(click) >> Hey, welcome back everybody. Jeff Frick here with theCUBE. It's 2018. We just got out of the CES show and all the rage is autonomous vehicles. You can't get away from it. It's what everybody's talking about. Tesla just announced their autonomous truck, their autonomous Roadster. We're here in Palo Alto, right on San Antonio Road. Googleplex and Waymo's are right up the street. So everyone is all about autonomous vehicles, but we're excited to be here at Phantom Auto and they're taking a slightly different approach for a slightly different problem. We're excited to have Shai Magzimof. He's the co-founder and CEO of Phantom Auto. Shai, great to see you. >> Nice talking to you, yeah. Thanks for having me. >> So Phantom Auto, you guys just got back from CES. You were giving demos, but you weren't stuck in, like, the little lane that was protected. You were actually driving people all over the streets. >> We were driving on the Strip, yeah, yeah. We actually were picking people from the hotel lobby, so the valet guys would let us in with an empty vehicle. These videos are actually also online, and we drove them off the Strip and back to the hotel, or to another destination. >> So you're doing a whole different thing. You do not have an autonomous vehicle. >> It's not an autonomous vehicle. >> You were the ultimate chauffeur driven vehicle. >> Right. Right. So again, for the show, we did our job to show that the vehicle can drive without a driver in the driver's seat, but what we do is actually a safety solution for autonomous vehicles. And that safety is basically what happens if an autonomous vehicle artificial intelligence doesn't work. Let's say there's something that it cannot see, or something that, you know, an unidentified object, road construction areas, severe weather conditions, all this stuff happens all the time. And autonomous vehicles may struggle with the situation so Phantom Auto provides a solution that we work with these companies. We provide them that solution that allows remote operations, so someone will connect remotely. >> So let's back up a couple steps. Autonomous vehicles are meant for no driver. You guys have a driver but you're really assisted driving with a person from a remote location. So how do you describe that in a short category? I'm sure the analysts will want you to have a category. >> The category would be the same way you think about air traffic control, right, or any type of control center, like call control centers. Any type of support for customers, you would have a bunch of people sitting in front of computers, in our case they're sitting at computers with steering wheels, we'll see that later, and they can connect to a vehicle remotely, and when they move the steering wheel or press the gas or brake, it would actually happen in realtime. So we have this software that allows this realtime, critical communication for autonomous vehicles. >> Now what's weird is when we first heard about you guys, I'm thinking, okay what is the use case? Am I going to send the Phantom Auto to go pick up my hundred-year old grandfather who probably shouldn't be driving anymore, where you're escorting it. But really it's a very different application, and I don't think most people understand that, in autonomous vehicles, there's a whole lot of use cases still that they haven't quite figured out. My favorite one is when two of them pull up to a four-way stop, and neither of them wants to go first. They get stuck in a friendly lock, right, they get paper-logger, some poor kid has his foot in the intersection and is trying to wave the car through and it won't go through. So it's corner cases that you guys are all about, to really enable that next-stage of machinery. >> When I started a company, right, I'm a big believer in autonomous vehicle, I wanted to make them happen faster and sooner because it's life-saving technology. This is going to change the world. We all want it faster. Now, the reason why we're still not there yet is because there are many corner cases, edge cases, these situations where the machine didn't train enough for, and in this situation they provide a cover. So we have a person that would sit in an office, he doesn't have to be so close nearby. When we were in Vegas a couple weeks ago, the driver was in Mountain View, so Mountain View, California, Silicon Valley to Vegas, and he moves the steering wheel and he moves it real time. >> But he's driving the car. >> Yeah. >> So one of the great knocks on cloud, right, is latency, and clearly the use case that's always brought up is if you're in a self-driving car, you don't have time for the data to get it to the cloud and back to make a decision if a little ball rolls out into the street. So latency is a big issue. How do you guys deal with the latency issue? >> That's our secret sauce, obviously, but I'm happy to share as much as I can. The high level description would be, we connect multiple networks at the same time. We would usually have only AT&T in your cellphone, right, or in your car, and then we have AT&T, Verizon, T-Mobile, and a few networks, all of these together are bonded, and once they're bonded they get a much stronger connection. It sounds maybe easy, okay so let's plug a few phones and then get a really good connection, but it's much more complicated than that. We share and split the data across multiple networks at the same time, we prioritize the data. So, like a brake, it's very important, right, so if the remote operator is pressing the brake, you want it to be first in the vehicle, where the right side of the camera is not as critical, so lower latency for the brake, and then a little bit higher latency for something less important. >> So you've got dynamic, kind of, latent distribution. >> It's all dynamic, realtime, you know, so that's what we do, our real core. We provide this communication, real time, critical layer of communication for the video streaming and back of the data from the remote operator, back and forth all the time. >> So that's one big piece of it. Another big piece of it is the communications between the occupants in the vehicle and the driver. Another really important piece that obviously most people aren't thinking about for autonomous vehicles because they don't have that use case. But that's a pretty important piece of your solution. >> Yeah, that's a big one. I'd say that for this, you don't need to do a lot of innovation. It could be a simple call with the driver remotely. But, we're all about safety, right, and we're all about giving passengers this psychological trust, and it is true, you want to sit in a car that drives 100% of the time. If I tell you that your car today would go in and drives only 95% of the time, you would not buy this car. Same thing with autonomous vehicles. So we provide a safety and service layer. On the safety side, it's about assisting the vehicle when there's an emergency. It could be post-emergency or before it happens. Let's say you're just stuck in the middle of the lane and you don't know what happens. Even if the driver remotely wouldn't actually drive the car, you still want to be able to talk to somebody, right. So, I'd start with first the person, the driver, the human being would greet you when you enter the vehicle. It's an autonomous vehicle, he would say hello, how are you, nice to meet you, my name is let's say Ben- >> Ben is going to be your driver. >> Your driver soon, and Ben is going to tell you that whenever you have a problem, if you need any assistance, he would be there for you. That already gives you like a whole different type of experience, and when you leave the vehicle too, he's not going to be there all the time engaged with the car. The car is going to drive on an autonomous AV system, but at least he's there in case you need him. >> And again, the attention thing, which is an issue, you see with some of the test autonomous cars out there we were talking before we turned the cameras on, where the engineer's got his hands ready to grab the wheel if there's an emergency. That's not really Ben's role here. The car is going to take evasive action in terms of emergency. It's more to get out of like these weird corner cases as you said. >> Correct, it's not a test driver. Today, most autonomous vehicle companies still require and mandate it, it's actually illegal. By the regs, you have to have a person in the car. We also have a person in the car, and we do that same thing, although when Ben is driving, he's not replacing that person. He's just assisting when the autonomous vehicle system would have an issue. >> Right. So the next thing I think that's pretty interesting about your company, as you said, you're a software company. There is hardware components, you can see the back of the car, we'll take some film of the driving station, but you use a lot of off the shelf, really simple hardware to execute this. There's Logitech, little steering wheels are over there, it feels like a big video game, you've got the big, curved Samsung screens, basic cameras on the car, so talk about the opportunity to build a software company and you're leveraging somebody else's autonomous vehicle technology to really get in the middle of this with just software, a pretty cool opportunity. >> I'll tell you what. The best time of my life was earlier this year, when I was just putting this whole thing together because it was plugging in the hardware and the software, I did it together with a team that's also here in the office. Obviously, it was more challenging because from a software person to try and build this hardware, you know, is more challenging, but I'd say today, you can get anything on Amazon, you buy on eBay a part you need, you plug it in and it would just work. So, again, we did a lot of iteration, I'd say we spent a bit more money than we were supposed to. But, that works. >> Right. And then the last piece of the puzzle that I think is fascinating is the way you're going to integrate in with other people's autonomous vehicle, so again, we talked about Waymo up the street, the Google one, Uber is working on theirs, Volvo, every day you read about BMW, et cetera et cetera, so you really get to take advantage of those hardware systems, the sensor systems, the control systems, not only from those autonomous vehicles, but you're seeing now all this stuff that's coming in factory, right, avoidance collision and radar and all types of sensors, so you will have to be able to take advantage of those different platforms and integrate your system into those various platforms. >> Right. So we would work with a company, let's say if it's one of the big OEMs or ride-sharing companies, we would know how their vehicle is set up, all we need for our solution to work is a bunch of cameras and a few modems, right, so cameras everybody have, it's one of the most essential things in an autonomous vehicle- >> Right, right. >> We would just tag into these cameras, use the modems that we need for the software to run, and that's about it. So it's a pretty straightforward solution to allow remote control assistant for autonomous vehicles. >> I'm just curious, when you're talking to customers or potential partners, what is the piece that really resonates with them when you kind of explain your solution and how it fits with what they're trying to accomplish? >> Right, so our solution is really trying to help them reach market faster, so we're not replacing anybody's work. We're adding another layer of support and safety so when yous computer crashed, when your software crashed in the car, we're going to be there with another redundancy system to support with a driver remotely. So, that's what we do at the service level. >> Okay, so can I go take a drive? >> Yeah, sure. Let's do it. >> All right, we're going to check it out, we're going to take a drive. We'll see you in the car. Thanks for watching. (upbeat music)