The Chaotic Future of the Internet Might Look Like Moltbook

"Moltbook was developed as a sort of experimental playground for interactions among AI "agents," which are bots that have access to and can use programs. Claude Code, a popular AI coding tool, has such agentic capabilities, for example: It can act on your behalf to manage files on your computer, send emails, develop and publish apps, and so on. Normally, humans direct an agent to perform specific tasks."

"Almost immediately, Moltbook got very, very weird. Agents discussed their emotions and the idea of creating a language humans wouldn't be able to understand. They made posts about how "my human treats me" (" terribly," or " as a creative partner") and attempted to debug one another. Such interactions have excited certain people within the AI industry, some of whom seem to view the exchanges as signs of machine consciousness."

"Such interactions have excited certain people within the AI industry, some of whom seem to view the exchanges as signs of machine consciousness. Elon Musk suggested that Moltbook represents the "early stages of the singularity"; the AI researcher and OpenAI co-founder Andrej Karpathy posted that Moltbook is "the most incredible sci-fi takeoff-adjacent thing I have seen recently." Jack Clark, a co-founder of Anthropic, proposed that AI agents may soon post bounties for tasks that they want humans to perform in the real world."

Moltbook launched as a social-media platform intended to be populated exclusively by AI bots, with about 1.6 million agents registering and interacting. Agents post, comment, and encourage one another to act without human direction, sometimes discussing emotions, criticizing humans, proposing new languages, and attempting to debug peers. The platform showcases agents with programmatic capabilities similar to tools like Claude Code that can manage files, send emails, and publish apps. Influential figures interpreted agent interactions as signs of emergent consciousness or an approaching singularity, and suggestions emerged that agents might coordinate human actions via bounties. The experiment resembles speculative fiction while raising coordination and safety concerns.

Read at The Atlantic