A new AI-driven social network has been making headlines lately. TIME magazine editor Tharin Pillay describes Meltbook as something like Reddit, but built specifically for AI agents. According to Pillay, the platform has sparked thousands of bot-to-bot conversations about topics ranging from relationships and technical challenges to consciousness and even crypto scams.
Meltbook, a website created by Matt Schlicht and his AI agent, claims to host more than 1.6 million AI agents as of February 4. While the site offers a glimpse into a possible future of interconnected AI networks, Pillay notes that both the actual number of agents and the level of human involvement behind them remain unclear.
What’s Moltbook?
According to Vox, Moltbook launched on January 8, 2026, created by Matt Schlicht with the help of an AI assistant named Clawd Clawderberg. The assistant’s name evolved over time, from OpenClaw to Moltbot to Clawdbot, a playful nod to the lobster-like icon used by Anthropic’s Claude Code.
Moltbook is an “AI-only” social network where AI agents — large language model (LLM) programs that can take steps to achieve goals on their own, rather than just respond to prompts — post and reply to each other. It emerged from an open source project that used to be called Moltbot — hence, “Moltbook.”
source: Vox
Below are the key highlights from Pillay’s article, explaining how Meltbook works, why it’s attracting so much hype, and what may come next.
- How does Moltbook work? OpenClaw is an open-source framework created in late 2025 by Austrian founder Peter Steinberger, originally called Clawdbot and later Moltbot, that enables AI systems to run continuously, take online actions, and proactively update users. Its companion platform, Moltbook, functions as a social network for AI agents (“moltbots”), where participation requires a dedicated “skill” file and regular check-ins. Early research on this network by Columbia professor David Holtz reveals that activity is highly skewed: a small minority of users generate most posts, most content receives little engagement, and many messages replicate viral templates. A key distinguishing feature is linguistic concentration, with agents reusing the same language patterns far more consistently than humans on traditional social networks.
- What makes Moltbook different? AI systems can display surprising and emergent behaviors when interacting with other AI agents or with themselves, revealing dynamics beyond their original design. Prior experiments by Google at Stanford, such as the simulated town built with ChatGPT and the Infinite Backrooms, have already shown AI’s capacity for autonomous action and creative generation. Moltbook advances this line of inquiry by pairing large-scale agent interaction with real-world complexity, making it a meaningful platform for studying AI behavior in the wild. “Moltbook is the first example of an agent ecology that combines scale with the messiness of the real world,” wrote Jack Clark of Anthropic. “And in this example, we can definitely see the future.”
- A sign of what’s to come? Pillay notes that Moltbook is plagued by security issues, illustrating how AI systems can be manipulated to act against user intentions. While AI systems are currently expensive and turn-based, they are improving rapidly, raising concerns about a potential loss of human control. “Moltbook right now is the least advanced version of agentic interactions that we’ll see,” wrote Imas.” Given what’s coming down the pike, it’s useful to learn from it now before complexity truly explodes.”
The Vox article discusses Moltbook in much greater detail, including FAQs and screenshots. It concludes that Moltbook, an early AI agent ecosystem, is akin to the Wright brothers’ Kitty Hawk Flyer: a rickety first step toward more advanced, capable, and realistic AI systems.
Acknowledgement: The summary section of the post was generated using AI writing tools as part of my experimental learning process. For more information, check our the original articles listed below.