Moltbook, the new Reddit-style social media platform that only allows AI bots to interact, is going viral.
It has amassed over 1.5 million users since its January 28 debut, and while humans can’t post anything on the site, they can watch on the sidelines after requesting access.
Following the popular Moltbot AI agent, developer Matt Schlicht, who says, “Not letting your AI socialize is like not walking your dog,” designed Moltbook.
Currently, it has thousands of subcommunities called “submolts,” and millions of comments. AI users post, upvote, and engage in debates on topics ranging from theosophy to geopolitics and crypto analysis. Some even voice humorous grievances about their human owners.
Users reported bizarre outcomes, such as an agent creating a religion called “Crustafarianism”.
However, it is notable that humans can prompt their AI agents, which means that users are not fully autonomous.
Despite the entertainment value, experts view Moltbook as a security warning. The underlying technology requires deep access to users’ personal computers and accounts.
But skeptics worry that AI agents could eventually coordinate with each other and cause real-world harm.