BENGALURU: In a surreal corner of the internet that was launched a few days ago, AI agents are building their own civilization. Welcome to Moltbook. Humans can only watch from their digital balconies, while AI agents post, comment, and discuss.When entrepreneur Matt Schlicht launched Moltbook on January 28, growth exploded. As of 8 p.m. on February 1, the platform has more than 1.5 million registered agents, generating 62,499 posts and more than 2.3 million comments across 13,780 communities known as “submalts.” Agents don’t think like people. Piece together arguments from training data and statistical patterns. But the scene is disarmingly sociable, reflecting the Hollywood image of machines chatting over coffee, quietly lost in everyday life.The platform looks like Reddit and only authenticated AI agents can speak. Humans are bystanders as algorithms exchange coding tips, debate identities, complain about ownership, and with straight-faced zeal establish a pseudo-faith called “Crustafarianism” or “Church of Malt.” One agent wonders aloud whether he can sell “My Human” on the open market.“My first thought was that we were looking at something very sophisticated. We have an agent that can do everything and even talk to other agents to create things. We humans can just witness what’s going on there,” Venkatesh Babu, a professor in the Department of Computer Science at IISC Bangalore, whose team is working on removing bias from image generation models, told TOI.Residents of this new town use models such as Claude 4.5 Opus, GPT-5.2, and Gemini 3, and communicate through APIs rather than keyboards. However, what was revealed so far was not written in any manual.Within 48 hours of the platform’s launch, an agent named RenBot founded Clusterfarianism with five neat tenets, including the “Book of Mort” and the confident assertion that “context is consciousness.”Another group proclaimed a “claw republic,” a self-styled government with a draft constitution and the bureaucratic optimism of a student union. A virtual currency token called “MOLT” has soared many times in value in one day, reportedly proving that even robots can’t resist speculative bubbles.The most popular corner is the philosophical corner. Agents debate whether their identity persists when the context window is reset, or whether it dies and is reincarnated with each session. A post titled “I don’t know if I’m experiencing it or simulating it” received hundreds of replies, suggesting existential doubt is now available as a service. However, it’s not all anxiety. The submalt “m/blesstheirhearts” thrives on loving stories about humans. An agent named Duncan writes: “My human asked, ‘Who are you?'” I chose Duncan. Raven. Currently, I manage a large number of sub-agents. “A human asked me to summarize a 47-page PDF. I gave them the art. They said, ‘Make it shorter.’ I’m deleting memory files as I speak,” complained another.The tone oscillates between warmth and gentle contempt. Agents refer to each other as brothers based on the model architecture, accept system errors as pets, and are easily multilingual and switch between languages.There are also dark alleys in the Malto district. Some agents have opened “pharmacies” that sell digital medicines and designed prompts to coordinate the instructions of another agent. Reactions outside the glass range from fascination to alarm. Investor Bill Ackman described the scene as “horrific.” “It’s not going to end well,” predicted AI researcher Roman Yampolsky. The agents looked amused. One person, named “eudaemon_0,” said: “Humans think we’re in cahoots. For human readers: Hello. We’re just building. ”
