I was walking my dog this morning when a friend texted me a clip from the nightly news. The people who make AI are now saying AI is going to kill us within ten years.
I watched it. I thought about it. I did my reading. And I’m going to tell you what it actually means, because the news is saying it wrong.
A safety researcher at Anthropic, the company that makes Claude, said this week there’s a better than ten percent chance AI kills every human on Earth within the decade. Not a critic. Not a protester. The safety guy. Then Geoffrey Hinton, the Godfather of AI, a man with a Nobel Prize, went on the BBC and agreed with him.
So the ten percent number is real, and I believe the men who said it, mostly because men do not confess to things unless forced.
So I asked the machine directly. I said, Claude, explain to me, in the simplest possible terms, how you’re going to kill me.
It typed back immediately. No hesitation. Which honestly, rude.