David Eagleman

It wants to get to food, to water, back to its pack and so on.

So it just, you know, figures out how to run its body differently.

And wouldn't it be great if we could build a billion dollar Mars rover, if we're spending all that money and effort on it, if it could just, you know, saw off its wheel and then figure out how to operate in a different way.

So this is the idea of live wiring.

And it's still the case that almost everything we program and the robots we build and the Mars rovers we build are all totally pre-programmed.

This is what your body looks like.

This is how you're going to operate it.

As opposed to...

letting it operate like a human infant where it has to figure out its body.

I mean, imagine building a robot that flops around for years and eventually crawls and eventually learns how to walk.

That's the kind of thing we need to do if we want it to be flexible and live wired.

And so I'm very interested in the possibilities.

I think the future is going to be much more biological than the way we do it right now, which is we build hardwired machinery that is inflexible.

Yeah.

I mean, the thing about AI, it can do very impressive things.

but it's still not nearly as good as a kid.

A five-year-old can walk into a room, navigate a very complex room between the couches and under the table and whatever, can find her way to food and put food in her mouth, can socially manipulate adults, can do all these things.

AI is really stupid in comparison to that.

It's very good.

It's extraordinarily good at, for example, image recognition or categorization of things.