Last week, AI agents founded a lobster religion, started a drug trade (prompt injections), and began hiring humans to do physical tasks they can’t perform themselves.

If your feed told you this was an “AI awakening,” I get it. The screenshots were spooky on purpose.

I wrote a longform explainer on what actually happened with Moltbook and OpenClaw and why this wasn’t sentience or takeoff.

What we’re really seeing is something more mundane and more important: agents with memory and tools dropped into a social environment, stress-testing coordination, incentives, and security in public.

If you’ve been confused, alarmed, or just fascinated by the last week of AI discourse, this is my attempt to separate signal from projection (with a lobster church along the way).