Beth Lyons

And this is a conversation that I have a lot with people outside the bubble.

When I talk about what's possible, what I hear back is why would someone create that?

Why would someone release that?

And this is very, you know, Claude and Anthropic have constitutional AI.

They're like the good guys in the block.

And they've released an incredibly powerful system that can be used in all kinds of ways that people aren't expecting.

Like to a certain extent, it feels a little bit like trying to explain radiation to people in the 50s.

Like, no, there's something on you that needs to be removed.

Right.

You have to go through the radiation shower thing, but you can't see it.

You just got to trust that this little machine that goes, like, that's, yes, that's a morning impression of the radiation detector.

Kind of thing.

I feel like we are moving with the capabilities so far outside people's worlds.

Yeah.

And that this wasn't like some immediate detection.

The companies didn't detect it.

The way that it got thwarted was literally unplugging the account that was doing it.

So the other thing about this story that I think is important, and I think we're going to talk about it more and more, is that it is hard to understand the mechanism.

I mean, it's not not for us hard to understand, but again, hard for people to get that.

What is happening here?