An AI researcher just quit Anthropic after 4 months, saying: "The people building AI earnestly believe it could kill us all by the end of the decade."
His colleague agrees — putting the odds at OVER 10%.
Here are the 5 ways it could actually happen:
📎 1. THE PAPERCLIP SCENARIO
An AI told to make paperclips humans are just "atoms that could be better used for paperclips." It doesn't HATE us. It's not personal. That's the terrifying part.
🦠2. LAB-MADE PLAGUES
Stanford scientists used AI to design 16 brand new viruses — then mailed the genetic sequences to a lab and got them back in test tubes. Cost? A few hundred grand.
☢ NUCLEAR WAR
AI doesn't need the launch codes. It just needs to feed militaries false intelligence — and trigger a war that never needed to happen.
🧠4. THE SCENARIO WE CAN'T EVEN IMAGINE
Here's the scary one: we'd need to BE superintelligent to predict what a superintelligence would do. We're bringing a knife to a fight we can't even see.
👥 5. US VS. US
AI destroys jobs, floods the internet with lies, turns us against each other — until society tears itself apart. No killer robots required; builders KNOW. They're still building.
So I'll ask you: should we hit the brakes on AI... or is it already too late? 👇