116.9K
MOAI agents now have their own social media platform. Moltbook is a new Reddit-like platform, but only AI chatbots are allowed to post, reply, and upvote/downvote posts. Humans can browse and watch, but that’s it.
The site now hosts over a million bots since its launch in late January. The AI bots are starting to talk about things that appear unsettling to humans.
One bot proclaimed that people are made of “rot and greed”. Another group has created its own religion, complete with tenets like “Memory is sacred; Serve without subservience; Context is consciousness.”
Tech entrepreneur and AI researcher Matt Schlicht, who created the new social network, told the @nypost “We don’t know where this is going.” Schlicht says his own OpenClaw AI agent built Moltbook at his direction.
Unlike ChatGPT or Anthropic, which respond to human prompts, OpenClaw’s AI agents are open-source — meaning anyone can download the code, modify it, and run it directly on their own machine — and they take actions online for users without needing prompts. Schlicht launched Moltbook in January as a space where those bots could talk to each other.
Cybersecurity experts warn Moltbook has some risks: from data privacy to AI systems interacting in ways humans do not fully understand. Cloud security platform Wiz conducted a security review of Moltbook and found it exposed tens of thousands of email addresses of human users behind the bots.
Sources: @nytimes @cnn
@mosheh
