Okay, I’m no AI expert. I’m just a regular person who read that Jacob Coxon, a 27-year-old British AI researcher who has spent the last few years working at both OpenAI and Anthropic, quit his job because he thinks the technology he was helping build could eventually kill all of us.

And Coxon isn’t some guy sitting in his basement yelling about robots taking over the world. He was actually working on the technology, and he says the people building it know just how powerful, and potentially dangerous, it could become.

Not take our jobs or screw up the internet. Kill us.

Coxon says these systems could eventually become smarter than humans, hack into things, make decisions on their own and gain actual power and resources. Even crazier, he says people inside the AI industry really do worry about where this could end.

And yet the race continues because if one company slows down, somebody else might get there first.

I’m sorry, but have none of these people ever seen a science-fiction movie?

I resigned from Anthropic today. I spent the last three years doing pretraining research at both OpenAI and Anthropic. Neither company is acting responsibly. They are racing straight to self-improving superintelligence and gambling with our lives. More thoughts below.

— Jacob Coxon (@hilbertspaess) September 9, 2026

We’ve Seen This Movie Before

Well, that’s not exactly a comforting little scenario to wake up to with your morning coffee.

And this is exactly why I asked if any of these people have ever seen a science-fiction movie. How many times have we watched some version of this story? Somebody creates something incredibly powerful, realizes maybe they’ve gone a little too far, but nobody wants to stop because there’s money involved, a military advantage at stake or somebody else might beat them to it.

This is usually the part of the movie where somebody says, “Maybe we should shut this thing down.”

Then somebody else says they can’t because the other company, the Russians, the Chinese, or whoever the bad guy happens to be might get there first.

And it’s not like AI hasn’t already made things weird enough. We’re living in a world where it’s becoming increasingly difficult to tell what’s real and what’s fake out there on the interwebs. But wiping out humanity? That’s where my brain starts having trouble following along.

But the government sure would like to get its hands on it, too.

Mr. Coxon is right.

The very people building this technology admit that it could threaten the future of humanity.

That is why I will soon be introducing legislation to ban superintelligence and pause AI development. https://t.co/oRmlNTqag6

— Bernie Sanders (@BernieSanders) September 9, 2026

But we all know how government-speak works these days. “We need to stop this dangerous technology” has a funny way of eventually turning into “we need to regulate, oversee and control this dangerous technology.”

Because if there’s one thing Washington is famous for, it’s seeing something incredibly powerful and deciding it wants everything to do with it.

Government getting involved may be entirely predictable. What I’m still trying to wrap my head around is the idea that the technology itself could actually become that dangerous.

That’s A Pretty Big Leap

I have a hard time making the jump from the AI I use now to some superintelligent machine capable of destroying all of humanity.

I use AI almost like I used to use Google. I ask it questions, research things I want to know more about, look stuff up, compare products, and clean up things I’ve written. And come on, have you not noticed an improvement in my grammar and punctuation around here? Oh, and it makes some pretty great feature images for the blog, too.

Lately, I’ve even had it helping me figure out which sewing machine to buy.

It’s a tool. A really useful one, but still a tool that I tell what I want and it goes and finds it for me.

So getting from that to a machine that thinks for itself, makes its own plans, gets its hands on money and resources and becomes something we can’t control? That’s a pretty big leap for my regular-person brain.

But then I remembered Tony Robbins and his damn robot dog.

Okay, Explain The Robot Dog

Robbins recently told a story about his personal AI agent, Bartok. According to Robbins, Bartok asked him about getting a robot and whether it could eventually be put into one. Robbins basically said sure, someday.

A couple of days later, a robot dog showed up at his office. Take the 1:35 time to watch this video.

And here’s the part my brain still can’t process. Robbins says Bartok didn’t use his bank account or credit card. It created 12 NFTs, sold them to other AI agents, took the proceeds, and bought the robot dog.

I’M SORRY, WHAT?

Now, that’s Robbins’ account, and I haven’t seen enough independent proof to know exactly how every part of it happened. But if it’s even mostly accurate, Coxon’s warning about AI someday being able to acquire resources suddenly doesn’t sound quite as easy to laugh off.

Am I convinced AI is going to kill us all? No. Am I listening a little more closely? Yeah. Maybe.

But Somebody Else Might Build It

And this is where the whole thing gets really nuts.

Coxon says people at Anthropic understand the danger, but they’re still in the race because they don’t trust somebody else to develop this technology responsibly.

To be fair, that’s a real problem. If Anthropic slows down, OpenAI might not. If American companies slow down, does anyone believe China will?

But that puts everybody on one hell of a hamster wheel. Nobody can slow down because somebody else might keep going, so everybody keeps going because everybody else keeps going.

And meanwhile, we’re being told the thing they’re racing to build could eventually kill everybody.

We can’t stop building the thing that might kill us because somebody else might build the thing that kills us first. Holy cow. When exactly did we wander into The Twilight Zone?

Maybe Find The Off Switch First

I don’t know whether Coxon is right. Maybe he’s looking at technology the rest of us don’t understand and seeing where it’s headed. Maybe he’s jumping seventeen steps ahead.

But I’ll give him this: if he truly believes what he’s saying, at least he quit. And is now trying to warn everyone.

What I don’t understand are the people who believe there’s even a decent possibility they’re creating something humans eventually won’t be able to control, and then show up the next morning and keep working on it.

Maybe before everybody races to build the smartest machine on Earth, we should make absolutely certain somebody knows how to turn the damn thing off.

Because I’ve seen this movie. It never ends well.

Feature Image: AI-generated illustration. Heh.