Yes, sure. So on the first, I would actually characterize it a little differently. I wouldn't say -- okay, I understand why you asked the question the way you did around Tesla, but I wouldn't actually qualify it as more of an opportunity than a risk. And here's what I mean by that. So autonomous cars, let's step back [indiscernible]. Autonomous cars are definitely coming, right? And if you're in San Francisco, you know it, you see it every day. Sometimes you see good things. Sometimes you see things are a little bit strange. But at the end of the day, they're clearly on the scene. But it's one thing, and I say a very expensive one thing to build a an autonomous vehicle, very, very expensive. It's expensive to build it, billions of dollars of R&D. It's expensive to operate it. These things are not free to operate. Absolutely not. They break down, they have repairs, they have shadow drivers in the back all sorts of things. So that's expense that somebody's got to bear. And then it's also very expensive to build a rideshare platform. And I think it's maybe tempting or maybe it's -- I don't know, easy, let's say, to hear someone say, "Oh, well, you just build an app. And you think, well, that's a rideshare network." Well, no, that's not a rideshare network. A rideshare network involves conversations with every airport in the United States to figure out how did you pick up and drop losses every municipality in the United States. It's pricing 2 million times a day. It's picking things up 24/7 even in bad weather, even when it's snowing. It's figure out how to do supply-demand management such that you get cars in the right place at the right time and so on and so on at quite a large scale. And this is obviously expensive. We spend a lot of money on it every single year, as you guys well know. So -- so when I look at it that way, and I look at -- and then here's where I'll kind of finish on that. If you think of autonomous, here's maybe the way I might break it down. Somebody has to build that technology, right? And they're companies that are just focused on that. Somebody then has to build the cars, either bring their own homegrown technology in the car or take somebody else's technology. That's the second thing. Then somebody's got to own these cars, right? In the Lyft business model, of course, we don't typically own the car. That's -- drivers own their own car, which is great. I mean it's very capital efficient for us. We can put 2 million rides on the road every day and not effectively own a single car assets. That means we don't have to spend hundreds of million dollars in depreciating assets called cars. And then someone has to build the network and operate the network, and that's what we do. So if I look at it in that context, I get excited about autonomous cars because I think great, it's going to be another way for people to get around. You can sort of think of it as another car that we could run into our network. And personally, if I were sort of running the world, I would sort of think of companies specializing maybe in one of those areas. And the companies that try to do more than one of those areas might find it very expensive and maybe not such a great use of capital and resource focus. So that's sort of general thought there. And you definitely ask another question. Oh, yes, yes, yes. I mean so this is where as the CEO. Man, I'd love to talk about stuff coming out, and I'm just completely unable to. But what I can tell you is this. So frustrating. Just keeps me up anyway, but what are you going to do? So -- but here's what I can tell you. So if you look at, for example, our on-time pickup promise, this is another innovation that I just think we're super proud of. So when we launched it last year, you remember the promise. The promise is if we're more than 10 minutes late to pick you up for an airport ride, we will pay you up to $100, no questions asked. Now we've done 2 things since then that are both pretty amazing. The first is it is now available in just about every major airport in the country. When we launched it, we were in the subset now or in just about every major airport. Second of all, we look very closely at that remediation rate, right? What percentage of rides end up not going well? When we last talked about it, that number was about 2%, now is sub-1.5%. And that shows you how operational excellence can drive so much value. And by the way, when we do end up paying this 1.5% of the time when we do end up paying, those riders end up taking another Lyft in the next couple of weeks more frequently than people that, that hasn't happened to. So it's just, I think, a good example of how the innovation we're doing really does pay back for riders and drivers in this case, giving drivers more business and riders more reliability. So you can expect to see more of those. We mentioned in the earlier remarks, some -- as we go into the summer, some other interesting things with airports. Yes, this is why I just have to stop talking or else I'll get myself in trouble. But that I hope gives you a little flavor on what we're working on.