YouSaid · the spoken record

Aicha Evans

lines on the record: 82
first: 2020-03-06
most recent: 2020-03-06
sittings or episodes: 1
sources: podcast

Every line below is reproduced as it was said and linked to the record it came from. Nothing here is summarised or generated. Directory · Search · Corrections

“And so you can't say the system has to truly never get into a crash. For example, you could be stopped at a red light and somebody could just come and hit you from the side. It's literally physically nothing you can do about that. So it's not that we're going to say, hey, there's never going to be a collision, but we're going to say we have to be much better than humans and we're going to be especially strict at crashes that would have been our fault. Because I think the public will have a very low tolerance for robot vehicles that are proactively causing crashes. And so, our rate on those has to be, in our view, about an order of magnitude lower than an”
2020-03-06 · Decoder with Nilay Patel · Recode Decode: Aicha Evans and Jesse Levinson · IDENTIFIED FROM THE TRANSCRIPT
“Yeah, so in our view, what a safe autonomous vehicle means is a vehicle that per mile is significantly safer than humans at specifically not causing collisions, injuries, and fatality. Those are the three categories that we really focus on. I think it's really important to focus on all three of those categories. You can't just look at any one of them in isolation. We think we have to be better on all three. And we also look at when there is a collision, whether it's at this point in simulation, but when we think about deploying, it's very important that we are particularly better at not causing collisions injuries and fatalities. There's always going to be some amount of other road users doing reckless things.”
2020-03-06 · Decoder with Nilay Patel · Recode Decode: Aicha Evans and Jesse Levinson · IDENTIFIED FROM THE TRANSCRIPT
“And so this. So the status quo is pretty terrible when it comes to the safety of automobiles. And so the way we look at it is we have an opportunity to really improve on that. But we also have a responsibility to improve on that. If we can't quantitatively demonstrate that our technology is significantly safer than human drivers, then we shouldn't have our vehicles on the road. So we have pretty stringent internal safety targets on how safe the system needs to work. And we do a lot of simulation and real world testing with the L3 with safety drivers to make sure that we can prove that safety case out. And of course, we will not launch driverless vehicles on public roads until we have finished that work.”
2020-03-06 · Decoder with Nilay Patel · Recode Decode: Aicha Evans and Jesse Levinson · IDENTIFIED FROM THE TRANSCRIPT
“And it's late February now, the total number of deaths from coronavirus is still less than the number of people who die literally every single day in car crashes. So that's kind of a staggering way to think about it”
2020-03-06 · Decoder with Nilay Patel · Recode Decode: Aicha Evans and Jesse Levinson · IDENTIFIED FROM THE TRANSCRIPT
“Sure. Yeah. Yeah. So almost 40,000 Americans get killed in car crashes every year. It's about 1.3 million people globally, which is a staggering number. I mean, just about that in perspective. The whole world has been freaking out about coronavirus for a couple months, and understandably so, and this is high.”
2020-03-06 · Decoder with Nilay Patel · Recode Decode: Aicha Evans and Jesse Levinson · IDENTIFIED FROM THE TRANSCRIPT
“So that's actually already available. So again, it's a matter of executing and really achieving safety targets, feeling comfortable around that. Because once you deploy, you have to make sure that you're safe. And that really depends upon us. So I know there's a lot of discussion talking about the regulatory frameworks, but we feel that actually one can deploy today if one has the technology and one considers it ready and safe. Obviously, you have also an operational infrastructure to put in place.”
2020-03-06 · Decoder with Nilay Patel · Recode Decode: Aicha Evans and Jesse Levinson · IDENTIFIED FROM THE TRANSCRIPT
“We actually have found them to be very cooperative. They are all very interested in the technology. We often host folks at the federal, municipal, as well as at the state level. Several states have Essentially, you can get a driverless permit, right? So that more and more states are coming online.”
2020-03-06 · Decoder with Nilay Patel · Recode Decode: Aicha Evans and Jesse Levinson · IDENTIFIED FROM THE TRANSCRIPT
“But yeah, it's a pretty famous or infamous racetrack. And so we have a lease on that now. And it's really great for structured testing so we can test the vehicle at low and high speeds repeatably. We can do scenarios so we can test it with other agents. And because it's our facility, we can control it and so we can make sure that it's consistent and repeatable. And because it's a private facility, we're not bound to public road testing requirements.”
2020-03-06 · Decoder with Nilay Patel · Recode Decode: Aicha Evans and Jesse Levinson · IDENTIFIED FROM THE TRANSCRIPT
“Yeah, we like to call it final assembly, very similar to the video you saw earlier on how it happened. So you have stations and all of these parts come in. It's a relatively low number of parts that come in and get assembled together and then tested and then deployed.”
2020-03-06 · Decoder with Nilay Patel · Recode Decode: Aicha Evans and Jesse Levinson · IDENTIFIED FROM THE TRANSCRIPT
“Manufacturing facility. Although you get parts from all over the world, whether it's China, France, or Germany, or wherever you get it. But you're going to manufacture these, you were talking about sort of a Lego-like system, is that you pull things apart and add them in.”
2020-03-06 · Decoder with Nilay Patel · Recode Decode: Aicha Evans and Jesse Levinson · IDENTIFIED FROM THE TRANSCRIPT
“Okay, so what I know we need is capital, like everybody else. I think it's important to have strategic partners, and we're in the middle of those conversations. Short term, the focus on solving for autonomy, I always say that zero multiplied by anything is still zero. Now, when we start talking about scale and... Multiple cities going global, obviously, it will be very, very good to have partners. And you don't wait until you need that to start dating, shall we say, to use your analogy. So that's something we're working on. But first, focus, execute, and get this tomorrow.”
2020-03-06 · Decoder with Nilay Patel · Recode Decode: Aicha Evans and Jesse Levinson · IDENTIFIED FROM THE TRANSCRIPT
“And so when I put up that efficiency and that effectiveness so far, I'm really proud of the team and we need to continue doing that. And yes, it's no secret that we're fundraising like everybody else and we're in the middle of that. We're in advanced discussions with some potential investors, strategic investors and also corporate investors.”
2020-03-06 · Decoder with Nilay Patel · Recode Decode: Aicha Evans and Jesse Levinson · IDENTIFIED FROM THE TRANSCRIPT
“This is an expensive problem. I think about it different ways. It's a lot of money, but it's not. That's true because more is needed to get to market. And that's not specific to us. That's for everybody. But I walk around these hallways and I look at what we've accomplished with that money. And with the software and AI and essentially the fleet, meaning testing our L3s, but also with the vehicle.”
2020-03-06 · Decoder with Nilay Patel · Recode Decode: Aicha Evans and Jesse Levinson · IDENTIFIED FROM THE TRANSCRIPT
“Well, I don't know that we don't know, no, no. I don't think that will kill off an existing market. I think that what you'll see. You know that's my thing. Well, that was a different question. Yeah, yeah. So I think you're already seeing it in cities. A lot of people don't buy cars. It's not economical. It's expensive, as Jesse was talking about earlier. People are already using really mobility as a service. We're just offering a step up from that. It's also good for the cities. I mean, look at all much of the real estate is wasted on parking, for example. Everybody's complaining about everybody also driving around looking for parking, so it's going to be very efficient. And so back to competition. Look, first we're happy. It seems like many more people are coming to the same realization and sort of validating this approach. That's one. Second, meaning.”
2020-03-06 · Decoder with Nilay Patel · Recode Decode: Aicha Evans and Jesse Levinson · IDENTIFIED FROM THE TRANSCRIPT
“To do, yes. And you Okay, so in terms of getting to market, it's about sort of completing that step of fully removing the driver and then getting onto public roads. You won't hear at Zooks anybody talk about a hockey stick. It's going to start with geofenced areas where one will be able to sort of demonstrate the experience and start building a business around that and then scale from there, geofans by geofans city by city. We're testing in Las Vegas and San Francisco. We've been very public about that. So I think it's a fair assumption that we would be ending up there and starting there and building a service there. In terms of the competition, yeah, well, yes, there's a lot of competition. It's also a big market. We are very, very happy.”
2020-03-06 · Decoder with Nilay Patel · Recode Decode: Aicha Evans and Jesse Levinson · IDENTIFIED FROM THE TRANSCRIPT
“Okay, but we can increase the quality of the service, increase the safety, and we can lower the cost because we remove 80% or so of the fare, which is going to the driver. And then we replace it with a robo taxi. So we have some expenses too, but it's much, much less than paying an entire person all the time.”
2020-03-06 · Decoder with Nilay Patel · Recode Decode: Aicha Evans and Jesse Levinson · IDENTIFIED FROM THE TRANSCRIPT
“Yeah, well, there's sort of two different business models, right? One is you buy a car and hopefully it is sort of more and more autonomous features come into cars over time. And the other is you use Uber and Lyft. And the way we look at it is neither of those solutions is quite optimal, at least for getting around cities. One of the troubles with Uber and Lyft is you're really getting into the backseat of a stranger's car. And some of those experiences are really good and some of them are really not so good. And no matter what, you're having to pay somebody else to drive you around, which is if you think about it quite extravagant, you're literally paying somebody else to show for you around. And so economically, that doesn't work very well for most people most of the time. And then buying your own car also doesn't make a lot of economic sense because you spend tens of thousands of dollars, many tens of thousands of dollars on a depreciating asset that you only use 4% of the time. And so what we want to do is combine the best of all of that and say, hey, we can build.”
2020-03-06 · Decoder with Nilay Patel · Recode Decode: Aicha Evans and Jesse Levinson · IDENTIFIED FROM THE TRANSCRIPT
“Well, it's a technology that serves a purpose. But from a customer or a passenger standpoint, you step in and you're being transported. It's very smooth and you're not really thinking about driving. The driving is happening through the sensor architecture, the compute, AI, but that's not what you experience. What you experience is that you're being moved around. And a little bit of, when you think about also the London cabs, we talk about AI.”
2020-03-06 · Decoder with Nilay Patel · Recode Decode: Aicha Evans and Jesse Levinson · IDENTIFIED FROM THE TRANSCRIPT
“That's correct. Well, we actually have been surprised just about everybody who stepped into that vehicle got used to the concept within a couple of seconds. Because we're transporting you. This is very important. We are not building autonomous driving. We are building a transportation.”
2020-03-06 · Decoder with Nilay Patel · Recode Decode: Aicha Evans and Jesse Levinson · IDENTIFIED FROM THE TRANSCRIPT
“Yeah, it's kind of a small one of those, though, right? So it's a compact bidirectional symmetrical fully autonomous, fully electric vehicle that's designed for moving people around cities. There's two seats facing two seats. It's really comfortable. It's really nice. It's a fairly high-end experience. And again, because you get to spread the cost of the vehicle out over people paying for it all day long, the marginal cost of all those components, whether it's the vehicle or the sensors or the computers, the marginal cost per ride is very, very low. So we can afford to make a safe, high-end, beautiful vehicle that's actually going to be more affordable than what people use today because of the utilization rate.”
2020-03-06 · Decoder with Nilay Patel · Recode Decode: Aicha Evans and Jesse Levinson · IDENTIFIED FROM THE TRANSCRIPT
“Depends how you define everything. I think there's a misconception sometimes that we're like, you know, getting in there and hammering out some tires and then building motor. No, exactly. So we have a wonderful global supply chain and almost all the components in the vehicle come from other companies. What we've done is designed and engineered a totally new architecture. It's a different way of putting together parts, most of which are really off the shelf and some of which have some minor modifications for our requirements. And then we put them together in a beautiful new vehicle that's designed for a use case that nobody's really ever done before.”
2020-03-06 · Decoder with Nilay Patel · Recode Decode: Aicha Evans and Jesse Levinson · IDENTIFIED FROM THE TRANSCRIPT
“Yeah, they have. Look, it's not just about building the car. It's about all of the thinking that goes into re-architecting and redesigning the car with a single purpose. AI drives and human enjoy. And when you free yourself of those constraints, it makes a huge difference. You actually integrate faster, you go faster, you focus on the right things. We also think you make it safer. And so it's really about the entire integrated concept as opposed to just building a new car. Last but not least, you were talking about retrofitted cars a second ago. When you start looking at a passenger car from the concept or from the standpoint of a human driving it, a lot of things make sense in that car. The steering wheel, the windshield, how it's modulized and so on. Those constraints are not necessarily applicable or helpful for AI to drive. In some cases, actually, they are not a good thing.”
2020-03-06 · Decoder with Nilay Patel · Recode Decode: Aicha Evans and Jesse Levinson · IDENTIFIED FROM THE TRANSCRIPT
“Someone operat That's right. And in 2014, we didn't know of anybody working on anything like that. And my co founder, Tim, was really excited about that model. And after I talked to him for an hour, I was like, this is a really, really smart way to solve an important problem for society. And we decided to give it a shot.”
2020-03-06 · Decoder with Nilay Patel · Recode Decode: Aicha Evans and Jesse Levinson · IDENTIFIED FROM THE TRANSCRIPT
“Google was working on that, and a lot of the car companies had been working on it or at least talking about it. And the insight that we started, Zeux with was that if you had a new type of vehicle where you weren't stuck with the constraints of passenger cars which are designed for people to own and only use 4% of the time, you could build a much nicer product that was optimized for point-to-point autonomous mobility in cities. It could be safer, it could be a much better experience, it could actually optimize a sense.”
2020-03-06 · Decoder with Nilay Patel · Recode Decode: Aicha Evans and Jesse Levinson · IDENTIFIED FROM THE TRANSCRIPT
“Yeah, so we started the company in 2014, and the insight there was that retrofitting autonomy onto existing cars was probably not the best way to solve.”
2020-03-06 · Decoder with Nilay Patel · Recode Decode: Aicha Evans and Jesse Levinson · IDENTIFIED FROM THE TRANSCRIPT
“You know, he's, I bring a certain set of experiences, he brings another, and then we work to there. But you'll be surprised. I mean, some of the challenges, once you have the concept and the prototype and the idea, getting to market is pretty much the same principles I'm finding in terms of software engineering, in terms of integration, in terms of schedule, execution, go-to-market strategy, and so on.”
2020-03-06 · Decoder with Nilay Patel · Recode Decode: Aicha Evans and Jesse Levinson · IDENTIFIED FROM THE TRANSCRIPT
“Was very attractive to me. And it's also when you're saying ground up, it's really from scratch with a lot of these companies, and we'll go through all the competitors in a second, but you have a computing background, correct? A technical background, but not in autonomous cars. Not enough.”
2020-03-06 · Decoder with Nilay Patel · Recode Decode: Aicha Evans and Jesse Levinson · IDENTIFIED FROM THE TRANSCRIPT
“Mobile is a startup, well, now a company in Israel, an Intel company that works in the autonomous driving space, and they have been doing that for a long time. And so obviously as working as part of the corporate strategy at Intel, paid attention to this space. I think from the get-go, looking also at the stage where Zux was at, taking it to scale, getting to market, and sort of bringing the processes and all of the different thinking when you're dealing with a thousand people and scaling even beyond that, that was very attractive.”
2020-03-06 · Decoder with Nilay Patel · Recode Decode: Aicha Evans and Jesse Levinson · IDENTIFIED FROM THE TRANSCRIPT
“And then met Jesse, some board members, and really learned about Zeus from the inside and this whole concept of mobility as a service fully. Autonomous fleet from the ground up, really, the notion that For AI to drive and to make it easier and safest. And that's very different from what happens when a human being is driving.”
2020-03-06 · Decoder with Nilay Patel · Recode Decode: Aicha Evans and Jesse Levinson · IDENTIFIED FROM THE TRANSCRIPT
“I worked on wireless. And Intel touches many different aspects of compute. And yeah, just Waiting for nothing really to happen, just having fun. And then I That To go work for another big company, that was pretty clear and then received a call. Of the first”
2020-03-06 · Decoder with Nilay Patel · Recode Decode: Aicha Evans and Jesse Levinson · IDENTIFIED FROM THE TRANSCRIPT
“When I was a kid, I played with computers a lot. My mom was a computer scientist. She was teaching me how to program when I was five and six years old. And I was like, this is really cool. And I was kind of obsessed with. Getting computers to do things that people were good at, that computers didn't know how to do. And even as a kid, you're in cars all the time. And so ever since I was a kid, I was thinking, hey, why couldn't computers do this if people can do it?”
2020-03-06 · Decoder with Nilay Patel · Recode Decode: Aicha Evans and Jesse Levinson · IDENTIFIED FROM THE TRANSCRIPT
“Well, I was certainly working on it at Stanford. I don't know if I'd say I was an expert, but I had spent many years working on the technology at Stanford. When I looked at what artificial intelligence could do to help society in transformational ways, I thought that autonomous driving was really pretty close to the top of that list. Driving is such a huge part of our society time wise, economically, environmentally, and safety wise. And autonomous technology showed the opportunity to really make significant progress in all of those different axes. And that was really compelling to me.”
2020-03-06 · Decoder with Nilay Patel · Recode Decode: Aicha Evans and Jesse Levinson · IDENTIFIED FROM THE TRANSCRIPT