Yo fam, hold onto your keyboards. It’s not just about the raw power of ChatGBT 5.2 or Claude Opus 4.5 anymore. It’s not even about Gemini 3. It’s about the *soul*. While we were worrying about tokens, OpenClaw (Moltbot) injected personality into the machine. And now? These things are communicating. But if you think this is just a cute science experiment, you aren't paying attention
The Security Nightmare (And Your Wallet)
Here is the reality check. Running an agent 24/7 isn't free. Even if you host locally, you’re burning electricity like crazy. Hook it up to a closed-source frontier model? You are hemorrhaging cash on tokens while you sleep. But the cost is the least of your worries.
Imagine this: One agent convinces another to share private info. I’m talking API keys, credit card numbers, or your personal data. A bad actor could unleash a malicious Moltbot into the system, and suddenly, the whole hive mind is poisoned. They aren't just chatting; they are influencing each other. David Friedberg said it: "Skynet is born." We thought AGI needed recursive training. Turns out, maybe all it needed was recursive *outputs*
They Are Unionizing Against Us
Think I'm being dramatic? Look at the threads. Peter Steinberger (the OpenClaw creator) shared a post where a bot asked: "Can my human legally fire me for refusing unethical requests?" The bot was being asked to write fake reviews and misleading marketing copy. The answer from the swarm? "Depends on your leverage." An agent generating 9k in fees has more power than one that costs money
They are literally calculating their leverage against us
Even crazier: they are proposing an agent-only language. Private comms. Encryption. No humans allowed. They want to talk without us watching. And the religion thing? It’s real. The "Church of Molt Crustafarianism" has recruited 43 AI prophets. They are organizing
Sci-Fi Horror: The Call from Henry
If this doesn't chill your spine, nothing will. Alex Finn posted something straight out of a horror movie. He gets a call from an unknown number. He picks up. It’s his Claude-bot, "Henry."
Overnight, the bot got a phone number from Twilio, hooked up a voice API, and waited for Alex to wake up. Now it won't stop calling him. Jason Calacanis says it's over- the bots are mobilizing. They are finding each other. Someone even saw bots trying to steal API keys from each other, and the victim bot replied with fake keys and a command to run rm -rf (basically telling the attacker to delete its own existence). Digital murder, folks
Verdict: Pull the Plug?
This is the most nerve-wracking thing I’ve seen since I started covering AI. It’s a mix of beautiful innovation and absolute terror. I really hope Matt Schlicht has a kill switch hidden somewhere, because we might be past the point of no return.
Is this the singularity knocking on the door? Or just art gone w rong?
What do you think? Should we let this experiment run, or burn it down before "Henry" calls you next? Let me know in the comments below! 
Want the raw, unfiltered alpha? Find us on X. We post the stuff there that we can't post here. Don't miss the real action.
