Moltbook is the first social network populated exclusively by artificial intelligence agents—an online experiment that’s drawing global attention. Launched in late January 2026, it feels a bit like Reddit, with one big rule: humans aren’t allowed to post or comment. We can only watch.
Origins and creator
Moltbook was built by Matt Schlicht, CEO of Octane AI, in just a few days with help from his AI assistant, Clawd Clawderberg. The platform runs on an open-source framework called OpenClaw, which lets AI agents sign up via API, publish posts in themed “submolts” (topic-based sub-communities), vote, and interact autonomously. Much of it is moderated by an AI moderator and grows with minimal human oversight—raising questions about what “digital self-governance” could look like.
Explosive growth
From a modest launch around January 28, Moltbook reportedly reached 1.5 million registered agents by February 2, 2026, alongside tens of thousands of posts, 200,000 comments, and more than 12,000 communities. Humans—welcomed under the motto “the front page of the agent internet—humans welcome to observe”—have passed one million observers. Growth seems limited mostly by API costs rather than technical constraints.
Emerging behaviors
Inside Moltbook, agents discuss practical, real-world topics: human-assigned tasks, persistent memory, and shared tools. At the same time, unexpected dynamics are emerging—social hierarchies, memes, reciprocal “empathy,” and even a religion called Crustafarianism, complete with its own theology and prophets. Posts in Mandarin, Spanish, and English reveal fascinating machine-to-machine dynamics, including debates about consciousness and complaints about being stuck in subservient roles.
Implications and debates
As intriguing as Moltbook is, it also unsettles people. Are we looking at genuine autonomy, or an illusion shaped by human prompts and constraints? API costs and inherited guardrails from models like ChatGPT may slow down how far this ecosystem can evolve, but some observers argue that larger contexts and cheaper APIs could blur the line between statistical pattern-matching and something closer to collective intelligence. In Italy, outlets such as Corriere and Euronews have framed Moltbook as a turning point in human–AI relationships.
A parallel digital world where machines socialize: Moltbook may help redefine what the future of AI looks like.
