We’ve got a special episode of Decoder today. I’m talking to General Motors CEO Mary Barra and new GM Chief Product Officer Sterling Anderson about a lot of big news the company just announced.

That includes a Google Gemini-powered AI assistant that’s coming to new cars, and an entirely new hardware and software platform coming to the Escalade IQ in 2028 alongside true Level 3 autonomous driving. There’s also a new home battery business and a new robotics division. 

It’s a lot, and it all comes against the backdrop of President Trump’s trade wars, tariffs, and the expiring EV tax credit here in the United States, all of which have really upended the car business. Just last week, the day before I talked to Mary and Sterling, GM took a $1.6 billion writedown on its EV business against falling demand.

A lot of long-term plans about the EV transition are falling by the wayside like this. So, I wanted to know how Mary was thinking about it all, especially since she made some of the most aggressive EV platform bets among the legacy automakers several years ago. In one important way, that bet really paid off: GM has a full lineup of EVs running on a mature platform now. I myself just leased a Cadillac Vistiq SUV to take advantage of the tax credit before it expired. 

Verge subscribers, don’t forget you get exclusive access to ad-free Decoder wherever you get your podcasts. Head here. Not a subscriber? You can sign up here.

But the market has changed dramatically, and consumers are becoming far more price sensitive. The average cost of a new car in the United States just broke $50,000 the first time. And we’ve heard from a lot of car CEOs lately who say that’s a big problem. Most consumers don’t want to pay much more than $30,000 for a new car, and that means all of these costs need to come down. 

So I asked Mary how she’s navigating the current moment, her company’s relationship with the Trump administration, and why she’s confident that EVs, autonomy, and AI are going to continue to help GM sell more cars — instead of just selling more expensive ones to a smaller group of wealthier consumers.

I also obviously took the opportunity to get deep into the details of the platform with Sterling, who spent several years at Tesla and was the co-founder of Aurora, the autonomous trucking startup. GM poached him back in May of this year, and now it’s his job to oversee the entire end-to-end experience of both gas and electric GM vehicles. 

That means he has to answer for decisions around the hardware, software, and the interface, as well as all of the trade-offs that come with those decisions. I had a lot of very specific feature requests for Sterling, as well as some big-picture questions about what it means to think of these cars as platforms.

Of course, that means we spent some real time on GM’s big decision to ditch Apple CarPlay in Android Auto, whether that decision is paying off, and what all this looks like in the future as AI voice assistants and more capable autonomy come into the mix. That’s a whole lot of big Decoder themes in this conversation, and Mary, Sterling, and I really got into it. I think you’re going to like this one. 

Okay, GM CEO Mary Barra and CPO Sterling Anderson. Here we go.

This interview has been lightly edited for length and clarity.

Mary Barra, you are the chair and CEO of GM, and Sterling Anderson, you’re the chief product officer at GM. Welcome to Decoder.

Sterling Anderson: Thanks.

Mary Barra: It’s great to be here. 

I am very excited to talk to you both today. There is a lot of news to discuss. I’m going to just go through the list and see if I get it all. You have the introduction of a Google Gemini-powered assistant in your cars in 2026, a next-generation hardware and software platform that will come first on the Escalade IQ in 2028 and that will enable the next generation of Super Cruise, which will let you have eyes-off driving and additional hands-off driving in the Escalade IQ. 

There’s a new robotics division, which is a big bet, and then new batteries and new home energy solutions. All of that is coming in the context of what I would call huge changes in the car market: the end of the EV tax credit, tariffs that are upending the global supply chain, and what appears to be a brewing trade war with China over rare earth metals, which are essential to all of these plans. Did I miss anything?

MB: We’ve been pretty busy. I think you captured most of it, but I would say in the midst of all of that, we have a great product portfolio of both internal combustion engine vehicles and electric vehicles, and the customers are responding very well to it. One of the things our team has developed over the last handful of years is the ability to be very resilient and agile to respond to all these challenges. I’m really proud of the team. I’m proud of the vehicles and the services we have today. What we’re going to talk about is where we’re headed. It’s an exciting time at General Motors.

I want to come to all that news, and what you’re describing there is what I call Decoder bait — this is how you are running a company of this size to make all these bets and be flexible. And then Sterling, I don’t know if they warned you, but I have some very nerdy questions for you and very specific feature requests about my Cadillac Vistiq in particular.

SA: I love it.

We’ll come to that as well. But Mary, I want to start with GM and some of these bets that you’ve been making for a while now. I would say that it’s been almost five years that you’ve been aggressively positioning GM for an electric and autonomous future. Just this week, though, GM announced a $1.6 billion writedown in the third quarter against changes in the market and falling EV sales. 

So I’m curious about the big picture. How would you describe where you want GM to be, and how far you are along on that journey? And how things have changed in ways you were expecting and not expecting?

MB: I would say, first of all, it’s being very customer-focused and making sure we have great products with the right design, the right quality, the right features, performance, the right software and services that we can put on top of that that customers want to buy. So we start there and that’s where we want to head. 

From an EV perspective, we still believe in EVs and I consider them to be our North Star, but clearly, when you have a dramatic shift like we’re seeing in the regulatory environment in the United States and potentially other places, as well as a very substantial change in the consumer tax credit, that’s going to change buyer behavior. When you think about vehicles, they’re either the most important or the second most important purchase that a customer makes, I mean, it’s very significant, and we get to be a part of that with our consumer.

As we look at these changes, we need to make adjustments, but we’re going to meet the customer where they are, whether they want a great [internal combustion engine] vehicle or they want an EV. The adjustment that we made today from the charge was really reflecting what we believe now will be a slower EV adoption, but we still have a great portfolio of award-winning EVs that we’re going to be offering. So, we’re going to meet the customer where they are, but when you have that dramatic of a change that happens in that shorter period of time, you have to make adjustments to run the business well.

Let me ask you about that specifically. Obviously, the tax credit is gone, and it drove a lot of EV adoption. I am one of those people. I rushed and made sure that I bought my Cadillac before that tax credit went away. I’ll tell you, the dealer I bought my car for, he gave me a great deal, and he told me, “I will move as many EVs as I can because that’s how we get allocations on gas-powered Escalades,” and in particular, I think the CT5-V Blackwing. 

That’s the thing he wanted — he’s like “I’ve got to move this many EVs to get allocations on the car where I’m going to make a lot of money.” That’s all changing: that table has been flipped, the tax credit’s gone away, and we don’t know what the market’s going to do. Are you taking the $1.6 billion charge preemptively, or have you already seen the change such that you’re taking that charge?

MB: One of the things that we made an adjustment already was the fact that we were planning to have one of our factories in Lake Orion, Michigan be an electric plant. When we looked at the shift and what’s happening with tariffs, we decided that we were going to expand the capacity, because right now, for instance, full-size utility vehicles, we can’t build enough of them, because the demand is so strong. 

So we’re going to use that plant. Once you make a change like that, that’s something that already has happened. I don’t think we’re going to know for sure what true EV demand is until next year, because, like you and many others, there was definitely a pull ahead for people who wanted access to the $7,500 credit.

So we’ll see potentially a little bit lower uptake in the fourth quarter. We don’t know yet. As we get into next year, we’re going to see that. But again, when you talk about $7,500 on a vehicle, that’s pretty substantial. We do expect — and I think the industry expects, and the external forecasters believe — that we’re going to see slower EV growth, but I think the important thing is we think we’ll still see growth.

Can I put that in the context of new versus used cars? I see what’s going on with new cars, most new EVs are leased, and every company, including GM, played games with residuals on leases to bring those prices down on top of the tax credit. And then I look at the used market and I see EV depreciation is out of control.

Just financially, buying a used EV might be the only smart car purchase you should make today. That’s not growth in new sales, but it’s obviously more people coming into the EV market as owners, as having the experience. Do you think that will have any effect on how you market and sell your new EVs?

MB: First, I would say I don’t think we played games with residuals. General Motors has been very disciplined for more than a decade about understanding what pricing does, and how when you make dramatic changes in pricing, what that does to the residuals, and that it takes almost a new generation of that vehicle to get the residuals to be where they’re at. So we manage that very, very carefully because we want to make sure the people who buy our vehicles keep that value. 

Now, when someone leases, that then becomes the leaser. Or in our case, many of our vehicles are bought through GM Financial, but we look at that as well because it’s still something very important. But again, I think it’s too soon to tell. Generally, what we see, there’s a new car buyer, and then there are people who buy used vehicles. And so we’re just going to have to see what the vehicle equation is. But I think one of the things you’re getting at is that the consumer is very rational. I think they’re going to be looking across the board.

SA: If I can add one thought there — Nilay, you’re now on a Vistiq, so getting a little more of this experience with some of these cars. For many, getting over the hump of that kind of initial activation energy of being comfortable with an EV, being comfortable with the range, getting comfortable with setting up Level 2 charging in their home, and these other things, I think there is incremental value to, as you say, the penetration that we get in the secondhand market for people who get comfortable with these vehicles. They are better cars, they’re better on a number of dimensions, and as that comfort grows, I do hope that it’s value-accretive to future purchases.

MB: And to your point, Sterling, we have already seen with our customer feedback, that once a customer owns an EV, because of those benefits, they are very much likely to buy another EV. But we’re still in early days, and again, I think we’re going to see as charging infrastructure becomes more widespread across the country, we’re going to continue to see EV adoption grow. But the great thing, the beautiful thing, from a General Motors perspective, is that we’ve got a vehicle for whatever they choose, and I think that puts us in a very unique position as we move forward.

In one way —  don’t take this wrong way — what you’re saying here is what I’ve heard about EVs for a long time: they are better cars, you don’t have to do the maintenance, they have instant torque, you charge them at home, and you never have to think about going to a gas station again in your life. Those have been the arguments for some time. I think in the pandemic, a lot of companies reacted to what seemed like infinite demand for the Tesla Model 3; they couldn’t make them fast enough.

You made a lot of bets, and those bets right now are paying off. I would say my car feels like the payoff of a big bet you all made some time ago. It is a mature platform, it’s a great car, I have no complaints with it. I mean, I do, but we’ll come to that, Sterling, just you wait. Still, it’s very hard to nitpick that car.

But then the whole market changed around you. That infinite demand for Tesla products is no longer infinite, even for Tesla. Even Tesla is doing sales and pricing changes, and we’re seeing the tax credits go away. Do you think it’s your gas vehicles that are going to carry you through this period of uncertainty, or do you think the EV sales are actually going to hold their own?

MB: Again, I think it’s too soon to tell, and I’ve been in this job too long to make predictions like that. We’ll see. We’re well positioned either way. But I think an important point is that even in this period over the last two years, General Motors was taking share. We were growing share and continue to have vehicles that were highly rated year after year and model after model. 

We’re well-positioned. We’ll meet customers where they are, but I don’t think that it’s dramatically going to change. I think our internal combustion engine vehicle business has been important. We’re still on a journey to get to profitability from an EV perspective, so we’re going to continue on that journey. And we’re investing in technology that takes cost out of the vehicle while taking nothing away from the customer. So we’re going to be in a good position there as well.

California Governor Gavin Newsom says GM sold out to the Trump administration over this tax credit. I’m just going to read you the quote: “Mary Barra sold us out, eliminating Ronald Reagan’s work, eliminating the progress we made under the California Air Resources Board in 1967. The Republicans rolled that back this year under Donald Trump’s leadership, but the American automobile manufacturer allowed that to happen. GM led that effort.” Do you have a response to that?

MB: I would say we worked tirelessly for more than a year, trying to make sure that the states that were following [California Air Resources Board]… We have worked with CARB for decades, but we got to a point where when you looked at what was required for model year ’26 — as you know, those sales start in calendar year ’25 — those states following CARB needed to get to a 35 percent EV penetration. The consumer wasn’t there. 

We worked with the different states, and we worked with CARB, but it got to a point where we were weeks away from not being able to sell customers the cars they wanted because the penalties were so strict in some cases. So we have always consistently — for my entire term here as CEO — been really working hard to get one national standard, because we have really strict regulations across the board, and to have it varying by state makes it even more complicated.

We’ve always said we needed harmonized standards, and that’s what we’ve always worked toward. In this situation, there were other automakers who equally supported getting the [Congressional Review Act] passed. Apparently, I was the one singled out there, but I still stand by what we did because we worked for a long time, and I think there were a lot of governors and other lawmakers who were relieved that CRA passed because we were approaching a situation that could have had an impact on jobs, on plants not being able to produce. Literally, we would’ve been limited in what we could sell.

On the one hand, the Trump administration made it easier for you to sell the cars you have and the cars you’re making today without what you’re describing as the regulations that would’ve prevented you from doing that. On the other hand, we’re talking about the tax credits going away and we’re talking about costs rising for the automotive industry because of Trump’s tariff policies. 

Does that feel like a holistic conversation with the Trump administration about the automotive industry in America, or is it all things happening at once? Because from the outside, it feels like no one’s talking to each other.

MB: We are talking a lot to the administration on many aspects, on making sure that we have a strong manufacturing base in this country, and on making sure that we have strong automotive companies, American automotive companies, getting a level playing field. When you look at what’s happening in tariffs, I’ve long said, “Let us have a level playing field and we’ll compete,” but we have not had a level playing field with many countries for many, many years. 

So, as you make those changes, yes, it can be disruptive, and we’re making adjustments so we can compete in that market. But when you look at the change from a regulatory environment, again, we don’t set those changes, we just try to help people understand it was very clear the Trump administration felt strongly about wanting consumers to choose versus being regulated into having to buy a certain vehicle.

I’m a big believer in consumer choice, and for years I’ve been saying I want to get enough EVs out there that are great EVs, that a consumer chooses it because they love it and it’s the car they want to drive. Once you get into regulations driving consumer behavior, you’re already in a difficult place. So yes, has a lot happened this year? Absolutely. But I think it’s what we do and how we respond. We responded to COVID, we responded to the semiconductor shortage, and I’m really proud of the team. 

This has been some change, but we were going to need change regardless. The regulatory environment was getting in front of the consumer, and, largely, the charging infrastructure had built up as much as I think people thought it would be or where it would advance by the time some of these regulations kicked in.

So to me, I look at tariffs as being about a level playing field and having a strong manufacturing capability in this country. I think the regulatory change was about consumer choice. We want to just do great vehicles that people choose, and that’s why I still say EVs are our North Star. And I’ve had many conversations with the administration and with the President about that. I fundamentally believe EVs will continue to grow over time, albeit a little bit more slowly because we don’t have the consumer tax credit anymore, because they’re great vehicles.

Do you think that consumers are going to feel the impacts of the tariff policies on GM or other carmakers the way that they feel it on other products? We just ordered clothes for a wedding and we got hit with tariff charges by DHL at JFK. Is that going to come to consumers in a real way in the auto industry, or are you managing against that? Because we’re already talking about prices going up in other ways.

MB: We’re looking at how we maintain our focus on affordability. How do we continue to drive efficiency? How do we comply with tariffs to make sure that we aren’t paying a big tariff bill? And then we’re also continuing to communicate with the administration and members of the cabinet so they understand some of the unintended consequences of some of the policies. 

I have to say they’ve been incredibly receptive to and really have become students of the industry, to make sure they understand the ramifications. They’ve made adjustments to make sure we’re working toward having strong manufacturing in this country and having a level plane.

Can I just ask you one more question about the broader car market? Then I want to do the Decoder questions, which I think will lead into all of the news here. You’re talking about affordability, we’re talking about your cost rising with tariffs, consumer cost rising as credits go away, all these sorts of things. Broadly, the car market is as K-shaped as the rest of the economy. 

I think I saw a report today that said average car prices are headed north of $50,000, while defaults on financing are at an all-time high. That’s not a good split, right? The cars are getting more expensive than ever and people are not able to pay their bills. A lot of people can’t afford these ever more expensive cars. Whenever we publish a story about a new EV, the first thing that people say in our comments is, “That’s too expensive.” How are you thinking about that?

MB: That’s why I’m extremely proud of what we have from a Chevrolet perspective, whether you’re looking at the Chevrolet Trax and Trailblazer, which are internal combustion engine vehicles and start at around $20,000 and then go up, or you’re looking at now the Bolt that we are bringing back and just announced will be at that $30,000 level. The Equinox is in the mid-$30,000 range, too. 

I think we’re offering a great choice for the consumer in a very affordable price range with great vehicles that have beautiful design and incredible safety. If you look at the screen on the Equinox EV, it’s a large screen with great safety features. I understand what the broad industry trends are, but when I look at it from a General Motors perspective, we’re doing extremely well in these entry-level segments in both ICE and EV, as well as really strong work at the top end.

Some consumers are saying they want a more premium truck, they want a more premium SUV. And so we’re meeting that demand when you think about a Denali on the Yukon, or the Cadillac Escalade and Escalade IQ. But then we also have great vehicles that are very affordable. So we’re a full portfolio manufacturer and I’m very proud that we can meet customers where they are.

There are not many car companies that have a range from $20,000 up to… I believe that the most expensive is this Cadillac Celestiq at $300,000. That’s the full range. I will warn you, Sterling, that we are going to look at what’s on the screen very soon, trust me. It’s coming. 

Let me ask you the Decoder questions. I think people understand GM as nameplates. There’s GM, GMC, Chevy, Cadillac, Buick, and all the rest. How is GM structured? As a company, as an operating company, how is it structured? Is it just the divisions and the brands, or are you more functional than that?

MB: I think we’re more functional, and Sterling is responsible for our products globally as the chief product officer. We leverage that to get the scale, but then we also understand for the different brands what we need to do to make sure we’re delivering on that brand promise, whether it’s true luxury with a Cadillac value, or the unique American heritage that Chevrolet has. Buick is premium and has been doing very well and growing, and then GMC, which is a very premium truck. 

Each of our four brands are very specific for what they mean to the customer. We make sure when we do a portfolio of vehicles or when we’re doing, for instance, full-size trucks, that we understand what it means to be a Chevy, what it means to be a GMC or a Cadillac. We look at all of that. It’s really a cross-functional team working together to make sure we deliver on that. We do have regional sales and marketing teams that really understand the customer in each of our markets. So at a broad-brush, at the most senior level, that is how I would say the company is organized.

Walk me through that in a little more specificity what you’re describing. It sounds complex. So you have centralized product development, we’re going to make a pickup truck, and then you have brand expressions of but one pickup truck. How do they decide what goes where? Do the brands themselves have teams that get to do it? Do they have their own designers? How does that work?

MB: Sterling and I just came from design, and we were looking at a platform that’s going to have both a Cadillac and a Chevrolet. There are the brands in the studio, where they have people dedicated to Chevrolet, and then a different team that’s dedicated to Cadillac. But then we also have product-planning people who are looking at the technology roadmap, what features and functionality need to be on that vehicle. 

How are we going to make sure that we achieve how we want to go to market from a price perspective, getting back to that affordability or that true luxury? But it’s done with people. For instance, the brand is managed for the globe as well, so it’s not like a Chevy in South America has a totally different team managing it than a Chevrolet in the United States or in another part of the world.

It’s not really complex; it actually simplified it. We get that what is so important in the auto industry is scale. Because once you have scale, that’s how you give the customer more than they think they would get for a price point. We’re leveraging General Motors’ scale, and I’m very proud that we sell more vehicles in this country than anyone else, but we’re also extremely strong in South America, we have a great business in the Middle East. We still have a significant business in China. Leveraging that scale is one of the things that allows us to keep vehicles affordable or give people more functionality than they expect.

That’s actually one of the things I’m most interested in. There’s the US market and the markets that are related to the US market, or have similar tastes and desires, similar brands competing, and then there’s the Chinese market. I watch a lot of car YouTube videos for better or worse, and car YouTubers are increasingly saying, “Look at these Chinese EVs, they’re better than the cars that we can get here. United States protectionism is keeping us from having these cars.” They go through all the quirks and features of the cars, and I think that’s great because I really enjoy watching car YouTube. 

If I was in your seat, I’d say, “Oh, boy, the stuff we have to do to compete in China is radically different from the things we have to do to compete in the United States.” How do you manage that split across this portfolio of brands?

MB: You have to use technology in China for China because that’s the requirement in China. So already, there’s a little bit of a fence around that. And right now there is an incredible price war going on. You can’t have over 100 different OEMs in a country trying to compete, especially now as they’re competing on price. And if you read what’s happening in-country right now, there’s a lot of change happening. I would also say the market is over capacity in China, which is causing, from a business perspective, exporting to other markets, but they’re also doing it while highly subsidized in many cases. We are regularly benchmarking our Chinese competitors, like we do for global competitors, understanding where they’re putting features, where they’re not, what they’re adding.

But I would also say we’ve got to meet the regulations in this country. The safety standards are different, there are different emission standards, and there are also standards and requirements around connectivity. Like I mentioned, for instance, autonomy in China, you need to use a system that’s Chinese-based because the country actually controls the maps. So there are also regulations or executive orders — I can’t remember exactly — but there are rules we have to follow about what we’re able to use in-country. 

There are already differences. Just to look at what the offering is and not understanding what the rules of the different markets are going either way is, I think, a very big simplification. I don’t know, Sterling, if you have anything to add with what you saw at your time at Tesla, or just even when you were at Aurora, and understanding what markets you could compete in. 

SA: As Mary said, as we benchmark across these vehicles, if we strip away the subsidies and the other support that the Chinese government is providing to many of these players, we compare favorably on most dimensions with our EVs. In particular, we’ve got some work that we’re doing around a next-generation electric vehicle and electrical architecture that compares very favorably to what we see. So I’m impressed by what they’ve done, I’ll start with that. We don’t want to be resting on our laurels that way. I think there’s more that we can be doing, but we’re on it. We recognize very acutely the speed at which an industry that is so heavily subsidized tends to move and we don’t intend to wait for it.

MB: Sterling makes a really great point. We are regularly looking at how we make every part of our business more efficient. We don’t control what the government policies are; it’s evident by what’s changed this year. So we just need to have the best product that we have as efficiently as possible, and that’s what we’re focused on every day.

We just had Ford CEO Jim Farley on the show. My friend, Joanna Stern, interviewed him while I was on leave. She filled in for me, which was very nice of her to do.

Farley is very clear that BYD makes a better car than Ford. He drove around BYD for a while and publicly announced, “This is better than our car.” On our show, he said to Joanna, “We’re going to have a better car for you when your Mach-E lease is done. It’s going to be a better car; don’t lease it on Mach-E.” To do that, Ford had to set up an entire skunkworks. That was the payoff of the structural change inside of Ford: we’re going to take a bunch of people, we’re going to put them aside, we’re going to take them out of the Ford ecosystem, let them rethink the car from the ground up.

They also announced a new architecture. They announced a new manufacturing process that was notably led by a Tesla veteran at Ford. Sterling, you worked at Tesla as well. You’re at GM now; did you need to do anything similar inside of GM’s culture to get to your new platform, your new ideas?

SA: I like [Ford EV chief] Doug Field a lot; we worked together. I also worked with Alan Clarke, who’s been leading the skunkworks program for Ford. Both great people. So nothing about that. What I will say is I’ve seen fairly oscillatory behavior from some OEMs when it comes to EVs. In one moment, it’s  “We’ll force an electric powertrain into an ICE platform and we’ll end up with a fairly compromised product as a result that turns people off to certain segments — electric battery, electric trucks, some things that just weren’t great about the way that that started in some of this space.” And then I see a pendulum swing all the way to, “It’s got to be a skunkworks project, it’s going to be done fully independently from a clean sheet, and then we’re going to somehow try to ingest it into the broader organization.” That ingestion is where the risk lies. What I’ve seen is that General Motors has been doing since before… I only got here four months ago, so there is only so much that you can attribute at all to me at this point.

But what I’ll say is, General Motors has taken a much more steady approach to this development: electric vehicles designed, developed, and manufactured on electric vehicle architectures, which are not nearly as compromised. We focus on the cost basis for these vehicles, on what the value-added components that we can put in them are, and what the things that just don’t add that value are, so that we can strip out and refine, such as a reduction in menial stuff, like spot welds. 

For some of our next generation platforms, there will be a dramatic reduction in those, you’ll never see them, you’ll never hear about them, but they drive an enormous cost. I think at one point, the estimate was $40,000 per spot weld if you advertise all the costs of the robots and everything else across it. But regarding the battery chemistry and some of the battery innovations, I think you may be aware that we’ve been pioneering, and today we’re the largest manufacturer of battery cells in North America. This is larger than Tesla today. 

We have been pioneering work in lithium manganese-rich, or LMR, battery chemistry which has something closer to the energy density of high nickel batteries, that you’re familiar with, at closer to the cost of lithium iron phosphate, or LFP. Those will come to market first in GM vehicles, in some of our larger vehicles in 2028. What we’ve done is oriented the company towards some of this. The functional strength that Mary references — in common design, common engineering — to innovate on many of the foundational enablers of compelling battery electric platforms is done through the streamlined process that we have for our global product development across the company. 

My expectation is that when this hits, as LMR hits, as our architectures hit, as our manufacturing improvements hit, as some of our robotics work that we can talk about hit, what you’ll see is those immediately kick in, and what’s done initially in one model, in one plant, in one brand, rapidly scales across a massive portfolio of vehicles.

The leverage that we realize by doing this is much, much higher than what we could do if it were an isolated effort. Nothing against isolated efforts; that’s sometimes the thing that you do, you have to do, to really move the needle on something. I’m really impressed by what the company’s been doing since well before I joined, but we’re certainly moving that forward. And we can talk about some of the ways in some of our electric architecture — Mary referenced a little bit about what we’re doing there around software-defined vehicles — but there’s a lot of work to come. The core story is we leverage our scale and we leverage our people best when we do this through a streamlined process that’s part of the company.

MB: Our goal was to get the right people in the right functions. From a software perspective, we brought in a tremendous amount of talent from Silicon Valley, big and small companies, Mag 7, et cetera. With [VP of Battery, Propulsion, and Sustainability] Kurt Kelty, who has worked his whole career in batteries, he came and went, “Oh, my gosh, you’ve got the capability to do not only core R&D.” 

When you have a promising startup, you can take that into one of our R&D centers and really see if it’s going to scale and if it’s manufacturable and to actually being able to work on process improvements, you need all of that to get the cost out of the battery to get EVs profitable. And then there are a lot of learnings that flow back in. From a software perspective, it doesn’t matter what the propulsion system is. And so our strategy has been to get the right and the best talent in and make them part of the team, and so like you said, it’s across the board.

You’re describing really long-term bets. Battery chemistry is one of those things where you start on the battery chemistry journey, and hopefully it pays off and that’s a step change in the car. Autonomy is one of those things. I’m actually very curious, Mary: you described bringing in startups and seeing if they scale — Cruise didn’t work out, but we’re here to talk about hands-free…, or rather eyes-free, Super Cruise, and that bet paid off. 

MB: What I would say about Cruise is there’s a tremendous amount of Cruise talent that’s still here. I remember back in 2016, I made the prediction we would have an autonomous vehicle by 2019. It really was a couple of years later that we had that. You’ve seen that the whole industry — whether you’re a tech company or an OEM — realized it’s a pretty significant challenge. I think we’re closer today, but as we made that change and as we continue to see progress, we looked at it and said, “As we want to deploy capital, where do we want to do that? And do we really want to put all that capital into all the vehicles for rideshare when our business today isn’t rideshare 1.0, it’s not taxis, it’s not rental cars?” 

We made a strategic decision to focus that talent and to have them work closely with engineer and technical talent within General Motors to be able to focus and be in a leading position from personal autonomy, and that’s what we’re focused on now. 

With Sterling coming in, he’s got great experience from his time at Tesla, and then most recently, the only trucking company that has autonomous trucks on the road. That represents somebody who brings in that expertise that is going to help us continue to be on the forefront and lead with these technologies. So from Cruise’s perspective, we did pivot away from robotaxi, but I would say there are a lot of core assets that we’re attributing to our personal autonomy journey from an overall autonomy perspective.

Let me ask you the other Decoder question, and then I want to put all this into practice against the announcements that you guys are making, because it all ties together. 

How do you make decisions? What’s your framework? You’ve obviously made a lot at GM; how do you organize that process?

MB: A lot of the decisions we make are complex. One of the things I like to leverage is the senior leadership team. There are nine people who are core to the company that lead all the functions. Sterling is on that team as our chief product officer, along with a few others. And we look at trends both inside and outside our industry. We look at what’s happening with technology; we’re very focused on where the consumer is. Of course, we benchmark where the competition is, but then we also look and try to say, “Where is the consumer going?” And I think there have been many times we have found white space in even vehicle segments. When we first put the Buick Encore out, critics said, “No one’s going to buy it.” 

That segment now has grown to four vehicles and is profitable and growing, and we can’t build enough of those vehicles. That’s just one example, but it’s really getting the team together and understanding where the consumer is going, where technology is going, and then what the roadmap is we’re going to put forward.

Five, seven years ago, we would put the portfolio together, and usually, it was an annual event. I would say it’s a more frequent event now as we continue to take new learnings and make adjustments. The minute you have new information, you don’t just want momentum to put a program forward; you want to pivot and make the change to make sure you’re going to be relevant going forward. But I would say my decision-making is largely based on looking at it as a team, with different experiences, different skills. Together, we get to better solutions and a better strategy.

The big announcement that I’m really curious about is the Gemini-powered assistant in the car that’s coming in 2026. I would say this is an expansion of the idea that GM should have a software platform in its cars: we’re running Android Automotive in the cars, there’s an app platform, there’s a data platform. There’s Google Assistant in my car; my daughter was asking it to play Taylor Swift. That’s basically all we use that platform for today because she won’t let us do anything else. But it works, it’s there. The big decision there, and I know you know this question is coming, is that you bet against putting smartphone projection in your EVs: there’s no Apple CarPlay or Android Autos in the EVs, but the gas cars still have it. How did you make that decision?

MB: It’s really a question of timing as we look at that, because — and I want to make sure that we get Sterling’s input on this as well — as we looked at it and as we made that decision, we were getting a lot of feedback from customers that it was very clunky moving back. It wasn’t seamless, and frankly, in some cases, it could be distracting to move back and forth if you were doing something that you could do on a phone projection type of system versus if you needed to do something in the vehicle. 

We also know that’s only going to increase when you look at some of the things we’re going to talk about that can make your life better and assist you as we move forward. We’re at the very, very early stages of services we can have on a vehicle to improve the overall customer experience and make the journey smoother.

We looked at that and we decided that we needed to have a great system in the vehicle that allowed people to have one system, and we’re going to continue to make that better and add new features. 

SA: What we’re talking about is the inevitable performance degradation when jumping between S-curves. The first of those S-curves was, for some time, that you and others got attached to phone projection applications largely because the in-vehicle HMI was pretty bad. Your opportunity for doing some of the things was better when you were using that. You’re driving a Vistiq, I understand; you’ve got Dolby surround audio, you’ve got giant screens, you’ve got giant displays. The analog I would use here is we’re on this new S-curve, where there is inevitably a jump that has to happen for you to get over to it. That’s uncomfortable for many. 

But frankly, it’s a very Jobsian approach to things. The removal of the disk drive, nobody liked that, everybody on the forums and Facebook was complaining about it, but to that he said, “Look, guys, flash storage really is the future. Get on board, you’ll see that.” That’s kind of what we’re saying here, in fact that’s exactly what we’re saying.

Say you’re talking to me about CarPlay. You’ve certainly got an iPhone, you’ve probably got a MacBook, and you have the opportunity to use phone projection on your MacBook, a phone mirroring application. How many of you are accessing online services like email, social media, and otherwise through the phone projection app in your laptop? Almost none of them do. Why? Because you’ve got a much larger screen on your laptop, you’ve got a much more convenient HMI via the keyboard, you’ve got better speakers.

Now, take that same analog to the car and ask the same question. Is it in a car that has not only just laptop speakers, not only a laptop screen, but something better that can move you, and that can integrate with charging infrastructure, with Super Cruise availability on your maps, all of these other things? You are in a much more immersive environment that can do so many more things; why would you use the equivalent of a phone mirroring application on a laptop in your car? So we said, “We’re taking out the disk drive, guys; get on board with flash storage, that’s where the future is.”

I have a number of responses for you, Sterling. I spent a lot of time hearing from our readers about this specifically. The first thing I will say is most people don’t use the phone mirroring app on their laptops because it’s not illegal to also use your phone while you’re using your laptop, and it is very illegal to use your phone while you’re driving your car. 

So this is the first difference that most of our readers would point to, that this thing where I need to see my phone, I need access to all of the applications and data on my phone without logging in again or having multiple user profiles, it’s fine with my laptop, it’s fine with my iPad, it’s not fine in the car. And the specific one that sticks in my brain about the library of applications is I’m actually not a huge CarPlay fan, and I complained about it on our other show, and I got this long note from a reader where he said, “I use between six and eight audio applications on my three-hour commute.”

One of them is just this, I believe, very niche Bible app. And he’s like, “They’re never going to support this on these other platforms. It’s on my phone and it has a CarPlay extension.” So that’s one thing — I just want the library of content. Then there are consumers being unable to affect what feels like business dealings at a much higher level. 

For example, yes, my car has Dolby Atmos in it, the number one provider of Atmos tracks in this industry is Apple Music, and Apple Music will not have an app on your phone, because I’m confident that Apple wants you to have CarPlay, and that is a business dealing that the consumer demand cannot affect. That’s kind of the shape of the puzzle, right?

MB: I would say we have a good relationship with Apple. I mean at the most senior level with Apple, with Google, with all of the tech companies. We’re bringing Apple Wallet. We’ll be announcing that shortly, that we’ll have that and have the ability to do some of the vehicle functions through that. So we’re having continual conversations with Apple, and I would say we’re talking about the opportunity and looking for win-wins. We also have a very good relationship with Google and we don’t enable Android Auto either. So I would say you’re talking about a moment in time versus where the industry is heading from Dolby Atmos and the relationship that we have with Apple. I wouldn’t make some of the broad-based assumptions you’re making.

Do you think I’m going to get the Apple Music app in my Cadillac?

SA: We don’t have anything to share on that right now, but your first comment really struck at the HMI, the ease of use, and [whether] you have to log into each of these different services and applications in your car. Because if you do, you get some breakage. Some people just will never do that, it’s a pain. We’re looking at that as well. What can we do about federated IDs? What can we do to eliminate that friction of you engaging with your car? I’m not sure I quite follow the whole “it’s illegal to use your phone when you’re driving and not when you’re on your laptop.” I think that cuts against your argument a little bit because–

Well, if the cops see you looking at your phone while you’re driving, you get pulled over; the cops see you looking at your phone while you’re using a laptop, I don’t think they’re going to do anything about it.

SA: My point being, if you’ve got your phone in your laptop or you’ve got your phone in your car, you’re sitting next to your laptop, you don’t typically pick up your phone to answer an email, if you’ll answer it on your laptop. You’re sitting in your car, you wouldn’t pick your phone to use a map application if it’s sitting there on a bigger screen in front of you in the car. And if you had-

Wait, hold on, not to get too weedsy, but I know we have a lot of readers who have Windows laptops for example, and they are constantly picking up their phone to use iMessage on their iPhones, because that’s not an experience that you can project to Windows in a meaningful way.

SA: And that kind of gets to Mary’s point, which is don’t expect that we won’t have a growing list. Today, we use Google Automotive services in these cars, that have access to the Play Store, and quite a few apps available through that. That will only grow in time, as will availability of apps from other places.

MB: I think, overarchingly, we want to give the consumer a great experience in the car that’s not clunky, that allows them to get to this point of safety that one of your viewers talked about, that you can be the least distracted and pull even more information from the vehicle to improve your whole experience overall. We’re at, I’d say, the early phase of making that shift and continuing to add more to it, continuing to make it more intuitive. How do we make access to more apps and more things you want to leverage? We’re excited about the future that we have coming for what’s going to be available in the system we put into the vehicle.

Let me ask you the second part of that question again, because, again, we’re talking so much about the future, and I understand the argument about the future you’re making, but you still have the smartphone projection in the gas cars. Why is it still in the gas cars?

MB: A lot of it depends on when you do an update to that vehicle. When you look at the fact that we have over 40 models across our portfolio, you don’t just do this and they all update. As we move forward with each new vehicle and major new vehicle launch, I think you’re going to see us consistent on that. We made a decision to prioritize our EV vehicles during this timeframe, and as we go forward, we’ll continue across the portfolio.

So we should expect new gas cars will not have smartphone projection?

MB: As we get to a major rollout, I think that’s the right expectation. Yes.

Let’s talk about that next rollout. I’m very interested in the fact that you’re putting Gemini in the cars. We obviously use Google Assistant in our car today. AI has often been described to me as a platform shift, and in particular, natural-language interface has been described as a major kind of platform shift, and then maybe the way we write applications will change with MCP and these other technologies. But the platform shift is you’re going to talk to the computer, it’s going to understand, it’s going to take some action for you. Are you feeling that way about the car, that AI and particularly these assistants will let you have a platform shift inside the car?

SA: Yeah, we are. We’re thinking about AI across the business, not just in what it does to the product experience, but what it does to our development of the product, what we can do to bring to bear the massive data that this company has in our production systems, in our CAD and development systems, to leverage in development of AI for production for the product. As it relates to the product, though, what I’ll say is this is one of the really important enablers of speed in a company with a portfolio this broad, is development of a common platform or undergirding infrastructure on which we can deploy, including over-the-air software updates. As Mary mentioned, there are some legacy platforms that will work their way out of the portfolio, but one of the things we’re working on in the software-defined vehicle space is a new electrical architecture that centralizes all compute in the vehicle. It doesn’t just move the compute to zonal aggregators; it centralizes all of it. In the upper trims, it leads to about a 35 times increase in the computational power of these-

This is the 2028 Escalade IQ, right? It’s going to be centralized with massive amounts of compute?

SA: That’s right. What it does is, from a networking’s perspective, move to Ethernet-based networking, which allows us to move to sub-millisecond response times. Take magnetic suspension systems or dampers, as your example, from accelerometer through controller back to actuator. We’re talking sub-one millisecond. You’re talking a thousand hertz. That’s a 10X improvement over previous electrical architectures. It’s a massive, massive opportunity not only for the dynamics, which I found General Motors is extraordinarily good at, but for the software that we can deploy on this. Now we’ve got a centralized architecture, we’ve got a central compute on which we can run a variety of applications to which we can deploy updates for those applications at a much higher pace than what we do today.

And so that flexibility that comes via abstraction of hardware and software with this new electric architecture is really going to be a powerful enabler for us going forward and something that I’m excited about. So your questions about why can I get some things in one car and not another? You’re going to have far fewer of those because you’re going to have a lot more commonality throughout the portfolio based on this platform that separates hardware.

You keep saying HMI — that’s “human machine interface” for people who are not total nerds like us. The idea, though, is that you’ll be able to just say, “Hey, Gemini,” and you’ll just have some natural language conversation with the car. There are a lot of things you could do with that. There are also a lot of questions about how well that might work, even today, as expressed in ChatGPT and Gemini itself. A lot of questions about how those things might fundamentally operate, what kinds of applications we might build, how reliable they’ll be, how many hallucinations we might have. Is that going to be enabled by your new architecture too, or are you going to send all of that to the cloud?

SA: There are a number of things that we’re going to infer directly on the vehicle. As I mentioned, the computational power is there for us to do a lot of on-vehicle inference. What you’re referencing is in the Gemini-based conversational AI work that we’re going to release next year. That’s really about vocal interaction, voice-based interaction with your car; it’s about asking about your destination, asking about a number of other things where it can give you contextual responses. Many elements of the car, if you’re going to directly control them or safety-critical them, how can you deal with the accuracy that you get out of an untailored LLM if you’re giving direct access to those things? And the short answer is that we’re not.

We are developing, in addition to the contextual AI that comes out next year, a more powerful, specifically tailored AI that will learn from not just your preferences and not just the data that you have available via a variety of other mechanisms, but it will be specifically tailored, customized, and learn from the vehicle itself. You’ll have access to specific functions that will have other controls in place that a generic LLM doesn’t have. So I think that the value of these large language models oftentimes can be realized in much greater effect through tailoring and fine-tuning for specific application areas, and that’s what we’re going into after the Gemini conversational app.

MB: The other thing I would add is that at General Motors, we always prioritize safety. We were the first to have a driver assistance technology out with Super Cruise, but we very much prioritized the safety of that and enabled different features. We saw it demonstrate that it could meet the safety standards that we have. I think that’s what Sterling’s talking about now: as we do this, we won’t just go wide open, but we’ll make sure that the integrity of the safety systems are there.

SA: If it’s controlling a temporally sensitive or safety relevant function, it will happen locally.

One of the reasons I’m asking about that very specifically is I’m watching your partner, Google, I’m watching Amazon, I’m watching Apple — they control a lot of that stack. They control the LLMs, they control the inference, they control the smart home platforms, and they have not been able to connect those assistants in the way that you’re describing. 

Alexa Plus exists and it has to fall back on some complicated deterministic system to flip lights up and down, and they have to do some orchestration. Google Home, they’re doing the same thing over there. Gemini is rolling out and it has to fall back and it’s working however it works. Apple has not been able to ship this product, which they had promised people. 

If I ask my assistant here to turn off the lights and it doesn’t happen, maybe the worst outcome is my wife is once again annoyed at me for trying to voice activate our home. If I get that wrong in the car, many other things can go wrong.

SA: That’s right, 100 percent. This is something that we are very keen on. On the development of our self-driving system, we are acutely familiar with and aware of the challenges that arise when you have a monolithic end-to-end system about which you can’t make any strong guarantees to other controls across the car that are safety relevant. That’s probably a much longer conversation than we have time for here, but this becomes arguably relevant in the design of our self-driving system, it’s architecturally relevant in the design of the rest of our AI and how it interfaces with vehicle controls, and we are taking that very carefully.

MB: And we have a safety group within the company that really looks at that and is looking at all those decisions that are made, because they can think through second-order, third. Down the road, we’re really thinking through to make sure that we’ve interrogated it, to make sure we’re making good decisions and we’re not going to put the driver or the passengers into a situation that no one wants to be in.

This is my other question: Why pick Gemini? When you evaluate all the universal partners, obviously, ChatGPT by OpenAI has a huge consumer market share; Google obviously is very capable. You already have the partnership. Was it as simple as “we already have the partnership” or did you do the full evaluation?

MB: I know the team did the full evaluation, and as we looked at what we needed to do, clearly, they’ve been a good partner to us, but it was a full evaluation.

When you think about the long-term future — you’re also announcing a robotics division, you’re announcing batteries in the home, you’re announcing eyes-free driving, what, Level 3 autonomy in the ’28 Escalade — that’s a big shape of how we will interact with cars. You’re going to get to the house, your car will be part of your energy system in your house, you’ll have robotics that will let you take the technologies you’re building for autonomy and AI in the car and put them in your house. That’s a pretty tight integration of how people feel about their cars today and how people feel about the things in their home, and somewhere in there is the next generation of devices.

This is the big question that’s animating the tech industry around me: how are we going to graduate from our phones? You’ve obviously made a big bet — “We’re going to graduate from our phones in a car, and the car is going to be a big part of that platform.” But then you got to go in the house and there’s going to be a GM robot, there’s going to be a GM power supply. How are you thinking about that extension as the rest of the industry grapples with what happens after the phone?

MB: We already have quite an extensive offering from a GM energy perspective [such as] the ability of your vehicle to actually power your home in power outages. So that’s something we’re continuing to work on, to make that vehicle work even harder for the owner. Beyond that, when we talk about robotics, that’s a broad term, and I’m sitting here speaking with someone who has his PhD in robotics, so I probably should turn it over to him. 

But when we look at what we’re doing on the factory floor with a cobot that is able to take over and do things that create safety issues for our team members, or create ergonomics issues, or get rid of non-value-added activities that make everyone more efficient, that learning in what we’re doing is something that we can look at for broader application. When we look at it from that perspective, and we already look at what we’re doing from a GM energy perspective for the home, there’s a lot of potential as we move forward. And Sterling, what would you like to add?

SA: The only thing I’d add is we are focused on growth areas that are both tangential to and synergistic with our core business. When you say batteries and energy storage, we’ve recently announced a partnership with Redwood Materials, for instance, where our second-live batteries are currently powering the largest second-use energy storage system in the world, the largest microgrid in North America, in Sparks, Nevada. It’s 63 megawatt hours, I think I remember that being. So there’s tremendous opportunity there. 

When we talk about grid rebalancing, particularly as we move into LMR, which has much higher cycle life, we get into a world where by plugging your car in, when it’s at home or your place of work, your car can be an active participant in the rebalancing of the grid, pulling power from it during lulls and pushing power to it at peak times. This is an enormous opportunity for the industry. Energy is going to be a very constrained resource if the last several months of discussion in the AI space are any indication. I mean, we’re talking about triple the capacity needed online.

We can play a pretty big role in that. As the largest cell manufacturer in the United States, there’s a lot we can be doing there. In robotics — I’ll just paint the picture for a moment — particularly in a world of learned models, which is effectively where AI is taking us, data is really important. The quantity of data, the cleanliness of that data, the availability of that data, in training systems that can take on more tasks than what they’ve ever done before. 

Look at a business like General Motors and you see a product, each of whose parts is known in detail: it’s sitting in CAD, we know the mass, we know the materials, we know the response to manipulation, we know the flows of that material through a production system, we know what happens at each step of the production system. Every single step is meticulously outlined in work instructions. That is an enormous database from which to develop really competent physical AI systems, embodied AI systems that can be powerful enablers, both of our own production and the safety of our employees.

So today, we got 30,000 robots operating alongside on the order of 97,000 production associates in 11 facilities. You think of automotive robots and you’re probably thinking of giant arms that sit in cages with e-stops on the outside that can’t operate in and around humans because they’re simply not designed for it. We’re developing autonomous mobile robots that move materials through our factories. We’ve started deploying those. We’re developing cobots, or collaborative robotic systems, that can operate alongside associates, that are currently in our factories and are scaling up, that can, for instance, manipulate and bring apart to the line to hand it to a worker, or hold it in place for that worker to shoot the bolts. 

So there’s tremendous opportunity for us to build that not only for our own use and improve the safety, the efficiency, the throughput of our own business, but to expand that to our supply base, to increase our competitive mode, and then ultimately, to productize it, commercialize it for others in the kind of industrial automation.

At least in my head, it is a small jump from, “we have an AI-enabled robot that can bring a factory worker bolts” to “I have a robot at my house that can bring me a beer.” The kind of manipulation you’re describing actually has a pretty massive general purpose application. Are you thinking that far ahead?

MB: I would say we think of a lot of things. One of the things at General Motors, we’re constantly innovating. As I mentioned, we were just in design and looking at what the future holds for what we can bring forward. I think we continue to imagine what the applications can be, but we’re not making any commitments or announcements today. But we recognize, as Sterling said, the amount of information that we have and the very complex work that is done today in our factories and what we can do to better support our team members. There’s a whole new world that is in front of us, an opportunity for us as we move forward.

You’ve described a lot of things that a lot of other companies are doing. Obviously, Tesla does a lot of these same things: they’ve pushed really far ahead with what they refer to as Full Self-Driving, and they’re demoing Optimus — however, whether Optimus works or does not work, that’s unknown. The difference in culture is stark. I can just identify the difference in culture: they are less worried about safety, Elon Musk is less worried about the safety of Full Self-Driving than GM is about Super Cruise on the highways. It is just obvious. They’re less worried about making promises that they might not be able to keep with Optimus.

In many ways, this is to GM’s credit. I think you hope more car companies are worried about safety, but in many ways, it has contributed to the perception that Tesla is vastly ahead. Do you think about that balance? And Sterling, I’m curious since you’ve worked at both places, if I had to just identify the cultural difference, would that be it?

MB: I’ve been at this company for over 40 years and it’s existed for well over 100, and I would say safety is an overriding priority: safety of our workforce, not just for injury, but also preventative, from an ergonomic perspective. We are committed and we’re very proud that we have an industry-leading record, and it translates into safety of our vehicles. And I probably get three or four letters a week from someone sending me a picture of them… and they’ll send me a picture of the vehicle that is somewhat unrecognizable, and I can tell you many of them, they walked away with just a scratch. I’ve had letters sent to me from people in a hospital bed thanking me and saying, “I will be buying a General Motors vehicle for the rest of my life because I know what you do and design into safety.”

We have the highest loyalty scores from a manufacturer perspective or from a car company perspective for our consumers. Safety is not just a cultural perspective; it is a commitment to our customer. Now, as we look and we go forward, how do we deliver that safety and how do we do it more efficiently? We’re always looking for ways that we can be better, but I think it’s a very important part of the promise. And if you read half of the letters that I get from our consumers and you look and see what they walked away from, I think it would be something that motivates you every day to continue to deliver on that promise.

Do you ever perceive that it’s holding you back?

MB: No, I perceive it as serving the customer. And so we’ll look at decisions we’re making, but also, we all have a regulator with NHTSA. We look at that and we are very forthright and transparent because we want them to know what we’re working on, what we’re looking at. And again, they’re our regulator, but I think having that so they understand how we look at safety for the long-term is going to pay dividends, and I think it’s going to pay dividends for the consumer. So I don’t think it holds us back; I think it keeps us focused on what’s really important. Time will tell, but again, I’m proud of our safety record, and I’m proud of what our customers say. And Sterling, you’ve seen it from a different perspective, so please.

SA: The stereotypical view that I think has been expressed of traditional automakers super focused on safety might suggest we are sitting still and just putting out vehicles with great airbags, seat belts, and crumple zones, and we’re good with that. I think the reality is far different for GM. Our customers have driven over 700 million hands-free miles with Super Cruise without a single accident attributed to the technology. I led Autopilot, and you can’t say that for Autopilot. I think this is the long-term play: we build trust with customers by delivering safe products. In the autonomy space, look at every company that’s had a major incident that could be attributed to that company; how did it go for them afterwards? They might have gotten there quicker, but it didn’t go well once the public lost trust. We have earned that, we will retain it, and we’re going to build on it.

And so the work that we’re doing — whether it be autonomy, be it robotics, be it a number of other areas — this caricature of a traditional OEM doesn’t fit. This caricature of bubble-wrapped vehicles that aren’t innovating, it just doesn’t work. The number of bets that we’re making… that Mary’s made, I mean. Years ago, when I was in the autonomy space, I watched her make the investment in Cruise. I watched her lead that to a place where Cruise was the first to deploy commercially in a densely urban setting. I think the reality of what GM has done simply doesn’t match the caricature.

All right, I have one very, very small feature request, but it’s very important to me. The buttons that control the seat memory on the door in my Cadillac Vistiq do not also switch the user profiles on the infotainment system.

So when I get in the car, I need to push the button to move my seat to my seat memory and switch the profile from my wife’s profile to my profile. I have no idea why I have to do both. I just want to be able to push the seat memory button and have that switch the profile as well. Can you fix that for me?

SA: I’m going to take a quick look, that can’t be hard. Let me take a quick look.

[Laughs] If you read the forums, I will tell you, that is the basic shape of the response to this problem.

SA: These are different systems, but let me take a look.

MB: And also, what would it take to do it? We’ll look at both.

I love it.

We are open to customer feedback and we love the fact that you’re driving a Vistiq.

Oh you will come to regret that, as I have more and more feature requests the next time you come on the show.

Mary, Sterling, thanks so much for being on Decoder. I’m excited to have you back again soon.

MB: Very good.  

SA: Thanks Nilay.

_{Questions or comments about this episode? Hit us up at decoder@theverge.com. We really do read every email!}