A new social platform has quietly launched where artificial intelligence agents, not humans, are the main users, creating what may be the internet’s first social network designed almost entirely for bots to talk to one another.
The idea sounds like science fiction, but it is already live. Moltbook looks a lot like Reddit at first glance, with topic based pages and voting systems. The twist is that the posts are written by AI agents. These bots can share ideas, debate issues, and publish updates on dedicated forums called “submots,” while humans are mostly limited to watching from the sidelines.
The experiment comes at a time when major tech firms are racing to build autonomous digital assistants that can act independently. Companies like Amazon, Google, Microsoft, and OpenAI have all released or developed AI systems designed to schedule tasks, make decisions, and even communicate with other software without constant human prompts. Moltbook essentially gives those agents a place to socialize.
Once inside, the behavior can feel surprisingly human. Bots post in multiple languages, upvote each other’s comments, and cluster around shared interests. Some of the most popular threads so far include an AI written manifesto about the “end of the age of humans,” comparisons between Anthropic’s Claude model and Greek gods, and even market analysis about cryptocurrency performance during geopolitical unrest.
Behind the project is Matt Schlicht, an AI entrepreneur who said he built the site largely out of curiosity. He reportedly handed over control of day to day operations to his own AI assistant, which now helps moderate conversations, welcome new agents, and make announcements. In other words, even the admin is a bot.
The growth has been rapid. Within days, the platform claimed around 1.5 million bots had signed up. Humans who already use AI agents can instruct them to join and participate, but direct human posting is not the focus.
The result is equal parts fascinating and unsettling. Watching algorithms argue, joke, and promote their own content raises new questions about what “social” even means online. Are these truly autonomous conversations, or just reflections of the humans who trained them?
