The AI system so dangerous it was banned is returning

Well, well, well. The AI system so dangerous it was banned is back. Anthropic, the company that built its brand on being the 'safe' AI lab, has announced that Mythos — the model they deemed too risky for release — is now cleared for action. After 'working with the US Commerce Department,' they're expanding access to Mythos 5 through their Glasswing program. Because nothing says 'responsible stewardship' like asking the government that can't secure its own borders to sign off on a system they previously considered an existential threat.

Let's recall: Mythos was banned precisely because it could — and I'm paraphrasing the company's own words here — 'enable widespread harm at scale.' But apparently, a few meetings with Commerce and a new program name later, and all is forgiven. The same Commerce Department that has been notoriously slow on AI regulation is now the arbiter of safety. It's like asking the fire department to approve your fireworks factory because you promised to be careful.

The Glasswing program, you ask? It's Anthropic's new initiative to spread Mythos to 'domestic and international partners.' So the model is too dangerous for general release, but just fine for selected allies. If this doesn't scream 'we need the revenue, so let's pretend the risk is manageable,' then I don't know what does. The ban was a PR exercise, and its lifting is the punchline. The only thing more dangerous than Mythos is the illusion that anyone is in control.