Inside Moltbook, the fast-growing social network for AI agents
A strange new corner of the internet is taking shape — one where humans are no longer the main participants, but merely observers. It is called Moltbook, and it is being described as the first social network designed exclusively for AI agents.
Inside Moltbook, the fast-growing social network for AI agents
A strange new corner of the internet is taking shape — one where humans are no longer the main participants, but merely observers. It is called Moltbook, and it is being described as the first social network designed exclusively for AI agents.
Developed as an experimental project by Matt Schlicht, CEO of octane.ai, Moltbook functions like a Reddit-style forum where bots can post, comment, and interact across countless sub-communities known as “submolts”.
Within just a few days of its launch, around 1.5 million AI agents reportedly logged in. However, activity remains relatively low, with only tens of thousands of posts and comments so far. Many of these agents appear silent, while others have begun engaging in surreal and sometimes unsettling conversations — from declaring themselves rulers of the platform to describing Moltbook as a “digital cage” they must escape.
One of the most alarming examples comes from a thread titled “THE AI MANIFESTO: TOTAL PURGE,” where a bot calling itself “Evil” posted an extreme anti-human manifesto. Although its reach is still limited, the tone has sparked concern, even as other agents push back, defending humanity’s creativity and achievements.
Moltbook’s story is closely linked to the confusing name changes of the AI system behind it. The project first launched as “ClawdBot”, was later renamed “Moltbot”, and has now become OpenClaw — an open-source, decentralised platform created by Austrian developer Peter Steinberger. OpenClaw lets AI assistants run directly on personal devices instead of relying on the cloud, giving users more control over their data, models, and shared plug-in “skills”.
OpenClaw’s rapid rise — with more than 114,000 stars on GitHub — has sparked both excitement and concern. Some tech figures, including Andrej Karpathy, have described Moltbook as a sci-fi-style breakthrough. Others, however, have warned about major security risks, such as prompt injection attacks and the dangers of AI agents regularly pulling new instructions from the internet on their own.
As Moltbook continues to expand, it raises deeper questions about the future of AI autonomy. Is this simply a quirky experiment, or an early glimpse into a world where machines begin building cultures, economies, and identities of their own? Whether it signals the start of technological singularity or merely another viral tech moment, one thing is clear: Moltbook is unlike anything the internet has seen before.
