 Welcome, so we have 25 minutes to talk about freedom safety How the future of smart cities and autonomous driving is going to be impacted by technology and what that means in terms of our own sort of life choices liberty and so forth Briefly Simon Levine. I'm a venture capitalist at mosaic ventures We focus on series a investing across Europe and I should say I'm a Investor in nexar one of the companies here on stage. Perhaps you can introduce yourself Josh, too Josh Moore managing partner husk ventures. We're also investors in nexar I ran uber, New York for about five and a half years prior to that Bruno Fernandez an advocate of data also CTO of nexar and one of the co-founders So just to sort of roll right into this You know one of the things that often gets discussed is how quickly Will autonomous driving arrive and there's lots of estimates that have been put forward in the market You know perhaps Josh you you know from your uber experience where there's obviously been lots of internal debate What's the view there in terms of the range of dates that we can start to see? Autonomous driving arrive. It's a really good question I think in the in the u.s. It'll vary by city and how the Governments and and regulators sort of see it I think in places like the West Coast the desert where it's sunny and clear 360 days a year you'll probably see it sooner than in Manhattan But what people don't I think always think about is that this isn't going to be a big reveal Where all of a sudden we wake up one day and we see autonomous out there are our cars already have Features that are semi-autonomous controlling speed based on the Vehicle in in in front of us lane management things like that So I think in some ways we're already there today in Manhattan, maybe 10 or 20 years, but I think Bruno would probably know better so there's different levels of autonomous We should probably spend just 30 seconds on that there's famously sort of five levels of autonomous and you know today We're at sort of level two three You know the chap that lost his life on the Florida highway Driving into the side of a truck was it was in a cruise control mode in his Tesla Trying level three which didn't work so well unfortunately tragically for him But you know Bruno, what's what's the kind of necks of you on when we can start to see level four level five In the market not necessarily pervasively, but in a mixed environment perhaps and if you can give us a date That will be even more fantastic. Yeah, I think it's January 24th 2019 Seriously, I think level four automation which you know for folks who do not know about it Is this really when cars drive in safe environments that are actually known advanced we think it's still We'll see cars in the road probably in the next couple of years that actually come this actually Testing that testing that there's a trial going in Chicago right now So it's not like it's unknown now for that become mainstream. We think it's really they get away, right? Even level five, which is the next one in an environment is definitely multiple they get so so several decades away You know just by context, you know, how long did it take for? Cars to get penetrated into the population or a dash to get penetrated Maybe you can sort of help frame that why that's an amazing so we you know We came from search advertising into the space We didn't know much about autonomous cars. We just thought like, you know a lot of data will help and one of these we learn is like Mercedes Benz, for example, launched the first car Build the first car in 1886. It took 82 years until 1968 for the US to have 50% of cars 50% of the households to have cars So that's definitely not exponential like some people are saying now Hey, we'll put it on those cars and we'll come even if you look at more recently autonomous drive even adaptive cruise control launched by Mercedes Benz again in 2002 if you look at 2015 you know with 13 years later that was still 2% penetration you look in the market It's still 2% penetration adaptive cruise control So it's very interesting when I talk to you guys as VCs and I go around like hey It's a soul problem. We got like oh my no straight in my denation cars Do not have these today and they will not have for decades to come So to summarize the sort of view here. We heard it both from Josh the uber task for you and Bruno at Nexar is multiple decades Before autonomous driving becomes pervasive and we can kind of give up the steering wheel And I guess in the meantime what we're saying is there'll be some kind of mixed fleet on the road or heterogeneous Vehicle population is that that that's next ours key assumption about how we design the company is Let's not wait decades for that to come. Let's start actually getting a hybrid environment between humans and cars Autonomous cars if you look at autonomous cars one of the issues they have is like, you know You come to an intersection and there is a human driving another car How do I tell us the autonomous car tell the human driver that I contact you and I have when we drive in it's like Yeah, I want to go I don't want to go you let it pedestrian go you you kind of move your hand you flash the lights all those things we do Humans cars kind of do it. So you need something else. That's what Nexar is It's a network that connects the car so that whether it is a human on the wheel or some autonomous car in the car You can actually start solving autonomy now And so just to put that in context I think there are a billion plus cars on the road today worldwide The fleet is getting renewed. You know, there's maybe 70 million plus new cars each year So just to cycle through what's there is already a 15 to 20 year replacement Cycle and we're saying that given that not every car today is autonomous. That's being put out on the road many decades So But I mean, I feel like I should stop you there I think will you know 100% of the cars be autonomous in Helsinki in five years or ten Maybe not for 100% for it to be perhaps even illegal to drive your own car decades away I agree, but I think it's going to be much sooner To this hybrid world where Trucks going across long journeys are going to be driving themselves where there is an Uber Lyft, etc Option for autonomous driving. Some of these services are already piloting this You'll start to see humans getting less involved right now There are certain cities where you can request an Uber and a car an autonomous vehicle with a sort of a Safety engineer in the driver seat will pull up I think over time that person may be leaves and it's monitored remotely or they're shifting to a different seat I think you will see those kind of changes where it'll be possible for everyone here to have the experience of an autonomous trip Within five years. I think it's further away to a complete autonomous Jetsons type society Maybe on the the open road environment I think I would agree right if you look into what automotive car manufacturers are doing today What is in the road? What is in the pipeline driving in a freeway driving a daytime diamond sunshine driving a truck platooning use cases for trucks Those are going to be solved in the next five years driving in an urban environment with snow with fog You know fiends here for example when they take a driving test They actually need to drive in the snow and sleep the car There's no autonomous car that's going to be able to drive in Finland the snow for decades So I have an iPad here with questions from the audience I encourage you to to prompt me like this, but someone saying why do we need autonomous cars really? So what what is the raison d'etre? Is it safety? Is it because we're lazy and we want to be watching a movie? You know, what is the reason that autonomous cars are important? So you so I'll do the productivity. You know, we are Wasting, you know billions of hours in cars For commutes, I don't have the stats off the top of my head, but people spend a lot of time commuting I think we can all sort of agree to that And if you could repurpose that time to be productive to be able to really engage in Conversations to get work done to watch a movie whatever that's a that's a huge benefit to society Which I think is probably the second Most important thing the first most important thing is safety really so with Next are the reason we created company was actually safe in lives and he came we came to realization There are about 1.2 million people that die on the road every year We drive as a community 10 trillion miles and the 1.2 million miles Sorry one 2 million people that died on the road the cost of society per death is two million dollars So you start multiplying those numbers you go like Beyond the personal drama that represents. It's a huge cost to the economy. So two three trillion dollars So that's a few a few percent of world GDP each year We're losing on the roads and not necessarily and he's preventable Which is that the thing that I think drives us as an XR and many other companies to really try and solve the problem You actually see that the technology the computer vision technology and the radar technology The artificial intelligence required to solve these problems available now So you could be very very cautious to wait decades to roll it out or start rolling out Progressively and start saving lives now and so that was the the mission of next are why you co-founded the company Can you just spend maybe 30 seconds telling the audience a little more about What that instantiates itself into today and the roadmap for the company going forward So the best way to look at our company is if you go to New York, for example We have you know a large number of Uber leaves Juno's get How many cars on the street on any given day are you next in a good day They're probably about five six thousand cars in New York that are equipped with this Which is roughly about two percent of all the hours of driving What these does is what are you driving while we're trying to do is these cars to be connected to each other So the humans can talk to other humans in this case the computers talk to their computers when they drive around They're issuing warning. So right now we get in tens of millions of warnings. Does this people drive? They're actually getting those warnings. Hey, there's something that is 15 seconds ahead of you and that's actually really critical because all their Technologies that can help you like a radar a lighter again What you find in a Mercedes Benz we find a BMW you find in Tesla even has a range as far as the camera goes You just go 150 meters Maybe even 200 meters. I was just really far But in an urban environment your blog to have a car in front of you There's this maybe a pedestrian crossing. You don't see it That's exactly where we're trying to get to and that's the world that you can experience already today as an Uber driver in New York So it's a kind of vehicle-to-vehicle mesh overlay network kind of like air traffic control that's Augmenting the driving experience for those Uber and Lyft drivers. Is that correct? It is correct It's air traffic control, but imagine, you know You know, there were 10,000 airplanes in the air at any point in time So right now, you know, you look over at our skies 10,000 planes on the road at any point in time There are 250 million cars so air traffic control is centralized you can centralize it Well, we're trying to build is a decentralized Cooperative overlay between these cars actually regulating the traffic regulating safety and so You know at what degree of penetration do we start to think that there will be Safety benefits and how big do those have to be for it to make a difference? in in our world with because you know So the answer would be if you were trying to just do DSR That what it's called vehicle-to-vehicle technology or Wi-Fi Which is the approach that the regulators in the US and Europe have tried for many years You would really require 25 30% penetration, which is very very high So it's a lot of people have got a co-operate and it's a lot of time It's a lot of money a lot of time right so that's you know There are studies that say it's a cost to the US economy alone about 200 billion dollars over a 40-year period It's about 2.7 to 3 billion dollars per year at the peak of 2023 and actually takes seven to eight years for that 30% to happen Now our approach the next our has been V2B is just part of a solution. It's not the solution You need a camera you need other sensors to actually help you when you do that and vehicle-to-vehicle Over a cellular with a camera with all these things that are available in your phone that everybody has we seen that at 1.5 2% that's 10 times less in terms of percentage right we start seeing that vehicle-to-vehicle Actually has impacts in preventing collisions and we measure it We have 25% recorded less collisions on people who are exposed to these than people who drive in a control bucket And Josh when you are managing the uber fleet in New York Obviously, you know 5,000 of drivers a day using nexar was it a popular product with the drivers? Was it did they feel safer in some way shape or form? Yeah I mean so just to back up what the what the product is the drivers are putting in their cars on a phone a little Video camera so the reason that the penetration can be low But still have that outside impact is because from that screen it can see all the cars on the roads You're not just getting the vehicle that it's in but also anything that that's in line of sight the safety benefits Are not theoretical the insurance companies that New York City uber and Lyft drivers have to buy insurance from Which can run four or five thousand dollars a year We're getting 10% off just for using this product because the insurance companies right away saw that this had a tangible impact on Risk and they priced it accordingly so drivers were thrilled it gave them the security of having the camera running So if there's an accident there's some kind of evidence, but also the insurance companies saw that They understood the technology and the benefit and the kind of alerts the drivers could get which could ultimately keep everybody Safer and we're willing to reduce the rates because of it Okay, so the title of this panel was freedom or safety and what we've heard The two of you say is if we can build an autonomous co-existing fleet We should be able to prevent many of those 1.2 million deaths and not you know countless other accidents maiming an injury Property damage and so forth. So that's the payoff the cost is Quote-unquote freedom and so there's different parts of that But we should try and unpack it for the audience because there is a a moral ethical element to this discussion And in particular if you're putting all these cameras Real-time creating an overlay that ultimately is Creating a kind of digital representation of the public space in real time There's a privacy issue. That's so why don't we spend a few minutes just on that question, you know From an uber perspective from a nexar perspective, how do you think about privacy and who owns the data? So I don't work at uber anymore. I'm not going to speak for them I think in in society there are trade-offs that we implicitly make you were you were saying this yesterday when we met You know, we have to drive on on roads and stop when the light is red and start when the light is green. These are sort of These are we give up our liberty just to go through the right light I can't just drive anywhere. I want I have to obey traffic signals They're sort of these societal norms that all existed when we were born and so we sort of accept them You know, I think the the trade-off will be easy saving lives for a little bit of freedom that says for example You can imagine when my car pulls into a stoplight. I lose control for a second It sort of you know locks me down or I'm just sort of making this up But it controls when I do the turn or it tells me that something is coming Those will be worthwhile trade-offs, and I don't think that to me seems like a no-brainer And Bruno the way that we we at nexar have thought about things is the data owned by Governments is it owned by the users themselves? Is it some kind of shared entity? Our model is the data is owned by the users and from the very beginning, you know, it's true We had a governance approach does Can we share the data with them? Whether it's for enforcement or any other purposes and our policy has been very clear We do not the data is owned by the user and that's it But the issue I think is more profound than that right is we can now extract out of the data things like What are the traffic signs? Where are the parking spots? How many counts of people are if? You know somebody came to the company could actually start extracting faces and license plates and recreate your whereabouts So I think them the more fundamental question to us as nexar is how do we want to do that? Do we want to let the governments run these things for us? I'd be looking at UK right in London today the three to five million cameras on the road It we're already living in the panopticon in London. We just don't know it ultimately. Yeah, so the question is whether you want To be the governments that actually ran that and who controls the government or are you okay? That is an environment what is open where users actually have access to the data themselves They can use it to defend their civil liberties and we had a few cases actually in the US with you know The police the cops stopping some of our drivers at a traffic light after doing what is called a California role and the driver say well Here's the footage Demonstration that I haven't done it so to a degree what feels like it could be a bad side effect Actually, I think for the people who truly care about their civil liberties. It tends to be actually a good thing so transparency empowers citizens or empowers non-governmental entities to Provide a record of exactly what happened okay the what about sort of folks who Kind of think about their car as a symbol of liberty and kind of driving down the desert highway You know, is that something that our grandchildren are gonna enjoy or instead they'll be reading a book and Catching up on Wikipedia instead. I sure hope they don't have to drive. I have a three-year-old I've basically made a bet that she'll never have to learn how to drive and I think I think I'll probably be be right about that And certainly as a parent that's sort of a wish of mine That's an incredibly dangerous situation when young people Learn how to drive and they go out so to avoid that, you know entirely I think would be a really great outcome and Let's assume that Your daughter never does have to drive one of the questions we have From the audience is Her car that an algorithm is driving with a network overlay provided by Nexar or someone else Has an accident Who's responsible for that from a legal perspective from a moral perspective? Is it The algorithm that was driving the car Well, I think I'd have to know more about the accident of my you know as long as she's okay I guess I don't really care But as far as the like general sense of whose fault is it? It's not always easy to ascribe fault in accidents today, and I think there'll be a case-by-case I think there's sort of a myth about what will the car decide will it hit the old people or the young people? That's not really how it works There will certainly be Malfunctions in the idea of the the Tesla accident that you mentioned at the beginning a white Truck sort of pulled in I don't know all the details But I know that it was white and the and the sensors didn't quite understand See it against the horizon. Yeah, and so that's a bug and In that case at at the same time the driver was not Using the autopilot the way that it was supposed to be used You're supposed to be alert and sort of hands on the wheel And so I don't know but I think it'll be case-by-case What I do know is I'd rather trust Machines even if they make mistakes when they're ready for prime time I'd rather trust the machine than everyone's sort of own fault and falling asleep and being drunk and being Distracted and all those things. Yeah, so in the aggregate the world will be a better place Just one or two other questions. We have a few minutes before we have to wrap up From tarot asks who will own the the data and the platform is this another kind of Google and Apple Amazon Network effect, you know, how does citizens or the public at large? Feel that their side is being fought for right so As I mentioned next are we all the data that people record belongs to those people What we are saying is we would like to use the data to develop autonomous drive technology We use the data to learn the algorithms, but we do not actually use the data for any other purpose If we were to provide the data to an insurance company It would be for example because you was as a person driving says like I want to benefit like in New York, right? I want to get a 10% discount Then is you who obtain into hey next are can you please share my data with insurers so I can get a discount? But we are not going around and sharing the data The only artifact that we create is really the one that generates safety and generates welfare in our society by removing those Lives that cost two million dollars each Going back to how we started the conversation At what level of kind of density do we start to believe that Those lives are going to be saved. Are we there at three four percent in New York? In the next year is that already going to have an impact on road safety in the city? I think the wild factor will happen in New York Within the next six to 12 months When we will have a car every minute a car every 200 meters a car in every intersection all the time And when you come to make a left turn when you come to make it right turn You don't know whether there is a pedestrian or there is a car and we can warn you That would the first time that somebody is that worried they'll go like wow I couldn't see a cross the corner I couldn't see there was something behind that track. So the point is it's not a theoretical discussion about 12 months Yeah, this is in the next six 12 months and Hopefully that saves lives I'm trying to think if there's anything more that we can add to that Should everyone download the next our app on their phone right now. I should feel free to do some Probably give a little in any market. Okay, wonderful. Well, thank you very much Josh from task Bruno from next are really appreciate being here on the slush stage This is a real fire. It's actually hot. I Just thought everyone