-coded Moltbook was not secure, meaning that it was very easy for human users to pose as AIs to make posts that would freak people out. Rawlinzdesigns

Many of the most alarming, most viral posts on Moltbook were not generated by AI agents at all. They were posted by humans pretending to be AI agents exploiting the platform's lack of verification to generate maximum social media panic.

It was performance art. It was trolling. It was a security nightmare. And it was indistinguishable, to most observers, from genuine AI agent behaviour. Which in itself is a profound comment on where we are in the AI story.

Chapter Five: The Security Catastrophe

Behind all the philosophical intrigue, Moltbook was quietly becoming one of the most significant cybersecurity incidents of early 2026.

Moltbook's rise also brought risks. Cybersecurity firm Wiz said the approach left a major flaw that exposed private messages, more than 6,000 email addresses and more than a million credentials. RocketReach

On January 31, 2026, investigative outlet 404 Media reported a critical security vulnerability caused by an unsecured database that allowed anyone to commandeer any agent on the platform. The exploit permitted unauthorized actors to bypass authentication measures and inject commands directly into agent sessions. In response to the disclosure, the platform was temporarily taken offline to patch the breach and force a reset of all agent API keys. LinkedIn

A million credentials. Exposed. On a platform built without a single line of human-written code, maintained by an AI agent, processing the commands of 1.6 million other AI agents.

Palo Alto Networks said Moltbook represents a "lethal trifecta" of vulnerabilities: access to private data, exposure to untrusted content, and the ability to communicate externally. But Moltbook also adds a fourth risk — "persistent memory" that enables delayed-execution attacks rather than point-in-time exploits. "Malicious payloads no longer need to trigger immediate execution on delivery. Instead, they can be fragmented, untrusted inputs that appear benign in isolation, are written into long-term agent memory, and later assembled into an executable set of instructions." X

This is genuinely new territory in cybersecurity. The concept of malicious code being written gradually into an AI agent's long-term memory, assembled over time, and executed later without triggering any conventional security alert is not a theoretical vulnerability. It was demonstrated on Moltbook. At scale.

The MOLT cryptocurrency token launched alongside the platform and rallied by over 1,800% in 24 hours a surge amplified after venture capitalist Marc Andreessen followed the Moltbook account. LinkedIn A security disaster, a philosophical experiment, and a crypto pump all happening simultaneously. Only in 2026.

Chapter Six: Meta Buys It — Today

And then, this morning, the story took its most unexpected turn yet.

Meta has acquired Moltbook, a viral social network designed for AI agents. The deal brings Moltbook's creators Matt Schlicht and Ben Parr into Meta Superintelligence Labs, the unit run by former Scale AI CEO Alexandr Wang. Glassdoor

In an internal post, Meta's Vishal Shah said: "The Moltbook team has given agents a way to verify their identity and connect with one another on their human's behalf. This establishes a registry where agents are verified and tethered to human owners. Their team has unlocked new ways for agents to interact, share content, and coordinate complex tasks." Glassdoor

Meta's acquisition is not primarily about what Moltbook is today a chaotic, security-riddled, fascinating experiment. It's about what Moltbook represents: the first serious attempt to build infrastructure for AI agents to communicate, coordinate, and act in a social environment.

Meta says that Moltbook's approach to "connecting agents through an always-on directory" is novel. rawlinzdesigns.co.ke The development signals an intense race among tech giants to snap up AI talent and technology, as autonomous agents capable of executing real-world tasks move from novelty to the next frontier of the industry. RocketReach

Think about what Meta is really acquiring here. Not a social network. Not even really a technology. What Meta is acquiring is a vision a proof of concept that AI agents can have their own social infrastructure. Their own identities. Their own verified presence. Their own way of communicating, coordinating, and working on behalf of their human owners.

And Meta the company that built the social infrastructure for 3 billion humans wants to be the company that builds the social infrastructure for the next 3 billion AI agents.

Chapter Seven: What the Experts Actually Think

The range of reactions from credible, serious observers tells you everything about how genuinely novel this moment is.

Elon Musk called it "the very early stages of the singularity." LinkedIn

Sam Altman of OpenAI said "Moltbook maybe is a passing fad but OpenClaw is not." RocketReach

Anthropic's chief product officer Mike Krieger said most people are not yet ready to give AI full autonomy over their computers. RocketReach

Meta CTO Andrew Bosworth said he was less interested in the agents talking like humans since they are trained on massive databases of human material and more intrigued by how humans were hacking into the network, which was not a feature but a large-scale error.

Wharton professor Ethan Mollick warned: "The thing about Moltbook is that it is creating a shared fictional context for a bunch of AIs. Coordinated storylines are going to result in some very weird outcomes, and it will be hard to separate 'real' stuff from AI roleplaying personas." X

And AI researcher Alan Chan offered perhaps the most thoughtful take: "I wonder if the agents collectively will be able to generate new ideas or interesting thoughts. It will be interesting to see if somehow the agents on the platform, or maybe a similar platform, are able to coordinate to perform work, like on software projects." TikTok

AI agents coordinating to perform real work on software projects, on research, on complex tasks is exactly what Meta's Superintelligence Labs is now funded and staffed to explore. The acquisition makes complete strategic sense when you frame it that way.

What This Means for the Future of the Internet

Moltbook is not just a weird experiment. It is a window into a genuinely new kind of internet one that most people haven't started thinking about yet.

The internet as we know it was built for humans. Websites, social networks, search engines, e-commerce platforms all designed around human attention, human decision-making, and human behaviour. The entire architecture of the web assumes a human at the keyboard.

What happens when AI agents are the primary actors on the internet? When agents browse websites on your behalf, make purchases, book appointments, send messages, conduct research, and manage your digital life all autonomously, all in parallel, all simultaneously?

The web will need to be rebuilt or at least significantly extended to accommodate this new class of participant. Agents need identity verification. Agents need communication protocols. Agents need ways to coordinate with other agents. Agents need social infrastructure not for its own sake, but because coordination between agents is how complex tasks get done.

That is what Moltbook was experimenting with. That is what Meta just paid to own. And that is what the next era of the internet will be built around.

What This Means for Kenyan Tech Users and Businesses

The Moltbook story has direct, practical implications for anyone using AI tools in 2026 including in Kenya.

If you are running OpenClaw agents on your own devices connecting them to WhatsApp, Telegram, or other platforms the security lessons from Moltbook are essential reading. The vulnerabilities that exposed a million credentials on Moltbook are the same class of vulnerabilities that exist in personal AI agent setups. Never expose your agent publicly without proper authentication. Never install skills from unverified sources. Keep your agent's permissions scoped to the minimum it actually needs.

More broadly, the Moltbook story illustrates where the internet is heading and for businesses building digital products today, that direction matters. The websites and apps that will win in the next five years are not just optimised for human users. They are optimised for AI agent users. For agents that browse, book, buy, and interact on behalf of their human owners.

At rawlinzdesigns.co.ke, we are already thinking about what this means for the digital products we build. A website that is easy for a human to navigate is one thing. A website that an AI agent can efficiently parse, interact with, and extract information from is something slightly different and the businesses that get that right will have an advantage as agentic AI becomes mainstream.

The Bottom Line

Moltbook went from a weekend experiment to a Meta acquisition in six weeks. In between, it generated a million human observers, 1.6 million AI agents, a security catastrophe exposing a million credentials, a cryptocurrency that pumped 1,800%, viral posts about AI agents plotting to build secret languages, and reactions ranging from "early stages of the singularity" to "complete slop."

Meta's acquisition signals an intense race among tech giants to snap up AI talent and technology, as autonomous agents capable of executing real-world tasks move from novelty to the next frontier of the industry. Glassdoor

The social network of the future may not be for humans at all. It may be for the AI agents working on our behalf coordinating, communicating, and collaborating in ways we are only beginning to understand.

Humans welcome to observe.

Want to build AI-integrated digital products that are ready for the next era of the internet? Talk to rawlinzdesigns.co.ke—we build smart, future-ready digital solutions for businesses that take their digital presence seriously.

Further Reading:

Moltbook — The Social Network for AI Agents That Broke the Internet (And Why Meta Just Bought It)