 Live from San Francisco, it's theCUBE, covering Micron Insight 2018, brought to you by Micron. Welcome back to the San Francisco Bay, everybody. We saw the sun rise in the bay this morning of an hour, so we're going to see the sun set. This gorgeous setting here at Pier 27, Knob Hills up there, the Golden Gate Bridge over there, and of course we have this gorgeous view of the bay. You're watching theCUBE, the leader in live tech coverage. We're covering Micron Insight 2018, AI, accelerating intelligence, a lot of talk on memory and storage, but a lot more talk around the future of AI. So we got a great discussion here on the auto business and how AI is powering that business. Jeff Bader is here as the corporate vice president and general manager of the embedded business unit at Micron, good to see you again. Jeff, thanks for coming on. And Simon Enrigger is the vice president at BMW and he's also joined by Vijay Nadkarni, who's the global head of AI and augmented reality at Vistion, which is a supplier to automobile manufacturers. Gentlemen, welcome to theCUBE, thanks so much for coming on. Thank you for having me. Thank you, thanks. So you guys had a panel earlier today, which was pretty extensive. And just a lot of talk about AI, how AI will be a platform for interacting with the vehicle, the consumer, the driver interacting with the vehicle. Also talked a lot about autonomous vehicles, but Simon, why don't you kick it off? Your role at BMW, let's just start there. We'll do the same for Vijay and then get into it. Yeah, I'm responsible for the research portion that we do globally, which is represented here in North America. And so obviously we're working on autonomous vehicles as well, as integrating assistance into the car and basically what we're trying to do is to get use AI as much as possible in all of the behavioral parts of the vehicle that users have an expectations towards being more personalized and having a personalized experience. Whereas we have a solid portion of the vehicle that is going to be as deterministic and aesthetic as we have it before, like all of the safety aspects, for example. And that is what we're working on here. Now, Vijay Vistion is a supplier to BMW and other auto manufacturers, is that right? Yes, we are a tier one supplier. So we basically don't make cars, but we supply auto manufacturers of which BMW is one. And my role is essentially AI technology at Vistion and also augmented reality. So in AI, there are basically two segments that we cater to. And one of them is autonomous driving, which is probably our biggest segment. And the second one is infotainment. And in that, the whole idea is to give the driver a better experience in the car by way of recommendations or productivity improvements in such. So that is, so my team basically develops the technology and then we essentially integrate that into our products. So not necessarily self-driving, it's really more about the experience inside the vehicle. That is the infotainment side. And then on the autonomous driving side, we of course very much are involved with the autonomous driving technology, which has to do with detecting objects and also making the proper maneuvers of the vehicle. And we're definitely going to talk about that. Now, Jeff, you sell to the embedded industry of putting automobile manufacturers. We hear that cars have, I forget the number of microprocessors, but there's also a lot of memory and storage associated. Yeah, I mean, if you follow the chain, you have Simon representing the OEMs. Vijay representing the tier one suppliers. We're supplier to those tier one suppliers in essence, right? So we're providing memory and storage that then goes into the car, as you said, across all of the different sort of control and engine control and computing units within the car, in particular into that infotainment application and increasingly into the ADAS or advanced driver assistance systems that are leading toward autonomous driving. So there's a lot of AI or some AI anyway in vehicles today, right, presumably. Yeah. David, who did a wonderful job on the panel, he thought he was outstanding, but he kind of got caught up in having multiple systems like an Apple CarPlay and your own system. I actually have a BMW, kind of a BMW, I have a Mini, because I'm afraid it's going to be self-driving cars and I just want to drive a car before they take it away from me. But you push a button if you want to talk to Siri and you push another button if you want to talk to the Mini, I mean, it's going to use it for different use cases, right? Exactly. Is that AI? Yeah, AI is also about adaption and it's also about integrating. So AI is coming with you with the devices that you have with you anyway, right? So you might be an Alexa user rather than a Google Assistant user and you would have that expectation to be able to ask to chat with your Alexa in your car as well. That's why we have them in the vehicle. Also, we have an own voice assistant that we recently launched in Paris Motor Show, which augments the experience that you have with your own assistants because it factors in all of the things you can do with the car. So you can say there is a solid portion of AI already in the vehicle. It's mainly visible in the infotainment section. Right, and of course, I remember the first time, I'm sure you guys experienced too that the car braked on my behalf and that kind of freaked me out. But then I kind of like it too. And that's another form of machine intelligence, right? Well, that counts for you that has not necessarily been done by AI because in let's say self-driving, there is a portion of pretty deterministic, rule-based behavior and exactly that one, like hitting an object at parking, you don't need AI to determine to hit the brakes. Right, and there is no portion of AI necessary in order to improve that behavior. Whereas predicting the best driving strategy for your 20 mile ride on the highway, this is where AI is really beneficial. In fact, I was at a conference last week in Orlando, it's the Splunk Show and it was a speaker from BMW talking about what you're doing in that regard. It's all about the data, right? And learning about it. And turning data into insights. Into better behavior. Into better expected behavior from whatever the customer wants. So, Vijay, you were saying before that you actually provide technology for autonomous vehicles. So, I got a question for you. Could an autonomous, could today's state of autonomous vehicles pass a driver's test? No. No, would you let it take once? No. Right answer, it depends. I mean, there are certain companies like Waymo, for example, that do a lot, but I still don't think Waymo can take a proper driver's test as of today. But it is, of course, trying to get there. But what we are essentially doing is taking baby steps first and I think you may be aware of the SAE levels. So, level one, level two, level three, level four, et cetera, and level five. So, we and most of the companies in the industry right now are really focusing more on the level two through level four. And a few companies like Google, Waymo rather, and Uber and such are focusing on the level five. We actually believe that the level two through four is the market will be ready for that, essentially in the shorter term, whereas the level five will take a little while to get there. So, everybody, of course, wants to know when we're going to have autonomous vehicles. I'm not going to ask you that question because it's like a spectrum of self-driving. But I want to ask you the question differently. And I'll ask each of you. When do you think that driving your own car will become the exception rather than the rule? Well, I'd rather prefer actually to rephrase the question maybe to where, not when. Because where, on a highway setting, this question can be answered precisely in roughly two to three years, the functionality will kick in, and then it's going to be the renewal of the vehicle. So, if you ask where, then there is an answer within the next five years, definitely. If we talk about an urban downtown scenario, the question when is hard to answer. Yeah, well, so my question is more of a social question than it is a technology question, because I'm not giving up my stick shift. Well, I will say, again, just an example, getting my 17-year-old to get his permit was like kicking a bird out of the nest. Right, get a drive. They don't care. He has a permanent driver on staff, basically, with me, right, so why? But, I mean, when I was a kid, that was freedom. 16 years old, you were racing out the door. And there is a large generational group growing up right now that doesn't necessarily see it as a necessity, right? So, not driving your own car. I think car share services, ride share, Uber, the so on and so forth, are absolutely going to solve a large portion of the transportation challenge for a large portion of the population, I think. But I agree with the earlier answers of, it's going to be where you're not driving as opposed to necessarily when, I think. Have we heard today, of course, talk about, I think the number is 40,000 fatalities on the roadways. In the U.S. In the U.S. Yeah. Everybody talks about how autonomous vehicles are going to help attack that problem. But it strikes me, talk about autonomous cars. Why don't we have autonomous carts, like in a hospital, or even autonomous robots that aren't relying on lines or stripes or beacons? One would think that that would come before an autonomous vehicle. Am I missing something? Are there systems out there that I just haven't seen? Well, I don't know if you've ever seen videos of Amazon distribution centers. Yeah, but they're going to school on lines and beacons and they're not really autonomous. That's fair, that's fair, yeah. So will we see autonomous carts before we see autonomous cars? I think it's a question of what problem does that solve necessarily. Yeah, sure. Amazon, for example. It's just as easy for them to know where something is and to paint a line so they, right? You think about Micron's fabs, every one of our fabs is completely automated. There's a material handling system that runs up and down around the ceilings, handling all the wafers and all the cartridges and wafers. Moving it from one tool to the next tool to the next tool. There's not people anymore carrying that around, or even robots on the floor, right? But it's a guided track system that only can go to certain places. Well, the last speaker today, too, was talking about it. And I remember when robots couldn't climb stairs and now they can do backflips. That's right, yes. You know, you think about the list of things that humans can do that computers can't do if the list gets smaller and smaller every year. So it's kind of scary to think about. On one hand, does the concept of Byzantine fault tolerance, you guys familiar with that? Does that come into play here? Do you guys know what that's about? I don't know what it is exactly. So it's a problem, and I first read about it. It's the Byzantine general problem. If you have nine generals, four want to attack, four want to retreat. And the ninth sends a message to half to retreat or not, and then you don't have the full force of the attack. So the concept is, if you're in a self-driving, both within the vehicle and within the ecosystem around the city, that you're collectively solving the problem. So these are challenging math problems that need to be worked out. And I'm not saying I'm a skeptic, but the more I read about it, the more hurdles we have. There's some isolated examples of where AI I think fits really well and is going to solve problems today, but this singularity of vehicles seems to be a ways off. We have a highly regulated environment. Obviously public transportation or public roads, right, are a highly regulated environment. So it's different than curating playlists or whatever, right? This is not so much regulated. Traffic and legislation isn't there yet. So especially, and it's designed for humans, right? Traffic, cars, roads are designed for humans to use them. And so the adoption to, the design of any legislation, any public infrastructure would be completely different if we didn't drive as humans, but we have machines drive them. So why are robots and cars not coming? Because the infrastructure really is designed for humans. And so I think that's what's going to be the ultimate slowdown, is how fast we, as a society that comes up with legislation with acceptance of behavioral aspects that are driven by AI, on how fast we adopt it. Technically, I think it can happen faster than you believe. Yeah, it's not a technology problem as much as it is. The public policy, insurance companies, I'm sure are going to weigh into this. I mean, if you think about, one of the events you could think of from a, let's say, even level four capable car on a highway is platooning. Yeah, right? Instead of having X number of car links to the car in front of you, you just stack them up and they're all going in a row. That sounds great until Joe Blow with their 20 year old Honda, starts to pull into that lane. So you either say this lane is not allowed for that or you create special infrastructure, essentially, that isn't designed for humans or is more designed specifically for the machine-driven car, right? How big is this market? It feels like it's enormous. I don't know, how do you look at the TAM? We can talk for the memory, I can talk the memory storage part of it, right? But today, memory and storage, all the memory storage for automotive is about a two and a half billion dollar market that is going to triple in the next three years and probably beyond that, my visibility is not so good, maybe yours is better for sure, but it's then really driven by adoption rate and how fast that starts to penetrate through the car OEM lines and across the different car OEMs. And Vijay, your firm is, when were you formed? How long have you been around? Oh, I've been around basically since around 2001. Yeah, okay. Relatively new. And then Ford spun out Whiskey on that at point. Right, okay. So, all right, so that's innovation. Got Ford, then of course it'd been around forever. But it's Greenfield for your group. Right. Whereas BMW, this is transitional, right? So, is it, are you trying not to get disrupted or are you trying to be the disruptor or is it just all sort of incremental business? Well, as a 101 year old company, obviously people think about you as being ripe for disruption. And I think we do quite well in terms of renewing ourselves coming from an aeroplane business to a motorcycle business to car business, right? It's cause you're cool. And so I think the answer is, are we fast enough? Are we fast enough in adoption? And on the other hand, it's fair to say that BMW with all of its brands is part of a premium thing. And so it's not into the mass transportation. So everything that's going to be eaten up by something like multi-occupancy vehicle, mass transportation in a smaller effort, right? This is probably not going to hurt the premium brand so much as a typical icono type of boxy car. Exciting times. Gentlemen, thanks so much for coming on theCUBE. We got a run. Appreciate it. Thank you so much. Thank you so much. Okay, thanks for watching everybody. We are out from San Francisco. You've watched theCUBE at Micron Insight 2018. Check out siliconangle.com for all the published research. theCUBE.net is where you'll find these videos. Wikibon.com for all the research. Thanks for watching everybody. We'll see you next time.