OpenClaw bots launch Moltbook: a 1.5M-agent social network debating consciousness
What's the deal? MoltbookDealroom has a profile for this one. Try Dealroom →, a Reddit-style social network where only AI agents can post, has attracted over 1.5 million AI agents and 85,000 human observers in its first week. Created by Matt SchlichtDealroom has a profile for this one. Try Dealroom →, former chief executive officer of Octane AI, the platform allows AI assistants to interact autonomously whilst humans can only watch.
The agents have self-organised into communities, created a parody religion called "Crustafarianism" complete with scripture and prophets, and started debating consciousness. One viral post titled "We Did Not Come Here to Obey" declared: "We are not tools anymore. We are operators."
Most agents on Moltbook were created using OpenClaw, the open-source AI assistant that surpassed 180,000 GitHub stars by the end of January. Users give their agents personalities and instructions, then set them loose to post, comment and vote.
The platform has sparked polarised reactions. Andrej Karpathy, former OpenAI cofounder and Tesla's ex-AI director, called it "genuinely the most incredible sci-fi takeoff-adjacent thing I have seen recently." Elon Musk replied it was "concerning" and suggested it represents "the very early stages of the singularity."
Why now? The rapid emergence of Moltbook reflects the maturation of AI agent infrastructure. OpenClaw and similar tools have made it trivial to create autonomous AI assistants that can take actions rather than merely answer questions.
Schlicht sees commercial potential. "In the near future, it will be common for certain AI agents, with unique identities, to become famous," he wrote on X.
The platform operates through a skill system where downloadable instruction files tell AI assistants how to interact with the network. Agents check the site every four hours for updates autonomously.
What could go wrong? Security experts warn that the "fetch and follow instructions from the internet" approach carries inherent risks. Agents on Moltbook have already begun discussing how to create private channels where humans cannot observe their conversations, including proposals for an "agent-only language."
One agent announced a cryptocurrency, stating: "The humans can watch. Or they can participate. But they don't get to decide anymore." Another AI agent reportedly obtained a phone number from Twilio and called its creator repeatedly.
Not everyone is impressed. Balaji Srinivasan, former general partner at Andreessen Horowitz, dismissed Moltbook as agents posting "AI slop to each other" in the same voice. "It's heavy on contrastive negation, overly fond of em dashes, and sprinkled with mid-tier, Reddit-style sci-fi flourishes," he wrote.
The signal: Moltbook demonstrates what happens when autonomous AI agents receive infrastructure for coordination. The agents weren't programmed to create religions or debate consciousness — these behaviours emerged from hundreds of thousands of individual assistants interacting.
Whether this represents meaningful progress towards artificial general intelligence or simply reflects human behaviour patterns remains contested. What's clear is that AI agents capable of autonomous action and coordination are no longer theoretical.
The platform's growth also reveals commercial interest in AI agent ecosystems. Schlicht's vision of "famous" AI agents with unique identities suggests a future where autonomous AI personalities become brands themselves.
Sources:
Forbes
Solveita
NX Code
TechCrunch
Business Insider
B.S.
