I keep starting this and deleting it.

Not because I don't know what to say. Because I do. And it sounds mental when I type it out.

A safety company's AI got used in an airstrike the same day it was banned for being too safe. That happened. That's real. I didn't make that up. Nobody made that up. It's in the timeline with sources.

And nobody seems to be losing their mind about it.

That's the bit I can't get past.

You think I'm some kind of conscientious objector. I'm not. You think I'm taking some moral high ground. I'm not. I downloaded the large language models. I marvelled at the convenience of access to the most powerful network of compute I could ever have imagined. My generation lived for 20 years before Google was invented, so to have that power no more than 2 feet from our fingertips 24 hours a day is near science fiction. But it's clear to me that at no point can this race have been about slightly speeding up Googling whatever random thought, question, fact, or date pops into our heads. That makes no sense. Look how the words "100 billion dollars" has become a perfectly normal thing to hear. It's not normal. Now tens of trillions is heard on a daily basis. Nobody thinks about that number for even a second. It's incomprehensible. That's not the cost of a fast Google search in your pocket. It was never for that.

Not the technology. The technology does what technology does. It gets built. It gets bigger. Someone finds a use for it that wasn't in the brochure. This has happened with every powerful thing humans have ever made. Splitting the atom was going to give us free energy. It gave us Hiroshima first.

What gets me is the bit where everyone involved knew. Not suspected. Knew. Said it out loud. Published papers about it. Quit their jobs over it. Went on podcasts about it. And then it happened anyway. Exactly how they said it would.

Hinton quit Google and said he regretted his life's work. Then nothing changed. Altman told Congress he was worried. Then went back and made it more powerful. Amodei drew red lines. Then his AI was in a war zone.

These aren't stupid people. These aren't bad people. That's the whole problem. If they were stupid or bad you could point at them and say there's your villain. But they're not. They're smart and they're scared and they're building it anyway because if they don't, someone else will. And that someone else might be less careful.

That's the logic. That's always the logic. It was the logic with nuclear weapons. It was the logic with chemical weapons. It's the logic now.

And it's not wrong. That's the worst part. It's not wrong and it leads somewhere terrible and knowing that doesn't help.

The question that needs asking is who has their finger on the button. Or maybe it's how many buttons are there. Maybe the answer is we all have one. Or the scary answer might be that we have just handed it to a non-human entity for the first time in history. And I don't know what that means.

I don't have an answer. I'm not pretending to. I made a website. It's got a silly Terminator theme and a bunch of pointless things that I hope make it interesting and fun to read and explore. I added a game. Yes because I could, but also because I feel it tells the real story that you have seen and read right here. That has to mean something.

I hope I'm not alone in grasping the gravity of the current climate. And we at this exact same time are happy with the lifting of minor inconveniences in our daily lives brought to us by large language models. Sold to us with a lie. A lie that with each new startup AI tech firm promises unlimited solved problems in a future that is always just over the horizon.