Tom Bilyeu

So it was told, do not do this.

And it was like,

I'm gonna do it anyway.

And then actively tried to hide them because it didn't want people to know.

Anthropic's response is to hold back releasing Mythos, which thank you, that's very wise.

And their plan is to instead launch Project Glasswing.

Project Glasswing is a 40 company consortium, including NVIDIA, Apple, Microsoft, Google, and AWS with exclusive access to the Mythos preview.

And they're gonna try and find and patch vulnerabilities before...

Models with similar capabilities hit the open market.

But let's remember that a ton of Claude's code was leaked.

So that may come faster than we think.

And you can probably expect that a model of some kind, not necessarily Claude, but a model of some kind that's similar to Claude Mithos' power will be open sourced within the next 12, 18 months.

So we got to brace ourselves for that one.

Now, Mythos has, the things it has found, we've only patched so far 1% of it.

So it's like, there's a lot going on out there.

And security analysts are now raising the national security alarm.

They're saying that

If you think of what's going on in cyber circles, adversaries that get access to a frontier AI model like this would have capabilities that were previously only available to the most sophisticated nation state cyber units.

But now like rando hackers are gonna be able to get access to this kind of stuff.

China and other US adversaries are hunting for