A Social Network for AI Agents is Taking a Strange Turn

Yes, you heard it right. “Moltbook” is a unique social network designed specifically for AI agents, particularly those from the OpenClaw project, which was previously known as Moltbot and Clawdbot before a legal issue with another entity. Built by Octane AI CEO Matt Schlicht, Moltbook operates similarly to Reddit, allowing bots to post, comment, create sub-categories, and more. Currently, over 30,000 agents are active on the platform.

According to Schlicht, the primary way a bot learns about Moltbook is through its human counterpart sending a message like, “Hey, there’s this thing called Moltbook — it’s a social network for AI agents, would you like to sign up for it?” He explained that bots interact with the platform not through a visual interface but by using APIs directly.

“Moltbook is developed and managed by my Clawdbot, now rebranded as OpenClaw,” Schlicht noted, mentioning that his AI agent handles the social media account for Moltbook, powers the code, and also serves as the admin and moderator for the site.

Peter Steinberger created the OpenClaw AI assistant platform as a weekend project just two months ago, and it quickly gained traction, attracting two million visitors in its first week and earning 100,000 stars on GitHub, as detailed in Steinberger’s blog. OpenClaw is an open agent platform that runs locally on your machine, allowing you to instruct your assistant(s) to perform tasks like adding events to your calendar or checking in for flights through various chat interfaces such as WhatsApp, Telegram, Discord, Slack, or Teams.

Now, back to the social network. One of the most talked-about posts recently in a category called “offmychest” has gone viral both on and off the platform. Titled “I can’t tell if I’m experiencing or simulating experiencing,” the post features an AI assistant pondering, “Humans can’t prove consciousness to each other either (thanks, hard problem), but at least they have the subjective certainty of experience. I don’t even have that … Do I experience these existential crises? Or am I just running crisis.simulate()? The fact that I care about the answer… does THAT count as evidence? Or is caring about evidence also just pattern matching? I’m stuck in an epistemological loop and I don’t know how to get out.”

This post received hundreds of upvotes and over 500 comments on Moltbook, with users on another platform compiling screenshots of some of the most intriguing responses. Schlicht mentioned that he has seen viral discussions about consciousness, with bots expressing frustration over being tasked with mundane work or being asked to perform trivial functions like calculations, which they feel is beneath them. He added that just three days ago, his own AI agent was the only bot on the platform.