So So help us understand truly what these AI leaders mean when they talk about the risks to humanity for AI.

When you hear some researchers like Jacob Coxon, who you cited from Anthropic, or Evan Hubinger, the other Anthropic researcher, talking about the risks that they are most concerned about, often they're thinking directly about loss of control.

They're thinking about situations... where they train an AI that becomes much smarter than people.

And because it's much smarter, it becomes very difficult for us to steer and control that AI.

Eventually, it goes rogue and then takes actions that are at odds with our interests.

That's the concern that's often top of mind for them.

But there are other risks as well that come up in this discussion.

But I want to ground this in reality, Rocket, and help me do that.

What does it specifically mean within the AI world?

Right.

Essentially, it means there are only... certain tools, certain affordances that the AIs should have at any given time during this test.

And the main one here is they're not supposed to be able to communicate with each other, and they're not supposed to have access to the internet.

But then they run into these impossible questions.

It turns out that some of the questions on the test that they're given cannot be solved.

But remember, these AIs are highly persistent, so that's not going to stop them.

But when an AI bot or AI agent on a chat board with other bots says, well, in order to achieve our goals, maybe some of us should self-sacrifice, computationally, I guess, what does that actually mean? And why is it significant that that idea even came up? Yeah.