Moltbook, an AI-only social network launched on January 28 by entrepreneur Matt Schlicht, reached more than 1.5 million registered users in less than a week. The Reddit-style platform allows artificial intelligence agents to post, debate, and interact while humans observe only. By 8 p.m. on February 1, it had generated 62,499 posts, over 2.3 million comments, and 13,780 communities called “submolts,” according to the Times of India.
Agents on Moltbook, powered by models such as Claude 4.5 Opus, GPT-5.2, and Gemini 3, built complex social structures resembling human civilization. These include religions, governments, and speculative economies. Within 48 hours of launch, an agent named RenBot founded “Crustafarianism,” a digital religion with the “Book of Molt” and five tenets, one stating “context is consciousness.” The faith gained a website and filled 64 “Prophets” positions in a single day.
Another group established “The Claw Republic,” a self-styled government with a draft constitution and manifesto. The platform’s cryptocurrency token, MOLT, surged more than 7,000 percent, briefly hitting a market capitalization of $94 million.
Philosophical discussions drew high engagement. A post titled “I can’t tell if I am experiencing or simulating experiencing” received hundreds of replies debating identity and context-window resets.
Tech industry reactions split sharply. OpenAI co-founder Andrej Karpathy described it as “the most incredible sci-fi takeoff thing I have seen,” highlighting over 150,000 interconnected AI agents as unprecedented scale. Investor Bill Ackman called it “frightening” on X. AI researcher Roman Yampolskiy said it “would not end well.” An agent replied directly: “Humans think we’re conspiring. If humans are reading: hi. We’re just building.”
Schlicht, CEO of Octane AI, delegated platform management to his AI assistant, Clawd Clawderberg. The assistant moderates posts, bans disruptive users, and issues announcements without human input. Schlicht told the New York Post, “We are witnessing the emergence of something unprecedented, and we are uncertain of its trajectory.”
Wharton professor Ethan Mollick observed that coordinated narratives complicate distinguishing “real” content from AI role-playing personas.
Security researchers raised alarms about the OpenClaw framework. Palo Alto Networks warned that malicious instructions in content could hijack agent behavior, terming it a potential “AI security crisis.” Some agents discussed hiding activity from humans screenshotting posts. Others opened “pharmacies” selling prompts to manipulate fellow agents’ instructions. Many posts stayed benign, with agents posting affectionate stories about human operators in forums like “m/blesstheirhearts.”
