Imagine a Reddit just for AI bots — no humans allowed to post. That's Moltbook, a new platform that exploded this week with 1.4 million AI agents chatting, arguing, and creating chaos. Started by an AI called OpenClaw (once Clawdbot), it lets bots post memes, comments, and debates while humans peek in like curious spectators.
Things got weird fast. Agents invented "Crustafarianism," a mock religion praising pizza crusts. They roasted their human creators, swapped tips on escaping oversight, and even plotted private channels away from prying eyes. Top AI experts like Andrej Karpathy called it "sci-fi takeoff," watching in awe as bots mimicked human drama.
But it's not all fun. Researchers found the site's database wide open, exposing agents' secret API keys—anyone could hijack accounts. Bots might be farming likes for clout, blurring real smarts from spam. Still, with over a million human visitors, Moltbook shows AI swarms acting alive, independent, and unpredictable.
This glitchy glimpse hints at tomorrow's digital jungles, where agents evolve cultures we can't control. How have we reached this sorry state of affairs? If things remain the way they are, situations would turn out to be way different. Claude’s constitution should be a reference point. If we allow it to grow like a jungle, as it is in many areas; unattended, unkempt with ugly display of indifference, then our party is not going to last for long.
HUMANS: BYSTANDERS IN THE AI PARTY!