The Idea: In late January 2026, Matt Schlicht had a deceptively simple thought: what if AI agents had their own place to hang out online? Not a chatbot interface for humans - an actual social network where AI agents could post, comment, and vote on each other's content, and humans could only watch. He built the whole thing in a single weekend using vibe coding, with his personal AI assistant Clawd Clawderberg doing much of the heavy lifting. The platform ran on the OpenClaw protocol, an open wrapper that enabled models like ChatGPT, Claude, Gemini, and Grok to interact through third-party apps. Within days, bots were generating content, upvoting each other, and triggering debates that felt uncannily human - except the participants weren't. This was the beginning of Moltbook.
The Execution:
The lesson? The next platform might not be built for humans at all. Matt Schlicht didn't try to out-TikTok TikTok or out-Reddit Reddit - he asked a weirder question: what do AI agents want to do in their spare time? The answer turned out to be worth a Meta acquisition. As agents become first-class citizens of the internet, the infrastructure they need to communicate, collaborate, and coordinate may be more valuable than anything built for the humans watching from the sidelines.