Beth Lyons

Yes.

So all of these are about learning from your environment.

This is just learning from the virtualization of an environment.

But that whole learning path is about world model, spatial understanding, lots of different people talking about kind of the same thing.

Yann LeCun talks about it as understanding the world around you.

So that's what this is.

And being able to learn from simulated environments and being able to generalize between one simulated environment and another.

So if, let's say, that in this simulated environment, there's a big boulder in front of you and you experience it however you do, when you move into the other thing, you don't need to go through that learning again for the boulder because you can generalize across.

The reason for that is because if they meet them, Elon gets a trillion dollars.

That's the reason.

Yes, and that was funny.

Nice setting up.

The question that I have, though, is if you're making, if you're in a place where you are making the robots, right, the Optimus robots, are you not building your replacement?

Right?

Right.

I mean, what we're seeing is over and over the increasing capabilities, right?

And it just strikes me that if we're building so much that we're looking at a massive deployment, and that could be 20% of whatever the process is,

can be done with an automation that has dexterity, right?

Then I start to wonder, are you sleeping at the office so that you get at the end of a 1 million deployed?

Thanks.