 Welcome everybody to the session on vehicles and data today. Thanks for joining us. I think that what I'd like to do is just go down and ask each of our presenters to do a quick introduction. Maybe three or four minutes about what they do and how they're involved in this space. I guess I'll start that off. I'm Stephen Zoff. I'm the Executive Director of the Center for Automotive Research here at Stanford. Backgrounds in the automotive space, eight years at Ford and BMW, and then with the US DOT on energy efficiency regulations. This is Kat Fe. Mark? Cool. We have a quorum of BMW guys here. Yeah. Listen, listen, set up that way. So I'm Mark Plachon. I'm starting up a new venture fund called Icebreaker Ventures with subtitle the Autonomous World Fund. So all things smart mobility and transportation related, especially autonomous car. And more importantly, how that ripples out and affects everything that touches and depends on transportation. It's part of what that venture fund is all about. Most people around here know me as the BMW guy because I was the, we set up BMW Ventures about five or six years ago. And I was the partner here in the Valley sort of helping out on that. And fell in love with how important autonomous driving is going to be to change and improve transportation and cities. Adam? Adam Langton. I'm with BMW North America. My title is Energy Services Management. And so in my role at BMW, I help develop energy products. So I do research and do product development in the energy space for products that are related to our electric vehicles. And that is primarily smart charging and stationary storage. Prior to joining BMW, I worked for the California Public Utilities Commission where I worked on electric vehicle policy and carbon cap and trade policy as well. Good morning. I'm Jay McFarland. I have not worked for BMW. Do you drive one of those? I own one once. I currently work at Lawrence Berkeley Labs and have a joint appointment at UC Berkeley focused on sustainable transportation. My background is in control systems and artificial intelligence and high performance computing. And I have worked for General Motors and helped launch the on-star system, which was the first at scale telematic system. And then that got me into maps and worked for Navtech, which was purchased by Nokia, which was then actually purchased by BMW Audi and Daimler. So I actually have worked for BMW. So I most recently moved to Lawrence Berkeley Labs. So that's my history. Okay, fantastic. So I think I wanted to start this off by just asking you all to say, today, where do we stand today? What is the most effective way that companies are using data that's coming off of cars? Mark, what's where do you think? You know, I think the number will start with a number. I think the number that's most important to think about the data as we're going to talk about it is how many lives saved. Because, you know, right now, cars and guns each kill about 40,000 people in the United States. Solving the car problem is probably easier. So let's get on with it. I mean, obviously data is the fundamental part of autonomous driving and the way it can then be used to implement improved safety as well as efficiency and other things is I think the motivating factor behind the data. I guess I'll offer maybe a local perspective on how that data can be used. I work on a pilot with PG&E where we're testing out smart charging from electric vehicles, trying to align the charging of electric vehicles with times when the grid is not congested or times when there's low cost and avoiding times of charging when electricity is expensive. So what we're able to do is in that pilot, we actually don't rely on charging stations for the data. We rely on the vehicle system for the data and we're able to communicate directly with the vehicle to get the vehicle to start or stop charging for that. So in this particular example, that vehicle data can be very, very useful and effective for managing the charging. We can understand how much charge somebody needs. We can get information about when they're going to depart on their trip and then we can use that information to optimize their charging. And so that's our side of the data. And if we combine that with data from the grid, from the utilities and other folks, then we can really get optimizations that really support the grid. Okay, so I'll come at it from a different perspective. So vehicles today are generating tons of data as well as your cell phones and map companies and the like and the Googles and the Apples are using that data right now to do the traffic and do all the routing, which is managing a lot of our traffic behavior right now. So that level, which is like pro data, which is, you know, just a small portion of the vehicle data is being used currently. Now, what's happening is now where these networks, when the vehicles get connected now, now we have an opportunity to extend that into the vehicle bus data. And so a lot of startups right now are sticking ODB to devices into the vehicles, which makes some of the car companies happy or not happy as it may be to get more detailed information. I think that data can be used for us to learn about how we experience the road network and learn more about how to control automated vehicles. But I don't think anybody is really doing that yet. I think maybe the car companies are, but most of the detailed bus level data is hidden behind the firewalls of the OEMs. But what's also interesting is these vehicles are so powerful with their sensors now, all of that data now can be put together to create the high definition map, which is what automated vehicles are going to need to do look-aheads when they are in their environment. So there's all kinds of opportunity coming. Everybody's working on it right now, I would say. Jane, I think that's a great segue, and especially the two of you coming from an OEM perspective. The number that Mark had shared with me on the amount of data created by AVs is on the order of four terabytes a day for just a couple of hours of driving maybe. And so that, to me, raises a number of technical challenges. One is just, well, how do you get that four terabytes off the vehicle into somewhere you can do something with it? And then two is, what do you actually do in terms of storing and analyzing that data? I mean, I have spent the last five years digging through that kind of data. So one of my passions right now for researchers in this room is quality in big data, because we don't manage that well. And I think we're going to have that in a big way in these new digital environment connected car environments. My personal opinion where we'll probably get some interesting debate on is that we're not going to move all that data. We would be foolish to move all that data up there in real time. So we are going to end up partitioning the data. There's going to be a lot of work done on the vehicle if we instrument them properly. There'll be potentially work done at the edge at the south tower level, and then there'll be work done at the cloud. And so there's the challenge is how do we partition this properly so that the work gets done in the right places so that we're not moving all this data all over the place because it's expensive to move, it's expensive to store, and it's expensive to analyze. So we're going to have to be really clever about how we manage the data going forward. I guess what I would add to that is for some of the projects like the energy projects that I work on, you don't need all that data. You don't need an enormous amount of data. You need a certain subset of the data that James is describing. But you need to have good quality data, you need that access to that. And in terms of using it, you need to make sure that you're respecting the privacy of the individuals and that you have cybersecurity elements in place also. So I think that's the other important part of how you store it, how you treat it. Yeah, I think that number, and I found that number and threw it out there, you know, this four terabytes to drive a car around is crazy. Obviously, that's an Intel number. And they put it out there as sort of the lead number so that their management can talk about how great their computing is compared to NVIDIA. Fine. The reality is, I mean, when you look at, you know, different applications need different data. But when you look at, for example, video cams that are looking out at the scene or video cams that are looking back at the driver or both, they only really need to, I mean, they're always generating the data, but they don't save it except 20 seconds before an accident and 20 seconds after an accident. Or 10 seconds before and 10 seconds after, you know, a hard stop or an emergency server or sort of interesting points to study. Otherwise, they throw the data away and don't worry about it. If you're talking about mapping, you know, Tesla cars, for example, are always collecting all the video they see, but they're comparing it against the map that they've got stored. And they only save the parts when suddenly it's different. You know, hey, the road's not where I thought it was going to be. Because there's a cone, there's a barrier, there's a stopped car, there's something. Okay, that's interesting. I need to send that upstairs so it can then be processed and redistributed to everybody who's got a map so they get a new map. But you only need a little tiny piece that this changed. So, so a lot of, I mean, there's a huge amount of data potential, but smart usage ends up sort of saying, I only need the pieces that are different. And that differentiating needs to happen locally and needs to happen really fast. You know, if you're trying, and there are real-time things and there are not real-time things. The mapping stuff you can deal with in seconds or minutes later, you're trying to put on the brakes before the car. You hit the car in front of you, that obviously has to happen locally. Nothing goes to the cloud. It's a non-deal in microseconds. So I think, Adam, you brought up something interesting there, the idea of customer privacy. And of course, nobody wants to see their data used against them. And this could be unintentional, it could be a leak, some exposure of data that's stored in the cloud. Or it could be something a little bit less perceptible, something like a rideshare company deciding how much to charge you, simply based on knowing your willingness to pay to a high degree. So what do you think that sort of the comfort level is in terms of people revealing data either intentionally or unintentionally? That's a good question. As far as like an individual's concern, I think it depends on where the individual is. It seems like in Europe, there's kind of higher awareness and sensitivity to how their data is used. And in the United States, it seems like from a consumer perspective, there's, at least right now, there's less sensitivity to that. There's less concerns about that. From like a business perspective, I guess the issue, it raises issues with how do you work with another company then? How do you share data with them and make sure that they're not going to use it either against you later on as a corporation or against your customers? I feel like from where I sit and I'm kind of new to this space in terms of the data, there's not a whole lot of clear rules on that, a lot of clear protocols and things like that between different companies that make that, that increase the comfort there. So it presents a challenge for us when we're looking to work with somebody because then we need to make sure, who are all the people that they can share it with? How do they manage it? How do they share it? And then, you know, without having a clear set of rules there, it slows the process down for working with them. I have a question. I want to ask a quick question related to data privacy. This is a real example. How many people think the, how full is your gas tank, your fuel gauge level, or your battery charge level is a big secret? Who thinks your cars fuel level or state of charge of the battery is a big confidential secret that you're unwilling to share? Who's willing to share the level of their gas gauge or the state of charge of their battery? Would you? Really? No kidding. None of you are Germans. There's a real battle we had at BMW. If you want to pull into a charging station that is allocating power among various cars and there's a limited amount of power to allocate, and it would make sense to sort of give the guy who's most empty more and the guy who's most full, you can charge them slower. They're all going to be there overnight parked in the basement of the hotel or whatever. And so we had a charging network that said we'll prioritize based on how empty you are and how urgent you need it. I mean if you're leaving right away then we'll give you a full charge otherwise we'll trickle charge you slowly. And that means obviously each car has to share its, how full am I? That was not allowed. BMW was totally unwilling to share that with the charging operator because that was private information for the consumers. And yet it was to the benefit of the consumers to be able to get a smart charge. Absolutely not. It went all the way to the top. I can't explain. If I can offer a little counter to that. So BMW has a privacy agreement with each of its customers and it's in regards to how we use their data and how that data gets shared. And so whatever, if we're going to share any data, there's a chance that you need to look at that privacy agreement and see if you need to change that and then you need to go back to the customer and get any privacy agreement. So there are some challenges with sharing that that you need to get the customer actually engaged and get that kind of commitment. I don't know if I wasn't involved in this particular issue so I don't know exactly. But that'd be one of the practical concerns is now you have to go back to the customer and get them to sign something. And then you also have to know with whatever company you're working with, how are they using that? And how are they going to protect it? How are they going to store it? Even if they're planning on, they have no intention to profit from it, you still have to make sure they're protecting all the cybersecurity and things like that. So I think those are challenges that we're just beginning to wrestle with. So I actually wanted to ask a question to the audience before Mark had jumped in. So Mark, here's really a stated preference question which is the comfort level of the audience. But I wanted to ask more of a revealed preference question. Is anybody here participating in the California road use pricing trial? Basically pricing for VMT usage? Anybody? Really? That's actually surprising in this group. How about in progressive snapshot discount or some other usage-based insurance? Wow. Okay, last one. How many of you participate in some crowdsourced traffic system like Waze? Anybody use Waze? Okay. All right, quite a few. So yeah, please. So let me get at that one. So people that have to manage traffic are working on survey data that's like five years old. So understanding mobility in the Bay Area, for example, who has all that data? 80% of mobility data is locked behind firewalls. The Googles and the Apples and everybody who is redirecting or directing us on our routes. Now think about what's actually happening here. Everybody now has dynamic routing on their phones. So who's managing traffic in the Bay Area? The Googles and the Apples and the Waze. And, you know, the cities have only a few levers to pull. They have ramp metering, they have HOV lanes, things like that. But those apps are redirecting thousands of people in minutes. So this is a huge problem is that the people who need the data to really understand the dynamics and the mobility issues of our freeways, and they're getting worse and worse, they don't have the data. So and what happens is everybody stands behind the privacy problem that we can't give the data. The lawyers will say we can't do it. You get to, you know, I have spent seven months working with someone already trying to get some data about mobility. Anonymized data, but we can't get over the hump. So just one more thing and I'll get there. So one of the things that I'd like to see is I'd like to see us to be able to decide where our data goes. Because we go, I agree, I agree, I agree, I agree every day on our apps and give away our location data. And the people who actually need it don't have it. The researchers who want to help change our transportation system, the cities, the government don't have that data. So I live in a state. Well, it's a tiny percentage, though, of all the people that are out there on those freeways right now. And of course, if you're in Europe or in the UK, those systems exist as well. I mean, it's very expensive at a certain time today, especially if you're coming from the other side. Well, in Europe, they don't put up with this. They just say you will give us your data. It's really the government's have the data. So I think that at least in the US, I know of three specific ways. And, you know, this is an energy conference. So I'd like to tie this back to energy goals. So there are three specific ways that at least some level of data is being used in pursuit of energy objectives. The first one of those that I thought I mentioned was actually looking at high level odometer readings. So in the CAFE standard, last year, US DOT switched from using an HTS, the travel survey data, to using odometer reading data to get a profile of vehicle usage to estimate benefits. That's one at least high level way. A second way that I wanted to mention was the road use pricing or VMT based fees. That's another thing that we're starting to see more and more of. Another one is what we're seeing with the recent modifications to the OBD standard in California. So starting, I don't know if people in the audience know this, but starting in the next couple of years, California will start to gather long run and short run in use fuel economy from vehicles. They know actually gallons burned, not just trying to estimate it from miles traveled and window sticker values. So I thought I'd ask our panel, you know, when we're talking about this environmental space or energy space, what are ways that you know of that there's already some level of activity here using data in the pursuit of environmental goals? Jane? A really good example is some work that's going on at UC Berkeley and the PAF organization looking at connected automated cruise control, so for energy perspective. And it's a very interesting study is going on that, you know, we as human beings are not very good at car following. And there's a study where they put a bunch of vehicles in a circle and had everybody follow each other and it goes unstable almost immediately. And then you can do automated cruise control, which is interesting, but it has a lot of variation. But if you connect the vehicles, if you put a connected automated vehicle in the flow of the traffic, it smooths the whole thing out. And that can have a really great impact on our energy use, because we have a lot of variability in these congested quarters. So that's a really good example where connected automated vehicles will really help with our energy use. Adam, how about you? I was just going to say it sort of echo that there's a local company that started by guys out of Stanford and Tesla, Peloton. Many of you probably know Peloton. That's about truck following truck, platooning trucks. And of course, that's the cool thing there is everybody knows the truck that's following gets about a 10% fuel savings, which is a big deal. But the front truck also gets a fuel saving. It's not like the front truck has to do more work to pull the back truck. The front truck gets a benefit of about 4%. The back truck gets about 10%. The combined benefit is huge. So these things work out. So on another angle on the energy side, I'd say on thinking about smart charging, the utilities, at least in California and nationally eventually are looking at getting a lot more renewables on their grid. And with renewables, particularly solar and wind, you can't control when that generation happens. So the utilities are looking at a whole different set of challenges to managing their grid. And electric vehicles can be a load resource that helps them balance that out and helps absorb those renewables. One of the ones that I'm particularly interested in is the idea of getting vehicle charging to align with solar in the afternoon. And that would be a case where vehicle data could be really valuable for doing that. If you're going to get vehicles to charge during the afternoon and absorb a lot of solar, you're probably going to need them to do less charging at night. So now you're going to need to manage multiple charging events to get that charging to actually take place in the afternoon. And you're also going to have to get somebody comfortable with the idea that they're not going to be full when they wake up in the morning, but they will make it to work and they will then be able to charge and maybe just get all solar energy in their vehicle. That would be a case where you need a lot of data and a lot of customer engagement. And maybe the customer engagement is even more important than the data aspect, but that's a case where you can take that data and you can use it to get the vehicle charging to happen where you want it to. So I think that's one where the vehicle becomes very important to making that happen. I just said a very nice story about saving energy. And I think we also need to think of the flip side that if we get automated vehicles, we could very potentially have much higher energy use because people are willing to work farther away from the workplace because there might be zero occupancy vehicles out there. And because we may never go to car sharing or ride sharing. So there's an opportunity on both sides coming and we have to get ahead of that. We have to start now thinking about it so that we don't end up creating more vehicle miles travel. I think I have a couple more questions. The goal is more vehicle miles travel. That's the economy we want growth. Well, I think that then we can agree to disagree and let's go. All right. So I'm sure we have some burning questions from the audience. If you wouldn't mind coming to the mic, please, so that we get your comments or questions on the video. Yes, sir. Oh, OK. Well, I have a question then, which has to do essentially with who's benefiting. So we're seeing, you know, in Silicon Valley, every day it seems that there's a new partnership. So for instance, the BMW Intel Delphi mobile I partnership or the here partnership, some of the others that we've been involved with. From an outsider's perspective, it's hard to understand who's really benefiting from the data that's being generated from those partnerships. So if you're, let's say an investor, for example, and you want to know who's really benefiting from this partnership, how would you choose to evaluate that? I mean, the goal, the goal has to be that the consumer benefits, right? They get cheaper transportation, safer transportation, more mobility, cheaper deliveries, you know, less congestion, fewer accidents. The whole structure of those platforms is for the benefit of consumers. But the only way consumers benefit, I mean, if consumers benefit, then those companies also make money. And that's the virtual circle. They get tangled up in controlling their data and preventing consumers from doing what they want. But we'll work through that. Adam, how would you evaluate that? Who's benefiting in these partnerships? Well, it's hard to say in those specific examples, because I don't work on the autonomous side, so I don't have as much insight to share there. But what I can say is that I do agree with Mark that the customer needs to be, in this case, the driver needs to be at the center of that. But to make it work, the driver needs to be at the center. So when we do our smart charging programs, we're always looking to keep the driver at the very center of those programs. We want the driver making the decisions in those programs. We want the driver to benefit. Because if the driver doesn't benefit or they don't have control of it, then they may not participate. And if they don't participate, then you've not achieved your goal. And that's in a smart charging context. I think they'll seem broadly those same concepts apply to other areas, potentially on autonomous as well. Jane? Well, I kind of think at the higher level, because I'm at now in the government level, is that we as the citizens have to benefit from this. And as I've already said, we're not because we don't have the data in order to design our transit systems better and design our transportation systems better. So I do think we have to work through it, as Mark said, and really find out how to anonymize effectively so that we can pull that data out and get it together. The OEMs will benefit as well if we do that properly. The HD map can't be built by just one fleet of BMWs. It has to be built by a fleet that crosses all of the links on the map. So at some level, everybody has to participate together. And as we do that, we need to find a way to engage government public entities in that process as well. Boy, that sure sounds like government wants to create a monopoly on data. I don't think we want to give them that much control. Why should we? I mean, why can't a consortium of three automakers, you know, I mean, if BMW, Audi and Volkswagen want to get together and share the day and make a map, Tesla wants to get together with somebody else and make a map. And three startups want to make maps and compete to sell their maps to others. There's lots of mapping startups now. Why shouldn't they? And if they all have to aggregate it together under government control, no thanks. There's nobody suggesting under government control by any stretch of the imagination. That's what you said. No, that's not what I said. I control it. It says that we need to understand mobility in our cities. And we need to we need to learn more about where people are going and why they're going there. I don't have a direct counter to that. But I guess we talked a lot about the data that's flowing out of vehicles and but there's also other data flows that are critical to this. The government actually has a lot of data that we would like to get that would help us in particularly in the energy space. So, for example, when it comes to smart charging, if we had detailed information about where the grid was congested or about renewable energy in real time on the grid, we could time the charging to align with that. That's data that utilities have that they don't share. So that's an example where if if that data was flowing out from at least from regulated entities, we'd be in a position to do a lot more as well. So I guess I just want to say that that can cut both ways. So what kind of tradeoff is that forcing you into it? So that let's grab that specific example. So you guys are running a program actively right now on effectively demand management if I understand correctly. And so so the lack of availability of information about the grid is forcing you to make some some tradeoffs, right? Yeah, you can rely on more general information that that is publicly available. Kaiso makes information available about what's happening on the grid, renewable energy and other elements. But if you if you don't have those details, then you we just simply can't do some of the optimizations and create some of the benefits. We just we can make estimates of where we think those happen. And that's useful in a research context in a limited sense, because it proves what functionality you have. So you can say, oh, if if I have this hourly data, I could do this. If I have this five minute data or real time data, I could do this. I think that's Adam, that's a great one to talk about for a moment. Because I mean, when I had my BMW electric car, the the rates tell you to charge starting at midnight, because that's your cheapest rate. And they penalize you if you charge in the afternoon because that's your peak rates. And yet, these days, the ISO actually wants you to charge just like you said in the afternoon, because that's when they've got lots of solar to get rid of. But if they want to charge you the highest price, then obviously incentives charge at night when the electricity is at 10 cents. So so we've got government sort of saying charge in the afternoon when it's the highest price. Is that because they want the revenue or because they really want to manage the grid? I would say it's more of like a legacy of the, you know, the rates are hard to change and takes a long time to change those. And we're coming from a historical system where the costs were highest in the afternoon. That's a slow thing. And then there's also concerns about, you know, customer fairness and equity and things like that that are trying to manage. But part of the reason those issues come up is because we don't have that data that allows us to do a better job. Yeah. Yes, sir. Hi. I'm old enough to remember when elevators had drivers. And we all take autonomous elevators pretty much for granted. Now we don't feel scared when we get in one. I live in Mountain View. I've been seeing Waymo driverless cars on the streets where I live for four years now. And I am sure that autonomous electric vehicles are going to be everywhere in 20-ish years. So my question has to do with if we go from a world where the average car is driven an hour or two a day to a world where the average car is driven eight or 10 hours a day, we will need about 25% as many cars as we have today. How are the automobile industry companies in the industry going to be able to cope with a reduction in demand for their product of 75% over a period of 20 years? Now that's a data formed question, perhaps not the kind of data that you were thinking about. But I think it's a critical issue because shrinking industries get desperate and they do crazy things like the coal industry is currently doing. I think that the winners will all be the autonomous companies. But I'd love to hear your take on how the shrinkage of the car fleet might play out over the next two decades. So I guess I'm not a panelist, but I'm going to feel this one anyway using the moderate privilege. Because I think that that's a very important point that you're bringing up. The idea that you need 25% the number of cars, that's a figure that you see different estimates of, does not mean the same thing as 25% of the vehicle sales. So vehicle sales are basically a function of three things. It's basically VMT, vehicle miles traveled or passenger miles traveled, let's say, the ratio of PMT to VMT, which is just another way of saying occupancy, and then how long do the cars last? So durability. And if the durability doesn't change that much and occupancy doesn't change that much, that means that vehicle sales will roughly track passenger miles traveled, right? So I don't see a shortfall in vehicle sales anytime soon. That having been said, that doesn't mean the profit margins will be maintained. So if right now the profit margins are not great on making cars, if we move to a more heavily mobility focused system, you can bet those margins will shrink. I think that your final point, desperate companies do desperate things is possibly a valid concern, but I don't think it'll be as a result of vehicle sales. I'd say desperate companies do creative things. You know, I think you look at automakers today and they're all trying to transition from being automakers to mobility suppliers. And they're launching their versions of Uber or Lyft or AAA just launched Gig and Volkswagen just launched another one. And so you're seeing companies saying, how do I get a different piece of the pie? Do I sell subscriptions for mobility? And you buy your transportation from me, but you also buy a lot of other services from them. Do you get your energy management and you back up power from the car company? Do you get your solar panels from your car company? You know, so I think you're right. Car companies are changing. And I agree. It's all going to be tied to how many vehicle miles you go. Electric cars may well last longer, you know, so 200,000 miles in my last 400, 500,000 miles. But then I'll need a lot of maintenance, a lot of service, a lot of fixing. You put cars into multi-use sharing and people abuse the heck out of them. Take a look at a rental car. So things are just going to be a little different. But those companies are getting very creative and very energetic about these investments in new business models. The example I can provide on that from the BMW perspective is we launched a car sharing service called Reach Now. The original version of that was actually here in the Bay Area. It was called Drive Now. So Reach Now is based in Seattle. So that's the example where BMW is forming a new kind of model for providing mobility services. The scary part is the data says we're not yet getting rid of our first car, but the ownership of second cars is dropping like a rock. And people are using Uber or Lyric as their second car. Yes, sir. Jambuzel with CalSTART. This term of autonomous vehicle kind of gets thrown around a lot. And I think we're at a point where what does it really mean? There is the level five autonomy where I think I'm even confused myself now. But where somebody is, you're in the car and the car is driving itself. And under a scenario like that, you could see an incredible increase in congestion. People willing, doing getting a lot of stuff done while they're sitting behind their wheels in the car. Volvo runs an ad where you push a button, the car takes over, you take a nap. And so you could easily then see this encouraging suburban expansion. More people living out in San Joaquin Valley commuting in the Silicon Valley. So there's that scenario. And then there's the scenario which you've been talking about more, the autonomous where we move a lot of people inside and move to a subscription service. Your thoughts about, and I know it's largely a data panel, even as it relates to data on public policies to encourage that we move in that right direction, that we get reduced congestion, we get a better mobility system, we get zero emissions, we're addressing the climate threat. What are the right public policies that we should be thinking about at this point? I'll get a start on that. I actually work on a program called Smart Mobility that is sponsored by the Department of Energy. And the challenge here, this isn't a system of systems. It's a very complex system of systems. You've got urban planning issues. You've got infrastructure issues. You've got multimodal. You've got the pesky humans in the middle of all of this just deciding I'm not going to do that today and not be predictive. So it's a very, very complex system of systems. And we need to start now thinking about these things, because one of the behavioral things is you can't start charging something. It's very hard to start charging for things that people have had for free before. So policy has to really start thinking about these things now. Well, research says that. We'll just say researchers say that. Parking on the street used to be free. Then they put meters in. I think it's an economic term called loss aversion, which I think is supported in the literature. So now I forgot what I was going to say. So anyway, I think that we have a lot of thinking to do about how to guide this. I am a Debbie Downer of Automated Vehicles just so you know, because I think level five is not ten years away. I think it's decades away because of the dynamics of the environment that it has to operate in. So the Waymo's and the like are great. There are little tiny pods that run around at 25 miles per hour when we start really getting level five. But a lot of people don't think level four will ever exist. And level five in a mixed heterogeneous environment is going to be very hard to manage because you're going to have to put intelligent agents inside those vehicles to operate cooperatively with all those pesky humans that are not automated. So I have the downer side of that utopia of all these little vehicles driving around because there's so much dynamics in the environment. I know that Mark will disagree with me on that. And computers will never win and go. All right. Yes, sir. A couple of comments. One, I think a lot of the conversation has been focused on the U.S. So I mentioned 40,000 deaths in the U.S. There's actually 1.2 to 1.5 million per year globally. And also the systems in all the other countries of the world are way different than they are here. Which brings up my question. So I wanted to get from my house today to here. That's all I care about. I just wanted to get from here to there. So that's what I want out of automated transportation. And in other countries, bicycles, for instance, are a big thing. I was just in Chicago. Didn't use a car for a week in Chicago, right? I used a transit system. I used Uber. So I'm just wondering about the data and how that would fit into my scenario. Getting from point A to point B might be bicycle, might be train, might be Uber, automated car, and how that would play into that. Well, I think that there are three big pieces of the sort of this utopian mobility future that we all really hope happens. And I would say first thing is we need to make reliable vehicles. And we're already doing that pretty well today. There are a lot of companies that know how to do that. The second piece, I would say, is developing a viable package of, let's say, automated vehicle sensors and algorithms that can operate. And then the third piece is essentially demand modeling, which is knowing where people are, where they want to go, and critically how much they're willing to pay. And that piece is held very closely. The second and third pieces really are held very closely by just a few companies right now. So I think that when you talk about really your question and the prior question about how do you get from point A to point B and how do you manage potentially an explosion in VMT, what I would look for really is use of curb space going forward, at least in urban areas. And the freeing up of parking spaces could be used for new travel lanes, right? Could be used for bike sharing racks, could be used for bike lanes, could be used for sidewalks, could be used for new retail space. There are dozens of things that you could do with that new space. And I think that really that's going to be the key to figure out how easy that mobility is in the future. Yes, ma'am. Hi. Ruth Marino. I just want to know what type of data is out there right now that would help both the private sector and the government to determine where to put new vehicle charging stations and how many to put out there? We've got some people that are equipped to answer that here. Go ahead. Yeah, that's a good question. You know, there's some transportation data. You know, like the National Transportation Household Survey was the one that sounds like they're moving away from now, but that's been used in the past to study that. One of the interesting opportunities related to data in that area is if we have utility data, very granular data on where the grid has excess capacity, substations or transformers that have excess capacity, if you have that data readily available, it's really easy for developers to see that and understand that. They could then identify where you could put a charging station where you wouldn't have an upgrade cost, because the upgrade cost can be $50,000, $100,000 in some cases. So it can be pretty significant relative to the cost of equipment. There's two kind of contexts that I think are important. You'd have fast charging where you probably have a little more flexibility over where you put it. So that becomes a lot more important in a fast charging context, because it tends to be more of a destination for somebody, because they can park there for half an hour or hopefully less in the future and be able to charge. So they could use that data to site. When you're talking about smaller, like level two charging, you have less flexibility over where you can install it. You're probably trying to put it where somebody's already parking. So you have less flexibility then, so it's probably less relevant for that. I think we'd want to look at what kind of travel data is available, but also other aspects, parking, data, and utility data, I think would be really important. In the labs, there's a lot of work in agent-based modeling to do simulations that will help you figure out where to site properly. But again, you need to know what the travel demand is on the network, because these vehicles are engaging with all the other vehicles that are out there, and so they're in the congestion path, and you need to have an understanding of where these vehicles live, which sometimes you can get it, sometimes you can't. So it's a big challenge for the modelers. But, wait guys, with 200-mile range cars, which is going to be the baseline sort of table stakes for electric cars, you almost never charge in the wild. The need for randomly located public charging is minuscule. Most charging happens at work or at home, and we know in putting a charger in at home, you know, in homes as easy, and apartments as doable, in some areas it's hard. I mean, there's just no place to do it. But the electrification of parking lots at work is happening. Think about what the charging companies, like ChargePoint, have learned. When we first started selling chargers, it was a big process of getting the facilities manager who ran the parking lot to buy a charger, and of course the parking lot manager doesn't have any budget. He's not an important player in the corporate budgeting process. That has completely changed. When a company hires a new employee, they know, I've got to give them a new laptop, I need to give them a cell phone, I need to give them a charging station. It becomes an employee benefit part of the budget, and they know that for each hundred employees I buy, I hire, today I need three charging stations because that's how many electric... The penetration of electric cars is about that. As electric car penetration goes higher, that ratio changes, it gets cranked into the budget, employee benefits, knows, I'm going to hire more people, I'm going to charge more, I'm going to electrify more parking spaces, move on. It's part of the process. I think that the one thing I would add to that question is that it's important to know the difference between the last thousand electric vehicle customers and the next thousand electric vehicle customers. Those people might not be reflective of each other. As Mark said, we're not dealing with 80 or 100 mile EVs anymore. We're dealing with 200 mile, 300 mile EVs now, and that means a very different thing. The people that are willing to consider those vehicles are very different people. Yes, there will be a huge reliance on home charging. Most people don't like going to public charging stations or workplace charging. What we need to be looking at is the travel behavior of current gasoline vehicle owners, and to see who we can do targeted marketing to for the next thousand vehicles. Yes, sir. These are general questions about the whole idea of transportation in the next 25 years. We have currently artificial intelligence going at a really fast speed. Right now we have 3 million truck drivers that maybe within the next 10 to 20 years they won't have their jobs. We will have probably 2 million people, 3 million people work for Walmart, Costco, all those companies that they don't need to commute to work anymore. Because Amazon is disrupting them. So there's a huge amount of disruption going on in automation in the whole industry that people in general are sort of... The whole idea of transportation is questionable. Why do we need transportation, especially massive, like transporting to work or something because there's no jobs, right? Some of these companies will actually get disrupted in a way that... Especially these jobs that are so routine. So the idea of automation, what do you think of that and how do you think this will disrupt the whole... Are we actually solving problems that wouldn't exist, like traffic in the next 25 years? Okay, so we've got a dystopian view of the future. No jobs, no transportation. Anybody want to grab that one? Is that what we're looking at? We're dealing with a non-existent problem because nobody's going anywhere? Well, I'll take that. I don't think that's going to happen. I think automation is a really good thing because it takes a lot of the mundane activities out of our world. We're still human beings. We still have social lives. We're still going to want to go places where it's just going to make our lives a little better. And it's not going to displace all the truck drivers. We will still need to have people in the trucks for a variety of reasons for a very long time. And we are social animals. The one thing that I would like to see is this notion of hyper-local thinking, is that we need to start building our communities hyper-locally instead of stretching out into suburbia that I have to go from my downtown condo in San Francisco all the way across town to go to the veterinarian or to the dentist or things like that. I think there's going to be a lot of change in our urban environments as we go forward. But I don't think it's going to know jobs and everybody stays at home and gets fat. It's exactly what they were worried about at every innovation along the way, right? I mean, if you have a combine and a guy could farm a thousand acres by himself, that'll put all the farmers out of work. But if you have that, then you create a thousand jobs in food processing and food delivery and food serving. So the problem with this wave, I think, is that a lot of this actually before, let's say we had horse riders or like we had cars, all these kind of things that the jobs actually were actually the job that needs to do artificial intelligence, like one or two people can do huge amount of automation. Even coal industry, I read an article that in Australia, like 12 people manage like a whole, like everything is automated. Even the coal industry that we think that, so we have like a huge amount of automation I think we're going to have that. So I think that there are concerns about legitimate concerns, in fact, that have sparked some studies here at Stanford looking at the interaction of AI and employment. I think that is a legitimate concern. And I think that's the point which I would read to basically be a breakdown of, well, look, our electronic connectivity and transportation complements or substitutes. That's a simple question. And that's been addressed by, I think it's Ed Glazer's group at Harvard looking at whether those are complements and substitutes. And what they seem to be finding is that the more connected people are electronically, the more they want to travel anyway. And I don't think there's an immediate concern that people are going to stop driving around. So, thank you. Hi, my name is Rod Sinks. I'm an elected official from the city of Cupertino. I helped found Silicon Valley. A clean energy and we're now providing 100% carbon free electricity to 11 cities in Santa Clara County. And I serve on the airport. And so my question with you, with regard to transportation is, we now are finally beginning to see hydrogen vehicles. That hydrogen largely is coming from Bay Area refineries. So it's a cogeneration process. You could argue that that, although the owners are happy about their greenness, that fuel is not particularly green. My question to you is, do you have a model on which of these two types is going to succeed, the all electric or the fuel cell powered car? And as an official, that serves on the air district here in the Bay Area, where we've provided major incentive to create the EV charging infrastructure, we also be looking to fund hydrogen. And do you think perhaps, in moving from if hydrogen vehicles take off, do you see a feasible scenario in California for building a lot more solar and then effectively creating hydrogen from that as a storage means, when we're generating way too much solar then we can consume on the grid. And then, yeah, consuming that as, for transportation. So I'm interested in how you see the dynamic and what advice you'd give me as a public official and either promoting the hydrogen model or taking it down. I think that's an interesting point. The discussion of hydrogen has largely been replaced by the EV discussion. Does that mean the hydrogen cars are going away or are they just been delayed? I guess my first reaction to that would be it's hard for the government to predict the future of these technologies and in any technology space, it's very hard to predict what will happen. Well, we have money for grants and we could cite stations in our community. We ought to be doing that. Real decisions for us to be making. Yeah, it's like, you know, I don't work on the hydrogen side, so I can't offer you a perspective of where it's going along. And I just would caution about the idea of trying to predict exactly where the future would go. And what is more effective, what California has been successful and at the state level is trying to set the outcomes that it wants. So if they say, you know, the outcome that we want is reduced emissions by X percent. And then creating incentives that align with that that are technology neutral seems to be the best way. I don't know what that addresses. Your concern is more of a local concern. You're saying, look, I've got to make investment decisions now. You put several hundred million in grants to encourage hydrogen. Do we cite those stations or do we say, hey, no way, we like the pure EV model? Yeah, I've seen a large number of product plans over the years and those do include hydrogen. I would say most of the bias towards hydrogen tends to be on heavier, longer range vehicles. And so that to me doesn't lend itself immediately to citing stations in an urban area. Absolutely. I mean, it may make sense for fleets. Yeah. Because, you know, a $30 million hydrogen station can feed a fleet, but you don't want to place those all over the place to deal with cars. It's nuts. It's an inefficient carrier of energy, not a generator of energy. It's being propped up because government keeps throwing $30 million apop at these stations. Adam? I guess the other thing I'd add to this, which I think gets missed a lot in this conversation about the all-electric or the hydrogen, is I think there's an important role now, especially for plug-in hybrids to play, because you can get a certain segment of the population that maybe doesn't care as much about the environmental benefits and isn't willing to make some of the sacrifices that are required for an all-electric vehicle. They can buy the vehicle and they don't have to use the plug if they don't want to. They don't have to have everything set up to accommodate that. And what I think we would find then is they will use it. They will use it when they can charge, when they have opportunities to charge. They'll learn it without having to make a huge commitment in the beginning. And so I guess I'd like to see a little more focus on that because it's going to get a different segment of customers to adopt the vehicles. I totally agree with you. I think the consumer studies are all right down that path because EVs have been flat for years and years and years and that's because they've been sold to everybody who is inclined to buy an EV. And so we need to end the Tesla jumped out by appealing in a totally different segment. And so the changes in the vehicle, the way people make decisions about vehicles, they went to a different consumer segment. I think the plug-in hybrids to me remind me of that Mitch Hedberg joke that an escalator can never go out of order, it just becomes stairs, right? So a plug-in hybrid can never go out of order, it just becomes a regular hybrid. So I think there's a lot of benefit there. But the good news is this year over last year, electric car sales are up 44% this year so far. And it's because we didn't make the right cars for the first two decades of electric cars. We made 70-mile goofy looking things nobody wanted instead of 200-mile things that are really better cars. I think we have time for maybe one more question. Sure, mine's pretty quick. I'm from AutoTech Ventures. I'm interested in car data today. It seems the most valuable data is behind the OEM firewall, everything else you can sort of collect with your phone. What's it going to take for OEMs to open that data to third parties, whether it be startups or governments? Well, I got to tell you, even as an academic, it's been extremely hard to get access to data from OEMs. That's something I've been trying to do for years now. So I don't have any great news for you if you're an investor. Maybe give them some money. Although I would say that data is only worth something if they use it. And many of the OEMs have not figured out how to use that data to their best advantage. So if you can bring to the table something that they don't know how to do yet, that's probably worth a shot. It's actually getting worse. I mean, BMW just announced that the OBD2 port would only be functional when it's static and a service garage. So that the service technician can use data off the port. But if the car is moving, then the port is shut off. So that cuts out all the people who made the OBD2 data things, the dongles, insurance companies, all the benefits that consumers were getting because they had hacked the OBD2 port, they're now going to invent you from getting anything. You've got a gentleman here from the ARB that you could ask to change that regulation for you. I mean, I was at BMW trying to push for openness on that stuff and obviously didn't get anywhere. All right, Adam, you want to respond? Yeah, I can't respond generally to that. I'd be happy to talk to you offline a little bit more and see if there's opportunities. I guess I'd go back to the idea that there is a customer privacy issue that becomes really important. There is also cybersecurity issue, which I mentioned before, which is also an important issue. So we've got to manage those in terms of dealing with this. All right, Jane, you want the last word? Yeah. Someone's got to figure out how to, they are not going to give the data out because there's also a lot of information about the design of the vehicle in that data if you really worked hard at it. That's what they say anyway. So we are going to have to figure out how to partition this data so that we can allow levels of access to it. And nobody's doing any of that work. You can figure out the privacy issue. You can do K-means privacy stuff. You can do all kinds of things, but nobody's doing that yet. And so as I said earlier, we're going to have to work through it because it will benefit people. It will benefit transportation. It will benefit cities. It will benefit a lot of people if we can get to it. So we just need to work our way through it and find a way. All right, so we're out of time. Please continue this after the fact. And please join me in thanking our panel here today.