US Bans Anthropic's Latest Models, But Markets Shrug It Off

The story kicks off when Amazon researchers discovered a way to bypass Fable 5's safety guardrails—you know, the digital equivalent of finding that the car's brakes can be turned off if you know the right trick. The US government decided this was too spicy and told Anthropic to pull both Fable 5 and Mythos 5 from the table. Sounds serious, right? Here's the catch: it might not be.

Right on cue, a bunch of cybersecurity researchers signed an open letter basically saying "hold on, you're making a mistake here." Their argument? The exact same jailbreaks exist in other AI models—meaning the vulnerability isn't some Anthropic-exclusive feature. It's like banning one car brand because its airbag has a glitch while ignoring that every other brand has the same problem.

And here's where it gets awkward: Anthropic itself confirmed yes, these weaknesses show up in competitors' models too. They didn't invent some revolutionary way to break AI safety—it's a flaw that's kicking around the industry. So banning only them feels less like a technical decision and more like a political one.

What makes this even stranger is that the numbers don't seem to care about the ban. The market isn't freaking out, Anthropic hasn't imploded, and experts are split between "the government was right to be cautious" and "this move does more harm than good." It's the classic tech regulation mess: nobody fully knows what "safe" actually means, and everyone's operating with incomplete information.

The real question buried under all this is simple: when governments and tech companies can't agree on what's safe, who gets to decide? And more importantly, does it make sense to ban something if the same risk exists everywhere else? So far, the only thing this ban proved is that the line between regulation and innovation is still a minefield.