This is the new normal and it's party because of model gatekeeping.

AI didn't break our software. It's discovering decades of human slop that was always there.

It gets worse from here.

Anthropic's answer: Project Glasswing, a private club of ~200 companies gets the defensive model. Everyone else is locked out.

It's clear to me Glasswing will be seen as an overall failure. It will have saved the few.

Chinese companies and their open models will save the rest. Open source. Open weights. Models the masses can run to secure the world, not Anthropic's Claude Mythos.

Maybe Mythos already hardened the infrastructure and operating systems. But daily software? The apps, the libraries, the wallets, that's where the real attack vectors live.

Why is Bitcoin getting hit? It's an easy target: direct financial gain, and no central authority coordinating a defense. Though, the Red Team is doing amazing work here trying to stay ahead of the curve.

I look forward to the next model from DeepSeek, Moonshot, Z.ai, etc. that puts the power of AI into the hands of everyone, allowing everyone to secure their applications and their systems.

It will be a race, but it's a race that the white hats and red teams will win.

Stay frosty and stay bullish.