 Welcome to the CEO Telecom Exchange Round Tables, both for our guests here at Telecom Exchange LA. Welcome, as well as for our viewers joining us as we stream on Facebook and on JSA TV. Our third panel today is on the very hot topic of smart machines. We're talking connected cars, robots, and the necessary infrastructure. I'm very honored to introduce Mr. Jeremy Kaplan. He's the Editor-in-Chief of Digital Trends, or DT. If you're not familiar, it's a leading consumer publication that reaches about 25 million readers a month, so many of us in this room, I'm sure. As Editor-in-Chief, Jeremy oversees a staff of 35 editors, video producers, and content experts, as well as dozens of contributors providing up to 100 pieces of news, analysis, and product reviews each day that spans all areas of tech. We're talking, of course, including smart machines. So quite a lot of writing and reading, I'm sure. So thank you, Jeremy, for taking time out of your day and for being here joining us. And thank you, everyone, for listening. Jeremy? Jamie said she's going to embarrass me. It's the mister that embarrasses me more than anything else. So yeah, Digital Trends, and for those of you who don't know, we're a technology news and review site, 25 million to 30 million users per month, and we focus on technology for the way you live, which increasingly is growing more complex as everything gets connected. And that's not just your smartphone, your smartwatch, but your stove and your fridge and whatnot. And I think that the challenge that we're all facing, we've all been seeing in the industry, is that increasingly, while these things connect, they also don't talk to each other very effectively. So I have a smartwatch, and it doesn't talk to my smartphone, which is like the most basic thing, and yet it's the truth. And last year, I bought a dishwasher, which is really neat. It's got this companion app, and you hold the app up, and it plays a little sound, and it goes online and diagnosis problems with the dishwasher, and it doesn't talk to anything else. I bought a new fridge. It doesn't talk to it. I bought a new stove. I bought all new appliances. They don't talk to each other. And I, of all people, I would think would have the things that speak to each other correctly. No, this is increasingly a problem. So we have here a panel of experts to talk about some of the infrastructure challenges we're facing. And it's a neat gamut that they run the gamut between the people that are building the infrastructure to the people that are building the protocols that run on the infrastructure to people that are using the infrastructure to people that can talk about what it all means for us. So I tend to introduce people with the wrong names. I call everyone Billy Bob when that's not really their name. Maybe you guys could each introduce yourselves and tell us just really briefly what your company does and how it relates to this space. I guess I'll start. I'm Doug Junkins. I'm from NTT. Specifically, part of NTT is applied R&D organization in Silicon Valley that's looking at technology that's emerging in the market and how NTT is an infrastructure provider can leverage that technology to improve the services and product offerings to our enterprise customers. Now, hi, my name is Arthur Lizinski. I'm the CEO of Umnitza. Umnitza is an IoT platform. We originally started tracking things that are more traditional, such things as mobile devices, servers, switches, and are now tracking a number of IoT devices. And maybe today we can talk about the similarities and why that's possible. Hey guys, my name is Elter. I am the CEO of Skirt and at Skirt we are trying to make mobility more accessible to everyone, which we do this by trying to increase the utilization rate of a car. So we are a mobility company. Right now, we let you book a car from a mobile application and we deliver the car to you. And on the background, we are building the infrastructure to move these cars around the city being able to deliver them, so like the physical infrastructure, as well as being able to do all these calculations to increase the average utilization rate of every car in our platform and basically trying to make one car more accessible by multiple people throughout the day or throughout the month. I'm Frank Toby. I actually have two hats. One is that I have a website called the Robot Report, which tracks the business of robotics. And the other is that I'm a co-founder of a company called Robo Global that produces an index of robotic stocks that has been licensed by a variety of exchange-traded funds that operate around the world so that people can invest in the robotic schema. Well, thanks for joining us, guys. So let's start off by talking about a topic that some of our earlier panels touched upon. Am I too loud? I get loud and I get passionate when I talk. Just wave at me if I'm talking too loud. Anyway, Jamie mentioned in her initial opening remarks the concept of smart networking. So we have smartphones, we have smart watches and smart cars and smart homes and we have smart lives. The smart, the network itself needs to evolve to deal with this. So Doug, maybe I'll throw this at you first. To what extent has the network evolved to date and to what extent do you think it will need to change to deal with the growing number of things that are online? I think that the network, I've been building the internet myself since 1995 with the one ISP or another growing. The network has always been evolving to find new ways to deliver services to customers to meet new demands. I think with the technology that's being developed now around software-defined networking and software-defined wide area networking, the capability for the network to adapt to changing workloads on its own in an autonomous fashion is actually very close to being a reality for us. Some people are already starting to look at network traffic patterns based on evolving use cases and being able to adapt their network topology or architecture in order to be able to better serve that particular application. I think as we continue to evolve to this world where every device is connected, there are gonna be patterns that start to emerge in terms of what's talking to each other. You spoke about appliances, maybe there needs to be a network of dishwashers that are communicating use cases or failure patterns in a way that's different than how your refrigerators talk. I don't wanna start to build silos or walls between the appliances. There's one, all appliances matter. So, but in any case, the network is gonna need to learn those patterns through observation and then adapt to meet those patterns. And we're starting to see the technology to do that now when we're able to set up ad hoc networks between particular devices based on their communication patterns. Yeah, I think maybe just follow up on that. One of the things that we're seeing is I think that a lot of that infrastructure technology is following the actual use cases we're trying to accomplish. So I think as we're trying to accomplish more use cases that infrastructure is going there. Trying to put the cart before the horse and coming up with all the ways we can improve infrastructure I think is good, but at the end of the day we're trying to solve a problem. And if we're really trying to solve that problem, we need the right infrastructure. And so I think the infrastructure players are catering to the problems we're trying to solve. I think infrastructure and I think wires. But do we have the right type of infrastructure present? Is there change to the protocols that we need or is everything just gonna work over wifi or over ethernet or whatever we happen to have? I think the world we live in today is everything is IP. The transport of how that IP is connected shouldn't matter. What we're seeing now is the way that we're adapting networks is we're building overlay networks on top of an IP infrastructure so that transport of the underlying IP can be ubiquitous across whatever transport technology you need in order to connect to a device. The networking we're talking about that's going to be adaptable on top of that is being built as an overlay that's abstracted from that transport. Yeah, I think there's a lot more work to be done. I think especially when we're talking about smart manufacturing, we're talking about stadiums, we're talking about a lot of capacity, a lot of data. I know VRs, videos taking into consideration, all this other stuff. Just based on our past trends, I think we're gonna solve that problem. And I think the more demand there is, the easier over time we're gonna solve that problem because there's a lot of money to be made in solving that issue for all the other things that are built on top and then provide all that value to the end customer. That very, very low layer has to be there and I think we'll continue to be there. Frank, we were talking about this before and Doug mentioned all appliances matter. But communication between different devices is one of the real challenges. Can you share that, Annick told me we were discussing before? FANIC is a large robot manufacturer. They have 400,000 robots around the world that are at work. They did a study with IBM to see how they could maintain their equipment better. And they found that if they took the streaming data from every robot and just analyzed it compared to all the others, they could find patterns of when the equipment was or screws or washers or devices were gonna break down. And they could then schedule a maintenance application. The problem is that those 400,000 robots which have perhaps another 10 sensors streaming so four million streaming devices are in perhaps 4,000 companies each with their own security system. And they don't want to give their data or allow their data to be analyzed even though it's beneficial for everybody involved. Not only that, they don't exactly speak the same computer language. FANIC sends all their data to Rockwell. Rockwell has a proprietary system that does not speak to anybody else. And if you have a company that has a Rockwell system and you want to send that data somewhere, who's to say it's secure or not secure? And that's the problem that they're having. Here they found that they could improve efficiency, productivity and runtime to the IBM standard of 99.99 and they can't do it because they can't get permissions from all the different players that are involved. And I think like when we're talking about IoT and we're talking about industrial internet and all this stuff, what's really happening is these manufacturers, the Rockwells of the world, KUKA's of the world, they're thinking in one way which is proprietary, this belongs to us, we built it, our standards, we integrate with our own systems, we don't integrate with other people because security, because product quality, all this stuff. And then you've got the entire other side which is software guys. And software guys are like, hey, no, go ahead and use it, open APIs, documentation, everybody work together, I mean that's where the open source came from, that's where all our software's in. And what's happening is these manufacturers are now getting pressed to say, hey, you need to have the software conveyed ability. And the software guys are going, wait a minute, what do you mean I can't install Windows 10 on everything? And so now what's happening is these standards that we used to have, and these new guys, they're coming together and two worlds are colliding. And so you have people who are in the manufacturing business going, why in the world would I open my APIs? And you got guys on the software area going, if you don't open these APIs, you will die. And so now I think what's happening is we're gonna find a common middle ground and we're just in the beginning of that. And I think that's why this space is so interesting. I think it's partially because doing these hardware things are becoming easier now, like manufacturing things, I can design something on my own, send it to someone in China and they start manufacturing this unit for me. So as software people, oftentimes recently, I wish I find myself trying to design some hardware unit and it's much easier to manufacture now. So I feel like these big guys are becoming more kind of influenced or forced to switch to more open model. I think it's like a trend in a lot of industries that over time, they have to open up and I think they're slowly realizing it. But yeah, I think though, I mean, the reality is these large established companies, giants or dinosaurs, maybe you might wanna call them, they're not gonna move overnight to being open. It's just not the way that their program, it's not the way that they've evolved to be successful. So what we can do as infrastructure providers is start to give them a path to how to open up by addressing some of their concerns, giving them a network, for example, where they can connect their devices outside of a factory manufacturing network, a closed network, so that we can actually start to collect those devices in a way that are secure, but they don't have to be completely open. We can create an overlay network that only ties their devices together into one of their analysis systems that they might provide as a service without, again, trying to make them completely pull back the bedsheets and expose everything that they have developed or have available. And they won't do that anyway. I think it's about coming to that middle ground. I think you mentioned Rockwell, and I think you think of all these like Johnson Controls, you mentioned them earlier, Honeywell, you think of all these companies that have been around for so long and everything is built around proprietary Siemens. And so I think eventually, they're gonna have to find that middle ground. I think companies like NTT and a lot of other companies are helping that happen. But I think there's gonna be a completely new type of person that we haven't seen before, which is somebody that really understands manufacturing, which in its own right is so many layers, and then somebody who also understands software and technology and combining the two. And I think that's when we're gonna see manufacturing automation to the next level. But it's so early. Talk to a manufacturing guy about technology and talk to a tech guy about manufacturing. I have no idea. And finding the guys that live in that middle, there's probably a handful at most, and that's gonna change. So Arthur, we're talking about this earlier. I was making the example of my dishwasher, which doesn't really talk to anything. And you made a great point that there are companies that are adding networking and ethernet ports onto old devices, and those aren't really thought from the ground up as being smart devices. And then you have something like an Amazon, Alexa, the Google Home devices. These things plan from the ground up for interoperability and for networking, which really transforms the entire space, right? Yeah, I think the retrofitted devices, the refrigerators, the robots, all these things we're trying to make smart have these limitations. And these new things, like you mentioned, like the Amazon Echo or these other things, just the way we interact with them is completely different. I don't need to use my phone to interact with the Amazon Echo, right? Like, I've just talked to it. So imagine now when this microphone becomes smarter, the building can become smart, I can interface with that building through whatever's already built into it. I can't go and talk to my refrigerator, right? Like, that seems crazy right now. But when the refrigerator is built, ground up with a microphone, it has a chipset and all these other stuff that it needs, the very first thing I'm gonna do is go to it and say, hey, order more eggs. And it's gonna go order more eggs. But we can't do that with the retrofitted stuff, obviously, because of the limitations that they provide. Why would your refrigerator not know to order more eggs when you're low on eggs, though? It should, if it has some kind of weight or it's doing some kind of visual stuff. Our co-founder, Trent Seed and I, we like to talk about all these other stuff we could do as if we weren't fully on one company. So we like to pretend we have free time. And one of the things we like to pretend is what it looked like to have the best connected fridge. Trent, who's just a genius, came up with this idea of having weight. What if you put your eggs into the same spot every day and then over weight, you can tell? Turns out, eggs weigh different, it's just a little more complicated than we originally thought. But yes, that's what's gonna happen. Automatic supply chain, reordering things based on quantity. In your own home, it's definitely gonna happen. So the concept of sharing information is vital in a lot of different ways. And it means different things for business versus consumer appliances. So for example, the concept of my furnace talking back to the network and saying, I need maintenance, that's fantastic, who wouldn't like that? But it does raise some privacy concerns, doesn't it? And as you talk about devices like the Echo, your more consumery devices where you have more personal information, I think you're more likely to wade into the privacy debate. Ilter, what sort of information does your company send back to the network? How do you monitor privacy like that? How do you worry about that? I mean, one thing we're looking at is these like, getting like, for example, diagnostics data from cars. It's interesting case with cars, I think, because they have these things called like OVD boards, which are actually standardized protocols. And they're fairly open right now. I was just thinking about why, and I realized that because these devices needs more maintenance and like, for example, like BMW can't have mechanics shops around the city. So they had to, I guess, expose some of these things already. So one thing we do is like we can use this open protocol to get all the diagnostics data from the car, their fuel level, if it needs oil change and all this stuff and we can store it in our own database and that way, we know that when a car needs maintenance, we work with other partners to get these cars actually. So we can't tell our partners that these cars need maintenance or service checkup, all these things. So it's interesting. The OBD2 connector is a great example of taking something that's old school and giving it this mass kind of appeal. Since 2004, I believe you probably know this better, it's required that all cars after 2000 have OBD2 connector and you can go and buy an OBD2 connector, plug it into your car and get data under your mobile devices, most likely everybody knows. But what's most interesting is all these cameras, all this OBD2 stuff started going into other stuff, like your tractors, like your other free, like I think coffee machines run on a can bus, right? So like all this stuff is becoming standardized. How do we get more of that to happen in other industries outside of the car industry where yes, you're right, uptime and efficiency and diagnostics is so important? I won't need another panel for that, that's a good question. You know, I think that's a big problem to solve when you're trying to figure out how to do the networking within these smart devices and then what the interface is. Where is that boundary between the network that's running the device, the can bus in a car, for example, and the external network, how you're connecting to that device. And can bus, I mean, obviously brings up all kinds of security concerns because it's a completely unsecure device in your network or in your automobile. And there have been remote hacks shown where if you can get access to that bus through, I think the easiest one that people have shown through entertainment systems that are connected, you can actually disrupt the whole operation of the car. So figuring out where that boundary from the external network to the internal network of these devices are and how to secure that boundary is something that needs to be solved before you're gonna start to see, you know, any sort of ubiquitous standard for those internal devices. So. Now's a good time to tell the story of George Hotz. George Hotz is a hacker and he was the first guy to hack an Apple iPhone. But in the self-driving arena, he wrote 2,000 lines of code that he put into along with physical hacks and he put into Uber cars in San Francisco. 10 Uber cars. And all of them had radar and a few other cameras. He then took the data from three months of 10 cars operation and ran it through his 2,000 lines of code, which wrote on its own new code that drove a car. He took a Bloomberg reporter out with him and they drove in the streets of San Francisco and on the freeway and at one point the car swerved out of the way of a bicyclist and the Bloomberg reporter said, well, you programmed for that, right? Says there's not a word of code that I programmed in this. It's coded by itself. There's 3,000 people doing the same task at Google. So the hacking aspect of this and also the evolutionary aspect of this and the speed and the exponentiality is something that I don't have a clue where, how fast it's going and where it's going, but I can see it happening. Well, we have a couple of people on this panel that have some direct insight. Gentlemen, how soon are we gonna get the self-driving cars? Joe Hart is actually my friend. We went to the same college. I think in six, seven years it's realistically gonna be everywhere in California. Not obviously the whole world, but I actually think in six years, a whole California, actually all cars will be autonomous, especially around the areas where we live right now. And the interesting thing with this is there's no transition between a normal car and autonomous cars. At some point, I think all the cars needs to be autonomous because mixing them is even actually the most more challenging than building these autonomous cars. So I'm guessing there will be a short phase where there will be one lane dedicated to autonomous cars, but eventually, I think all the five roads will be dedicated to autonomous cars. If you wanna drive yourself, if you're driving, you just go to a racing track or something and drive your car. But all the cars, I think in a fairly short future, it will be autonomous. I think specifically with this election and all this, I see how quickly things can change. And I think things can change really quickly, even things that have so much legislation like our road. So I think it's completely realistic. Back to the security thing for just one second. I think what's so interesting about security, and especially with all these connected devices that people are missing from a tech side, is that if you think about end points, as they call it, these end points that connect to the internet, whether they're your mobile device, whether they're Visa, POS systems at Target, whatever they are, they all have the same issue. They're all connected to the internet. So if we think that we're gonna solve that problem, or we already have solved that problem, it's incorrect. Ask anybody that works in IT at any corporation in the world and ask them if we've solved the endpoint security problem. The problem is gonna magnify as we have more cars and it'll be a cat and mouse race. There will always be dangers of this stuff. You can bus the robots, the cars, and your mobile device. They're always vulnerable to these hacks. And I think the question that we need to ask ourselves is what is the right infrastructure for that? But I think our fear and those concerns should not hinder innovation. And I think that's really important is I think we need to continue moving forward with these end points and find better ways to secure them. Yeah, and the idea of having all of these end points that you have to secure and manage and everything, that brings into the evolution of the management systems for managing end devices. Today, there may be some MDM type solutions that you use on your phones and your laptops, but it's not evolved to the point where we can actually manage what software version, what patch level, what security parameters are set on millions and billions of devices that are being connected to the internet. So one of the things that really needs to be worked on is how it is that you understand what assets there are out there in the network, how they're being managed, what versions there are for it and how they're being secured. Until we can centralize that, we're gonna live in a world where these sort of hacks are, I didn't mean for this to turn into a security panel, but the reality is that's where we need to be in order to be able to make the next step to this world of devices that are connected and communicating themselves. And just to add one thing, I just wanna drive this point home, it's not actually a technology issue. It's not actually something that is a technology is gonna solve, it's a process human issue. So when we think about managing end points today, the reason that we have such security discrepancies because we can't answer simple questions like where is the device? How crazy is that? That the IT teams all across the world don't know where their laptop is. I'm sure that anybody here has worked at a big company, walked away with a laptop. Guarantee it or something that the, and that's the security risk. You can upgrade all your firewalls, all you want. You can invest into all the security software in the world. If you don't have hygiene inside your company or even at home, based on how you manage your things end to end, no software or tool or agent you can install on the device is gonna do you any good. It's about hygiene and it's about process and that's the most unsexy thing about IoT, but it's so true and it's really important. But the challenge is that consumers don't want to give any bits of data up and we need to pump out an incredible amount of data. Consumers are worried about the Alexa phoning back that the lights are out because then therefore I'm asleep and somebody might come in and grab me in my bed or something, but meanwhile we need to know detailed information about how long the lights were on and which switch was on and uptime and all that stuff. That's the security issue, that's the real risk. Yeah. But anyway. We're in violent agreement. Another topic I wanted to bring up here because we have very important stuff. Speaking about Alexa, which is a fascinating way to control your devices, we have this world of bots coming out these days that are doing, that are controlling our devices and are chattering back and forth with each other and creating this vast amount of background network traffic. Is this a security concern? Is this something we should worry about? And from an infrastructure perspective, is there anything we need to do to support this, to think about data prioritization? Sure, I'll start again. A big chunk to throw at you. I think that from a security perspective, I think that it is data that's gonna have to be made visible if we want to be able to really adapt to this world of connected devices. So I think we've pretty well beat that home. In terms of looking at the network infrastructure as it exists today, I think the way that the internet has evolved over the last, let's say, 10 years has been this massive consolidation of resources into large data centers where everything is getting centralized into extremely large pools of compute resources. As I think it was Mark on the first panel mentioned, that world doesn't really adapt well to a IoT world of connected devices where we need to actually start to learn how to distribute those compute resources out as close to those devices as possible so we can collect and consolidate and correlate data near the edges or else it will inundate the network as well as making sure that we have the ability to have very low latency access to some of those compute resources. So when you don't have enough compute power on your mobile device or in your refrigerator to determine what the pattern for ordering eggs is, today's Tuesday so they're not gonna have eggs tomorrow and we wanna wait until Wednesday to order them, for example, those sorts of learning algorithms and artificial intelligence, the refrigerator may not be capable of making those decisions itself. Maybe latency for that particular application doesn't matter and you could do it in data center in Singapore if that's where your compute resources are. But there are gonna be applications where you need to have an instantaneous sort of response to a question from a device and being able to distribute those algorithms out towards the edges is gonna become very important, especially for use cases that we aren't even necessarily dreaming of yet. So I think that's the important shift we're gonna see from an infrastructure perspective and a network perspective is the distribution of all of this compute power that we've now spent 10 years consolidating. Yeah, it's a concept of load balancing, it seems like, you know, and where do you balance the load? I have a buddy of mine who runs a speech recognition like a Siri competitor and what they do is they put it into little toys for kids and you can interact with the toys and of course the toys aren't gonna be connected to the internet because they have to work whenever and so this interaction has to happen on the chipset on the device and their biggest cost is that chip that they have to run on the device and when we're talking about edge computing, it's basically, you know, are we running that information on the device and how much of it are we running and I wonder how much of that question is the application provider versus the actual infrastructure and does the application provider have the necessary tools to provide the best end service to the end customer? Yeah, I think privacy is another issue that is actually very related and it's gonna be solved with these, like having these computations happening on the client side. Like Apple started to do this thing where like they can actually, they have this thing called differential learning so like they can do a lot of the AI or machine learning stuff on the phone or even if they need to go back to the data center, they kind of like anonymize your data and send it and then it comes back. So like, for example, in my case, I'm fairly open-minded about all these technologies and stuff but like having Alexa at my house still makes me uncomfortable. It's like some device listening to my conversations and it goes back to Amazon's data centers but if I knew that the device was actually just keeping all those like listening the sounds it collects in the device in the client, I would be a lot more comfortable. I think that's something that is going to be kind of required for privacy as well because otherwise, like it's gonna get really uncomfortable for people. What we've got, did you wanna jump in? Well, I was just gonna say that once 5G comes along and can really speed up the latency factor, we hope but that can change a lot. Because right now, like your toy example, the chips provide as much as you can possibly get locally but when you have a mobile robot and you're carrying power around, you're limited really on what you can do and right now, going to the cloud is totally inefficient. Totally inefficient. And I think the drone example, the military drone example, those drones fly by themselves except for the first two minutes of takeoff and the last two minutes of landing. That's done by on the ground remote control. We've got a matter of minutes left. If anybody has any questions for our panelists, this would be a good time to throw it out there. Applications or potential implications with the FTC and the FCC, are you guys concerned about that at all? We're not because we're a B2B business and as far as we're concerned, if the customer's willing to give us money for it, we're willing to give them our product. I wonder how much that legislator affects you guys specifically. Are you guys affected by any of that legislature? Not really. We're the same. And this is labeled as a telecom exchange so that's why I'm asking for a telecom perspective. Right. Maybe not necessarily from the data perspective. Yeah, so as the telecom representative up here on this panel, obviously NTT looks very closely at what the FTC and FCC are doing as well as other legislative bodies around the world to make sure that the services that we're offering fit into whatever that those regulations are. And even to the point of having people in Washington to help work with the FCC and other governmental agencies to make sure that regulations that are coming out are technically feasible as well as commercially feasible. So I don't think that anything that is taking place in terms of the evolution of the network is outside of what the FCC has purview over but I don't think anything is massively different than the way that we've handled the growth of the internet for the last 20 years either. Did that answer your question? My rather. No. I don't have a better answer for you then. That was very fair. Any other questions for our panelists? I'll make a comment rather than a question. I serve on Beverly Hills has decided that autonomous vehicles are imminent, are disruptive, are going to change all of our land use and our planning. And so we've been studying them for a year. The comments about how imminent it's going to be driven by safety. Okay, now for the first time in many years the fatalities on U.S. roads are increasing but where you need to pay attention to is Southeast Asia. I think half a million people die a year on roads in China. And so this is a technology issue. So autonomous vehicles will exist in Southeast Asia. They will exist in China and it's the same cars, the same connectors. And you mentioned George Hotza, for comma AI for 200 bucks, you were very close to being able to put a device into your Audi that was better than a Tesla. It was, you know, I mean he took it off the market and it's a went after him a little bit. But you know, autonomous vehicles will be here, they'll be driven by safety and you'll be able to put them in your cars. What I disagree with you a little bit is there are so many cars around. I don't know how we deal with the interaction of autonomous vehicles and human driven vehicles on the road and it's a real dilemma. I think what you'll find in Southern California is that the HOV lanes will be AB only. Okay, and interesting things happen. I mean, I've run through numbers, I get 20X capacity if I put an autonomous vehicle lane only 20 times capacity. Others will argue for 10, but all of a sudden you've increased the capacity of an infrastructure we already have. So ABs might only be on highways, okay, they might be replacing chains, but I agree with your analysis. I think they're gonna be much more imminent than anybody thinks. That's great news for LA. It is. And you know, the safety issue is an interesting one. You know, the federal standard is if it's twice as safe, we'll look at it. Okay, so if we're only killing 15,000 people here, that's 15,000 lives, say it. How do we deal with the one or two because everybody will go robot, just kill somebody, right? You know, as soon as something goes wrong. You know, the dilemma here might slow things down. I would pay attention to China in particular. We've interviewed just about every major automobile manufacturer we can. We're singularly impressed with Volvo and Volvo is a Chinese company. Pay attention to China. Just an opinion. I'll give you the opposite opinion. Yep. You have the ability to have self-driving aeropines for what, 25 years? Self-driving trains for what, 35 years? You still don't have it? So you don't have the safety factor to driving it. It's the safety factor that is going to drive. I think it's gonna drive autonomous vehicles. It only takes one train accident to get every presentation. Well, I think if you go to Palo Alto, California, or Mountain View, California, you're gonna see just as many self-driving cars as you're gonna see normal cars. And I think that's maybe just to that region, but those aren't going anywhere. Well, can I also point out that the car industry, is the industry that successfully killed electric cars the first time around, wiped out trans and public transport in California also? This is one hell of an industry. And you're thinking five years, you're gonna be able to, no. This industry needs a lot. I mean, you're also not looking at quite very, very hard. Commercially, flying a plane, having a pilot there doesn't really cost that much. The flight takes 12 hours. You pay the pilot for a day. It doesn't really, also commercially, there's not business incentive to replace that pilot with someone, but there are billions of cars right now. And actually, the times that people will say by not having those drivers is huge, compared to times when people say it by not having pilots. Take your analogy, every autonomous vehicle has to have a four-wheel drive. No, it won't have to. No, no. Not the level fives. Yeah. I think it's a matter of discussion that in five years, we'll just be a new discussion. So I think we'll wait and see. I, for one, will buy a self-driving car as soon as it's available, because I have work to do and I can't be driving. If I just had a comment, there's a reason, at least for my city, the reason we don't care whether it's five years or 20 years, is because we go look at the impact of autonomous vehicles. So let's say it's 20 years. If anybody thinks it's not gonna be in 20 years, it's gonna be completely wrong. I think it's the time frame. We build buildings for 100 years. We build streets for 100 years. So our urban and land-use planning has to change now. Parking for any significant-sized residential building is 40% of the cost of the building, okay? If you're gonna remove 40% of the cost of the building, you don't need to park cars anymore. That's a whole change, a whole change in land-use. And so, yes, we need to worry about the time frames, but we need to do our planning now. And it's a very, very unique approach. And your business, every single manufacturer we've talked to does not believe they can sell cars for very long. Yeah, the ownership will die too, like, oh, I think there will be fleets of cars in the future owned by these OEMs themselves. Cars spend 85 or 90% of the time in the garage. Somebody's gonna come along and as you get home, they're gonna say, if you give me your car, I'll pay 50% of your car. That's his business, yeah. And as you get to work, they're gonna do it. So you're not gonna buy cars anymore. And on that note, we have to wrap up here. How about a round of applause for my panelists?