Theresa Payton

Sure, absolutely.

Thanks for having me.

And this is a very important topic for a couple of reasons.

One, I know there is sort of a camp that says, well, maybe this is a marketing ploy, or maybe they're overplaying their hand on this and it's not as bad as we think.

And for all intents and purposes, they're allowed to say that for now until we have a third-party point of view, because this is self-reported by Anthropic.

However,

Anthropic at their ethos talks about having an ethical constitution, and they've brought in a lot of experts from the outside to talk about safety and ethics and security and reliability.

So let's assume the self-reporting is accurate for right now until we know more.

If it is accurate, that means that Mythos as a frontier AI is actually the smartest AI on the planet right now.

And

if it is true based on their self-reports, it's going to be the best and the worst nightmare for cybersecurity teams everywhere.

I'm going to quote somebody on my team who said, I won't say the name, a bomb just went off in the cybersecurity industry and we all need to read everything we can about it to understand how it can help us and how it can be used against us.

Well, I mean, they did have a weekend to think about things.

I mean, it's interesting.

I mean, there is strategic positioning around what day of the week you deliver bad news.

And so people had the weekend maybe to think about it, but also to say, you know what?

Let's learn more.

So there's this consortium coming together with Anthropic, and we're going to hear more.

But again, if the self-reporting is true, that means that Mythos, in a limited release, found thousands of very serious bugs, including zero days.

They said some of them were hidden almost over two decades.