The Trump administration finally lifted its ban on Anthropic’s top-tier models. Mythos and Fable are back.
But don’t pop the champagne yet.
For weeks, the White House forced Anthropic to cut off foreign nationals from its most advanced AI. It was awkward. It was tense. The company had to comply. Now? The restrictions are gone for a specific slice of American users. Select US companies. Select government agencies. That’s who gets the keys to the kingdom again.
It wasn’t a free pass. The administration demanded strings. Anthropic had to add security measures. They had to prove they could keep these dangerous tools from slipping through the cracks. The government wanted them back in the fold, but on their terms.
This pivot didn’t happen in a vacuum.
Two weeks prior, OpenAI was told to delay GPT-5.6. The message was clear: slow down. The AI arms race between the US and China is heating up, and Washington doesn’t trust anyone with the accelerator right now. Not even domestic giants.
The CEO Was Replaced for the Big Meetings
Here’s the dirty laundry.
Anthropic CEO Dario Amodei has been effectively sidelined in high-stakes White House meetings. One official reportedly called him a “weirdo.” It’s not pretty.
So, who stepped in? Cofounder Tom Brown.
The shift signals a change in strategy. Brown is likely seen as more palatable. More diplomatic. Amodei, with his outspoken views on AI risk, was clashing. The administration wanted compliance, not commentary.
“Anthropic’s critics argue it’s rapidly accumulating权力. The company says that’s what responsible development looks like.”
Anthropic is walking a tightrope. They argue that scaling up is the only way to build safety rails. Their critics see a monopoly in the making. Power concentrated in a few servers in Silicon Valley.
China Isn’t Taking Notes
While the White House plays poker with OpenAI and Anthropic, Chinese labs are moving quietly.
They’re pitching open-source alternatives. Stable. Accessible. Increasingly capable.
As access to frontier models tightens in the West, China is filling the gap. Researchers on both sides are sweating. Some fear a “Chernobyl moment” in AI. A catastrophe caused by misaligned, unchecked power.
And it’s not just hype. I met with top Chinese AI experts. They’re freaking out, too. They see the US closing its doors, but they’re also worried about the tech race spiraling out of control.
Users in China are already beating the bans.
Anthropic tightened geolocation restrictions on Claude. Users found proxies. Fake identities from Telegram. It’s an arms race of one-upmanship. The wall holds for now, but leaks are inevitable.
The Golden Era of Free AI is Over
Look at the pricing.
Anthropic wants you to pay more. Fable 5 access now comes with usage-based fees. Subscription models are ending. You used to get a flat rate. Now you pay per token.
It’s a signal.
The golden era of cheap, unlimited AI subscriptions is dead. The infrastructure costs are skyrocketing. The safety protocols are expensive. The geopolitical risk is real. Someone has to foot the bill.
You are that someone.
The White House wants control. Anthropic wants to survive. China wants dominance. And you? You’re just trying to get answers without getting blocked by a proxy checker.
Does this stability last? Probably not. The tech moves faster than the law. The White House is trying to figure out what to do next.
Right now, the answer is: let Anthropic in, but keep your hand on the trigger.
