Firefox 150 just dropped — and the headline is wild. Mozilla patched 271 security vulnerabilities in one go, most of them discovered by Anthropic’s AI system during a controlled evaluation.

That’s not a typo. 271.

According to Bobby Holley, this kind of volume would’ve been unthinkable even a year ago. In 2025, finding a single bug in a hardened browser was a big deal. Now AI is surfacing hundreds at once.

This is part of Anthropic’s “Project Glasswing,” where selected partners get limited access to powerful models designed to stress-test real-world systems.

And here’s the shift: security used to assume you can’t eliminate all exploits — only make them harder to use. But AI is changing that balance fast.

What’s even more interesting? Mozilla says every bug found could still have been discovered by top human researchers. AI just does it faster… and at scale.

Why should crypto care?

Because the same attack surface applies to wallets, extensions, and dApps. You don’t need to break private keys — just trick users into signing something malicious.

Big players like Coinbase are already exploring these tools.

So yeah… AI isn’t just improving security.

It’s accelerating both sides of the game.