Leadership Insights: Transforming the Driving Industry

Transcript

Mike Vizard: Hello, and welcome to the latest edition of the Digital CxO leadership insights series. I’m your host, Mike Vizard. Today we’re with Padma Sundaram, who is director of software for GM. And we’re going to be talking about what does it take to get a company that has traditionally been obviously very hardware focused become a software company because your cars, trucks and whatever else you’re driving around these days is all pretty much dependent upon some type of code somewhere. Hi Padma. Welcome to the show.

Padma Sundaram: Hey, thanks Mike. Thanks for having me. Excellent question. I think the automotive industry is going through a huge digital transformation. What I mean by that is software is not new to the car. There’s plenty of software in today’s cars. If you take an advanced vehicle, you already have like; some studies have said like 150 million lines of code in an advanced automobile today. So what is the big thing about digital transformation in a software defined vehicle? But essentially, these traditional vehicles, even though they have a lot of software, software is very much deeply coupled with the hardware, meaning hardware and software go together. And making any upgrades post production to that software becomes next to impossible, if not, at least a trip to the dealership. So because they’re so tightly coupled, and every module of software, there’s so many – like 150 modules in the vehicle, that’s what they estimate. And you have this software talking to each other, modules talking to each other, there’s so much complexity. So what this software defined vehicle and this transformation brings is, now you’re trying to develop an end to end platform that is completely hardware agnostic to software, and your software and hardware are decoupled. Now once you’ve had that, and you have a standardized API from the platform, the software can evolve on top of it. Multiple criticality software’s can exist, multiple operating systems can exist. And this way, because the car is also the platform, it is digitally connected to the cloud at all times, you can continue to do over the air upgrades, updates and provide new software content, or better software content to the customers; the customers can even get absolutely new content that they never had when they purchase a car. So this allows the car to be continuously refreshed. And it also connects well with their digital life in general. And it becomes like an extension of their digital life.

Mike Vizard: We hear a lot these days, of course about all things related to cloud-native and microservices. And the assumption is is that all that stuff is somehow running up in the cloud, because well, it’s called cloud-native. But in reality, you’re using a lot of the same concepts to build applications that run in the car. So is the car essentially now a micro cloud of some kind or another?

Padma Sundaram: We are intending to go there. I think cloud-native technology is a critical enabler for us for software defined vehicles, because it allows us to develop fast, iterate fast testing; the cloud deployed, and we get to monitor the cars, right? We have connectivity to monitor the cars – if something isn’t working, we can quickly fix it through our backroom. So yes, cloud-native provides us the features, and the flexibility that’s necessary to continuously provide new feature capabilities, and new offerings to the vehicle.

Mike Vizard: As we go forward, also, the car is getting smarter or the vehicle and we’re putting more AI and machine learning algorithms in there as as that evolves. But the models at least as I understand that today are trained up in the cloud, and then there’s an inference engine that sits in the car but at some point will the cars and vehicles themselves actually start to learn and be trained with some sort of AI modeling capability locally?

Padma Sundarum: Absolutely. Artificial intelligence is critical to the decision making of automated drive resisting systems. For example, you need perception systems to provision and other sensors to understand the world around you, to perceive it, and to predict the environment and to plan it. So there is a lot of artificial intelligence, and the car continues to collect data and we continue to train it offline. So the vehicle continuously is learning essentially, and becoming better, with the continuous data collection and perceiving and solving the problems for the environment. But artificial intelligence machine learning are critical pieces for us if you are to get into autonomous and automated driving, which we are already doing; if you think of think of it – super cruise – that has artificial intelligence. And we already have a partnership with Cruise Automation. And we are working there with Cruise to make sure we are deploying self-driving vehicles. And we just did in San Francisco, and we’re planning to do it in other cities as well.

Mike Vizard: How personalized will I be able to make these environments? And I’m asking the question because at least I have a dream in my head somewhere that says I’m gonna subscribe to a car service. And every year someone is gonna come around with a brand new car and all my settings in that car will just automatically reset as I moved from one to the other. I mean, can I get to this kind of digital vehicle lifestyle?

Padma Sundaram: Absolutely, I think that is one of our goals. You know, usage of smartphones has already, you know, users have certain expectations. So when they move from one car to another they carry their settings, how the car feels, and because they have individualized the car, and we allow them to individualize the car. Absolutely, that is one of the key features when you go in, and you sign in into the car, just like your phone; it carries your settings, your options, what you prefer, all those things carry with you from car to car.

Mike Vizard: One of the things that I do perceive and of course, I’m sure my friends in the software development community will take some exception when I say this, but the cloud and all the horsepower that we had over the years allowed us to get a little sloppy on programming. But as we move to start putting software inside of vehicles, the amount of horsepower or hardware that you have is somewhat limited. So do I have to be a better programmer these days to create software for these platforms, because I have to get a lot more done with a lot less compute?

Padma Sundaram: Well, that is always true; you’re always going to have more feature capabilities and more need for content than what your hardware allows. Although having said that, with the advanced compute that we have these days, with GPU, CPUs and accelerators, we do have a lot more than what past cars had. And in terms of cloud computing, I’ll just say this. Cloud computing has always been there; those technologies have always been there. But it didn’t come to the cars. The reason being, cars have unique requirements in terms of real time requirements, in terms of safety, functional safety, in terms of cybersecurity. So now you see companies starting to focus on that. And there are many initiatives where companies are partnering to figure out how to do this. And there are hardware companies that are coming up with this very powerful compute that allows us to work in the car. But since the car is also connected to the cloud, there’s always some offloading that we do. And we can get things done in the cloud as well, something that’s not real time.

Mike Vizard: How will manufacturers of vehicles compete with one another as they go forward into this brave new digital world? Because it used to be kind of about the specs of the car, but maybe it’s going to be about something more than that, as we go forward it is going to be more about the experience, I think?

Padma Sundaram: Absolutely. You hit it on the nail. I think traditionally, if you see it has been more about horsepower, how fast my car can go, how does it accelerate and things like that. But more and more, you will still have those features. I’m not saying those will go away; those will always be there, because there will be a fan base for that. But in addition, the car becomes now more about the experience. And experience is not just limited to your driving, but what kinds of modes you want, how do you want the car to accelerate and apply brakes for specificity for automated features, level two and above features? And then it’s also about how do you want to have the experience of temperature and visibility? How do you want to control that? And entertainment – there’s a lot of entertainment for users and vehicles. And so you will see it becomes more and more about the user experience, end user experience. And they can provide us feedback. And because the car is connected, we can use the feedback to further improve and understand which features are working, which features are not working. And basically that allows us to improve those features that actually add value to the vehicle as well as the user.

Mike Vizard: Just about everybody I talked to is struggling to find the talent to build this brave new world that we’re talking about. So how do you all go about attracting the right skill set, the folks that know how to build these kinds of applications; where do they come from?

Padma Sundaram: Absolutely. That is, you know, if you think about what are the challenges in this digital transformation, having a sustainable talent pipeline; a continuous talent pipeline is one of the challenges, no question about it. The reason being as we are becoming, or I’d say, automotive companies are becoming more and more technology oriented. You’re competing for the same technology resources that traditional technology companies are competing for. So now you only have a limited pool of candidates, and everybody’s competing for that. So how do you distinguish yourself? Well, I’ll just say from my personal experience, and that can probably be something that I can share with people who are looking for what does an automotive company provide, one is GM has created an environment of startup for these critical organizations within a larger organization. It’s like an island within a larger company. For autonomous, we have a separate startup like organization for software defined vehicles; we learned a lot from our Cruise experience, and brought all those learnings that work into this software defined globalization. So what that does is, it brings you the stability of a larger company, the backing of a larger company, but allows and gives you the flexibility of working in a startup. So to me personally, I can innovate, I can fail fast, I can be flexible, I can adapt very quickly, and that’s the startup advantage that I get in this environment. At the same time, I have the stability of a larger company. So that’s that is the key thing. And what you’re doing is actually impacting the world, with two cars. And I always say when I go into these recruiting events, you know, technology and cars are exciting things. But when you bring them together, oh, my goodness, you can do really magic. So that’s the thing that you get to do here; you get to apply your technology. And you get to really enhance that core experience.

Mike Vizard: You guys are also working closely with Red Hat. And some of that focus is on cloud-native technologies. And I think part of the issue as we go forward is you’re making an effort to use standard technology. So it becomes easier to both build the apps, but recruit folks, so you don’t have to go find as many hardcore specialists. So is that a fair assumption is that that’s part of the thought process?

Padma Sundaram: Well, talking about Red Hat. So Red Hat – our collaboration and partnership with Red Hat is a significant moment, especially a collaboration between a traditional automotive company, and an open source technology provider. Right? And that is huge, if you look at our platform that’s going to power and enable us to define a vehicle, that’s a Linux based platform. And we intentionally did that so that we can attract third parties, and open software providers to very quickly build and integrate software for us. And that can add value to the vehicle. And of course that software will be going through certification and things like that. But this is what we want to do. We want to attract third party app developers to provide the apps, develop the apps and create very quick innovation that can add value to the user.

Mike Vizard: That really doesn’t sound much different than the smartphone that you were talking about before where basically, it’s a platform and you want people to build apps for it.

Padma Sundaram: Absolutely, you can think of it as a data center on wheels, a powerful data center on wheels.

Mike Vizard: So what’s your best advice to folks who are going to build those apps or who might want to come work for GM? Is there something that you look for, or are there characteristics or attributes, where you’re gonna go, yeah, these are, this is the perfect fit?

Padma Sundaram: Sure, I think, obviously, we are looking for talent who want to make a difference at the end of the day. And this opportunity of working in GM gives you a unique opportunity because of the scale that GM has. GM, once it has this platform, we’re going to deploy it across four of our brands globally. And the impact of that you can see touching so many people around the world and continuing to improve that – it’s not like a done deal that we put the car out and it’s done; it is a living thing. It’s a continuous learning machine. And you get to learn with it. And you get to collect the data and learn it and continue to improve. So it’s an exciting period for anybody who wants to be in that space of making a difference. to innovate and learn. This is the best opportunity in automotive space, I’ll say the best time to be in any automotive company like GM.

Mike Vizard: How do you strike a balance between the capabilities that you’re describing and privacy? Because, how will we admit that perhaps I may not drive at the speed limit as much as I should, but you know, I’m not sure I want to share that with everybody?

Padma Sundaram: I think the data, the data we’re talking about really is not worried about your individual style. That’s not what we’re worried about. What we are looking at is, how is the feature performing. Is the feature performing to its intended function? Is it providing value if somebody is not even using the feature when they opted to get it? And that gives me clear feedback they don’t like it. And so that is the idea here in terms of using the data to improve the feature; we are not worried about any personal data. That’s, that’s none of our business. And we will respect privacy, like anybody else. We don’t want to get into that. Our goal is to improve continuously the user experience. We want to be able to make this car the best as it can be. And it’s going to continue to evolve. So you might have purchased a car. And then one year later, two years later, it’s going to continue to be upgradable, updatable, and maybe more refreshed than when you got it in terms of features and capabilities.

Mize Vizard: Well, I think the bigger issue is not so much that I didn’t like it, it’s just I couldn’t figure out how to make it work.

Padma Sundaram: That is a good point that also gives us the feedback of how useful a feature is, right. So if somebody got it, and they’re not using it, well many things could be wrong, but they just don’t like how it functions or how the user interfaces. And we want to make sure it’s user friendly. And it’s able to do and be able to take their inputs and provide the functionality that it’s intended for. So yes, that’s also a critical input for us.

Mike Vizard: The last question: We have user interfaces in the car today. But I often wonder if that will give way more towards speech or any interfaces where, you know; I’ll have a dialogue with the car, and I might even give the car a personality, but it’s going to be my digital assistant on the road, but how much of the future of user interfaces is going to be some kind of spoken conversation?

Padma Sundaram: Absolutely, we are looking into all technologies. Voice activation is one of the critical features we are looking at. So we do want the car to be like a assistant where it can respond. So that’s all part of the digital experience.

Mike Vizard: All right, so I guess I want to get in my car someday and have it tell me the three things that are gonna get me in trouble today. And we’ll see what it comes back with. Padma, I want to thank you for being on the show and spending some time with us.

Padma Sundaram: Absolutely, thank you so much.

Mike Vizard: And thank you all for watching this episode. You can find this one on the Digital CxO website along with all our others. And we invite you to check them out and come back and visit us again next time. Thank you all.

Padma Sundaram: Thank you. Bye.