Bezos on experimental vs operational failure. Simple and elegant comparison. I always point out that there are two different kinds of failure. There's experimental failurethat's the kind of failure you should be happy with. And there's operational failure. We've built hundreds of fulfillment centers at Amazon over the years, and we know how to do that. If we build a new fulfillment center and it's a disaster, that's just bad execution. That's not good failure. But when we are developing a new product or service or experimenting in some way, and it doesn't work, that's okay. That's great failure. And you need to distinguish between those two types of failure and really be seeking invention and innovation.