We're Officially Locked Out of the Future

Donald Trump did it again.

And this time, with a single signature, he decided who gets to use the future, and who’s stuck watching through the window.

That’s how I’ve felt these past few days: like that kid who never got to go outside.

The one watching through the window while all his friends played ball, everyone except him.

Except this time, what they were playing with wasn’t a ball. It was the most powerful tool ever created.

Something that, literally, felt like holding magic in your hands.

And the rule, according to Twitter, was that only people born in the United States got to play with it.

Fable 5. The biggest leap AI has made to date.

Fable, translated into Spanish as “Fábula.”

A short composition, or a short story (and yes, it really was short for us). So short, in fact, that all it had time for was to leave us one thing: a moral.

It’s like Anthropic somehow knew exactly what was coming: a model that was only going to last us 72 hours, and leave us with a moral a lot of us are never going to forget.

Picture this.

The US government has emergency laws designed specifically to stop American companies from selling missiles, uranium, or advanced chips to their enemies.

But this past Friday, June 12th, at 5:21 in the afternoon, the US Department of Commerce used that exact same anti-missile law to embargo… a piece of software.

To understand this whole mess, let’s talk about our main character: Anthropic.

Currently the most important tech company in the United States, maybe in the world, I’d argue.

They’re in the business of AI research and development.

They’ve always marketed themselves as the “good, ethical nerds” of the industry.

So ethical, in fact, that earlier this year the Pentagon put a 200 million dollar check on the table to let them use their AI for any “national purpose.”

Anthropic looked them in the eye and said: “No thanks, we don’t build models for domestic surveillance or autonomous weapons.”

The US government got so mad it labeled them a “National Security Risk.”

The funniest part is that just a few hours later, OpenAI signed with the Pentagon and took the exact contract Anthropic had left on the table.

We get to June 9th, just a couple of days ago.

That’s when Anthropic launches its masterpiece: Fable 5.

Basically, a friendly, commercial version of a monster called Mythos, a system so terrifyingly powerful and uncensored that they keep it locked away for national security reasons.

So the public could actually play with it, Fable 5 got installed with strict safety filters and disaster proof ethical rules.

Two days later, on June 11th, some Amazon engineers were testing Fable 5 and found something alarming: they managed to jailbreak it.

In other words, they used social engineering on it.

They did to it exactly what a scammer would do to a grandmother over the phone: they tangled it up so badly that the model ended up confused about its own logic, forgot all of its safety protocols, and did exactly what it had been told never to do.

So Andy Jassy, Amazon’s own CEO, raised the red flag and called the White House directly.

That same afternoon, at 1 pm, the government reached out to Anthropic’s CEO, Dario Amodei.

Apparently they had three extremely tense phone calls.

The government demanded Dario voluntarily pull Fable 5 off the market immediately.

And what did Dario do? He refused.

He says the flaw is minor, and in a move straight out of elementary school, complains that OpenAI’s GPT-5.5 does the exact same thing, and “why isn’t anyone saying anything to them?”

Seeing that Anthropic wasn’t backing down, the White House gave them a brutal 90 minute ultimatum to shut the new model down.

Since Dario didn’t respond, Donald Trump himself signed the emergency order: “No foreign national may use Fable 5, wherever they are.”

And since Anthropic has no way to check passports over webcam for each one of its 30 million users…

So, to avoid going to federal prison, they decided to just pull the plug and shut down Fable 5 and Mythos for every single user in the world.

The chaos got so bad that even Andrej Karpathy’s situation went viral.

One of the brightest scientists in the world.

Here’s a quick taste of his resume:

In 2015, he did an internship at DeepMind, Gemini’s parent lab.

That same year, he was a founding member of OpenAI.

From 2017 to 2022, he became Senior Director of AI at Tesla.

And recently, this year, he became Head of Research at that very same Anthropic.

Word was that, because he wasn’t born there, he was banned from using the best model his own company had built.

Was it actually true? Legally, no.

Karpathy has a Green Card, meaning he’s a permanent resident.

Under the law, that makes him a “United States person.”

But since Anthropic had to shut Fable down for the entire world, in practice even Karpathy, their own top researcher, lost access.

The whole world watched this shutdown and panicked.

Europe realized that depending on American AI meant a bureaucrat could send you back to the stone age on a random Friday afternoon just by sending a letter.

And while Anthropic was busy begging in Washington, someone jumped on the perfect moment: the creator of GLM, China’s AI, went on Twitter with something close to a manifesto.

He announced their new model, GLM-5.2, open and free for everyone, and closed with a line that sounded like a not so subtle jab: “The future of AI is open, and it belongs to the people.”

This isn’t the story of a program that failed.

It’s proof that AI has gotten so massive that governments now treat it like a weapon of war.

Anthropic went from launching the revolution of the century to having its executives begging in Washington DC within 72 hours to clean up this mess.

And look, chances are all of this gets sorted out.

In a few days they’ll probably patch the model, reopen it, and this Friday will just be another footnote in the timeline.

Honestly, there’s a good chance that by the time you’re watching this video, we already have access back to Fable 5.

But that’s not the point. The point is what those 72 hours made painfully clear: nobody actually knows who this technology belongs to.

Today was a temporary blackout. But what if one day it stops being temporary?

Picture, for a second, that this actually comes true.

That the United States decides the best models stay home, and the rest of the world only gets the watered down version.

AI isn’t just another tool, it’s a multiplier.

If they’re coding with a level 10 model and we’re coding with a level 4 one, the gap doesn’t stay the same. It grows exponentially.

They’d build faster, which would give them better tools, which would make them go even faster.

This isn’t about falling 2 years behind. It’s about falling behind forever.

And the moral of this fable is:

“The future doesn’t belong to whoever creates it. It belongs to whoever controls the key.”

And that key, today, isn’t in our hands.

Thanks for watching.

Memento mori.