--:-- Host: Disruption is a word often used in the technology industry, and it might seem that the path from disruptive technology to market dominance is a fairly simple one. And yet the nature of disruption and how different innovators and companies harness disruptive technology hardly goes in a straight line. In this episode from September 2018, Benedict Evans and Steven Sinofsky use the example of Tesla and electric vehicles to dig into the nuances of what disruption looks like in practice in the tech industry and beyond.
--:-- [Music fades in and out]
--:-- Host: Hi, and welcome to the a16z podcast. In another of our hallway conversation episodes, Benedict Evans and Steven Sinofsky talk about Tesla and more broadly the nature of disruption overall. How disruptive is Tesla really? What exactly are they disrupting from the dashboard to car makers to vendors to energy source to autonomy overall? And how much does each innovation matter? Looking back at other examples historically, what kind of disruption matters most for market dominance?
--:-- Steven Sinofsky: Good morning, I'm Steven Sinofsky.
--:-- Benedict Evans: I'm Benedict Evans.
--:-- Steven Sinofsky: What we thought we would talk about today is a little bit about the nature of disruption and in particular about Tesla and the rise of electric vehicles and sort of how disruptive is Tesla? We're in an interesting time, like certainly economically, if you look at the main US car companies, Ford, Fiat Chrysler, GM, all of them are worth less than Tesla in market cap individually. Certainly, if you look at the past year, Tesla has been worth twice as much as them at some point. So crazy, interesting world.
--:-- Benedict Evans: But we use "disruption" a lot in Silicon Valley and in technology in particular. What's interesting about looking at Tesla is it may or may not really be disruption in the way the book is written, which is sort of, "Oh my god, there's a whole new thing and then the old thing just goes away and it's all new players."
--:-- Steven Sinofsky: I thought it was interesting to look at how people think about Tesla here. On the one hand, you have the narrative, "Oh my god, they're doing this stuff that the car companies can't do." On the other, you have the narrative, "Oh my god, they're making the cars in a tent and bits fall off when you drive down the highway. Ha ha ha, it'll never work." This reminded me a lot of hearing similar conversations around, for example, the iPhone. "This is a terrible phone. Will add touch really easily. No one will ever buy a phone for that much money."
--:-- Benedict Evans: Sorry, I had to toss that in there.
--:-- Steven Sinofsky: Exactly. On the one hand, people dismiss in both directions. They dismiss the new thing because it can't do what the old stuff does very well and don't realize that you might be able to learn that. But on the other hand, you also dismiss the difficulty of the old stuff and dismiss what the barriers to entry might actually be.
--:-- Benedict Evans: I wanted to pull apart, well, what are the different things that are happening here? Are they disruptive? Are they not? Are there barriers to entry? Which bits are there barriers to entry and barriers to entry to who? One of the historical comparisons I use, and Steven has sort of scars on his back around this, is if you look at what Apple did in the PC industry, Apple contributed to creating the PC. We had one of the first popular PCs. But Apple did not win PCs. So you can talk about whether Apple disrupted IBM, but Apple didn't actually get the benefit from that. In fact, PC companies didn't really get the benefit from that. PC companies became low-margin commodity companies, and the people who got the benefit were Microsoft and Intel.
--:-- Steven Sinofsky: What's so interesting about that disruption is, when can companies turn some technical innovation into a competitive advantage? And when is a company's go-to-market or technical innovation itself become a hindrance to adoption? Our tech industry is littered with examples of the innovator failing to become the dominator, so to speak. My favorite one is Replay TV. It gave us the DVR and then gave us TiVo, and now we all just have DVRs everywhere. Certainly, the mainframe and the mini computer led to the PC, and the revenue numbers all came and dwarfed it. And then, what you know so well in the phone industry is very similar.
--:-- Benedict Evans: So there is a question of, is this new thing fundamentally difficult for the incumbents to do? But also, is it important? There are sort of four things that I talked about in the blog post. The first is that Tesla kind of has to learn the old stuff. Tesla has to learn how to make cars at scale. There was a period when people said, "Oh my god, they're reinventing manufacturing." Actually, no, they just bought a second-hand robot factory. Tesla has to work out how to do cars.
--:-- Steven Sinofsky: It's almost worth a pause there just to remind people that other people's jobs are really much harder than you think they are. We tend to, even in the software hardware world, I've yet to ever meet a company that makes hardware that thinks software is really, really hard. And I've never met a hardware company or a software company that thinks they can't just go do hardware and buy it in China. This notion that to innovate in cars, you need to understand cars more and manufacturing. And to innovate in the software in cars, you need to understand software more if you're living in Detroit.
--:-- Benedict Evans: This is the thing that people in software don't really know enough about cars and people in cars don't really know enough about software. But to the point on Tesla, clearly there's this whole conversation now about the production hell and they're making cars in tents and the panel gaps are terrible and they're having fires in the paint line. Tesla has to learn how to do what Detroit already knows how to do and what Japan already knows how to do, which is to make cars reliably and efficiently. Or at least Germany and Japan. Well, Detroit doesn't make cars anymore, but make cars reliably and efficiently at scale. That's just a condition of entry. Tesla gets through production hell. That doesn't get them victory. That just gets them to continue, that just keeps them in the game.
--:-- Steven Sinofsky: What they also have to do is be doing something that the existing car industries can't do or will struggle to do for deep structural reasons. They won't just be able to hire engineers and just add that. That stuff has to be in some way fundamentally important. It has to be a profound reason why you would buy a car. And it also has to be something that other tech companies will struggle to do, which is to the Apple versus Microsoft point or the Apple versus Dell point. Apple did stuff that IBM, for the sake of argument, found it hard to do, but Dell did it better in partnership with Microsoft and Intel. Equally, HTC were the first people to make Android smartphones, but HTC turned out not to have the right positioning in the marketplace to take all the rewards from that.
--:-- Benedict Evans: You can look at the beautiful product and you have to unpick, okay, how are they going to make millions of them? What is it in that that is difficult as opposed to easy for other people to do? Which of those things are fundamentally important? And which of those things also will not just BMW – you can't just say, "Well, BMW isn't going to be able to make software." You also have to say, "Well, BMW isn't going to be able to buy those from some combination of Huawei and Shenzhen and Google" in order to get you to, okay, Tesla is going to be the new BMW or the new GM.
--:-- Steven Sinofsky: That's a very important point. Instead of looking at this as disruption, another way to look at this is to use an old phrase that existed before disruption and just refer to it as a secular shift. This is a shift. We're going to all be in electric cars and electric vehicles and electric transportation at some point, which is very different than it's disruptive because disruptive tends to focus on the micro, like one company versus another company. Whereas if everything is going to move to this, it's not clear that it just means that only the companies that are currently doing something are going to benefit. Going back to the DVR example, it turns out DVRs are like a commodity now. Everything that can receive video has the capability of just being a DVR.
--:-- Benedict Evans: We should dig through what those separate components are. The analogy that I use, what I try to do is break it apart. There is the electric itself, which is the battery and the motors and the powertrain and the controlling software for that. Not exactly the most revolutionary technology. Lithium-ion batteries are not something that got invented by Tesla and Panasonic five years ago. There's that. Then there is the sort of one level up, all the integration of the control systems around the car. And then there is the dashboard on the car and the experience, the broader experience of buying a car. Do you go through a dealer? Are there charging stations everywhere? Do you have lots of fiddly little buttons or just one beautiful touchscreen? We look at those and think, well, how are the dynamics of each of those going to play out and how hard are they for new people to play in?
--:-- Steven Sinofsky: If we start with electric, the analogy I thought was interesting here was to look at multi-touch. Apple was actually not the first company to sell a mobile phone with a capacitive multi-touch screen. I think there was an LG one a year earlier and maybe a couple of others. But Apple was the company that said, "Oh my god, we can actually use this to totally change what it is to be a phone."
--:-- Benedict Evans: It's an important point. They weren't first at using technology, but they were the first to integrate the technology and pull it all together.
--:-- Steven Sinofsky: As we go through and discuss each of these, it's worth saying that I'm fairly optimistic on the prospects of being able to solve this equation. Others are going to be fairly pessimistic. This is really about analyzing that conversation, not debating the winner.
--:-- Benedict Evans: Exactly. So you have this insight into a new piece of technology. "Okay, we could use this to make a phone. Lithium-ion batteries are going to get cheap enough that you could use them to make a car." This is the foundational insight of Tesla. But if you go into a store today, there are 1,000 phones with capacitive multi-touch screens. So clearly, just using a multi-touch screen of itself didn't get you anything because everyone could buy those. Even BlackBerry was selling phones with capacitive multi-touch screens. Within that, we split that out. On the one hand, the legacy companies, Nokia, BlackBerry, Palm, struggled to make a phone with a good capacitive multi-touch screen. On the other hand, in partnership with Google, Samsung and a lot of other people found it really easy to make phones with capacitive multi-touch screens. Today, the entire industry makes these things.
--:-- Steven Sinofsky: This is also just like, all of a sudden, everybody adds a notch to their phone. Something that appears like, "Whoa, that's going to be super tricky." All of a sudden, a supply chain appears, other people with expertise appear, and you have a lot of innovators building the same thing.
--:-- Benedict Evans: This is the thing. If you look at the way the PC industry works, the way the mobile phone industry works, indeed the way the car industry works, it's not that there is one company that has to work out how to make this thing. It's not that Bosch is going to have to learn electric. It's that you have a whole ecosystem of hundreds of different companies, hundreds of very big companies full of good engineers, who have to work out how to make this thing. Many of whom have been making batteries and electric motors for a long time already, just not quite the same kind. As you look at electric, it seems pretty clear to me that on a five or ten year view – and bear in mind cars are on a five to ten year replacement cycle, so it doesn't have to happen that quickly – the entire car supply chain will have reoriented around electric. Even more than that, the entire electronics industry that already does electric stuff will reorient around making components for electric cars. If you look at the teardowns of, say, a Chevy Volt, an awful lot of the value in that comes from people that were not traditionally car manufacturers or car component suppliers. They're not the traditional two ones. It's all LG.
--:-- Steven Sinofsky: What's super interesting about that too is that the expertise at existing car companies is in acquiring those technologies, building them out, establishing those relationships, negotiating the contracts, and getting all of that to happen. For even GM, that makes the Bolt and the Volt, it's not like there's this massive lithium-ion group at the company. So you sort of think, okay, the car companies are going to be able to go out and buy these components just the way they buy their existing components. There's not some fundamental intellectual property here. There's also no disruption story. It's not like they're sitting thinking, "Oh, this is a terrible idea and we don't understand this." It's an integration into their existing manufacturing process.
--:-- Benedict Evans: On the other hand, if you're a German company that makes gearboxes for the car industry, you're not going to be able to switch to making lithium-ion batteries. It's a totally different business. Your gearbox business is either going to disappear or you're going to shift to marine engines.
--:-- Steven Sinofsky: And if you're an auto supply store on the corner, these cars don't need parts anymore. That's a thing to go short right now if you're in the business of speculating about timelines.
--:-- Benedict Evans: If you're in the business of making radiators for cars, that business is going to go away and you're probably not going to replace that with a business making electric batteries or power control systems. That will go. That's not even disruption. That's just your whole industry disappears.
--:-- Steven Sinofsky: That's the secular shift. It's like all of a sudden horses are now centered around a different set of technologies in different places that you use them.
--:-- Benedict Evans: But that's a different layer in the stack to the car manufacturers. The analogy I used in my blog post was that, for the sake of argument, the Internet was radically disruptive to travel agents, but not disruptive at all to airline companies. Airlines still sell tickets. They sell them differently to different people, but they still run planes. That hasn't changed their business that much. For cars, an interesting view of this is the way that the change in the focus on safety permeated the car industry. There was an era in the 60s when nobody worked on safety. Then one manufacturer, particularly Volvo, picked up on safety. Then the Germans picked up on safety. You partially do regulations in Europe and things like that moving faster. And, "Oh my god, the American companies are not going to be able to have anti-lock brakes. They're not going to be able to have airbags. They're not going to be able to have all these things." It turns out, now there are dozens of companies that contribute to that supply chain, and it's just part of every car. You can't even differentiate on safety anymore because they're able to build that up.
--:-- Steven Sinofsky: That's a car version of multi-touch.
--:-- Benedict Evans: There's an interesting question in this, which is there is a new thing as opposed to, are there fundamental structural reasons why you're going to struggle to adjust to this? If you look at what happened with phones, Nokia was totally oriented around what the mobile operators wanted. They were totally oriented around optimization of component cost, around having a huge supply of building blocks that they could use to make 50 or 60 different phones every year.
--:-- Steven Sinofsky: Hundreds of models.
--:-- Benedict Evans: Hundreds of models. It's mind-boggling how many models they had. Their whole structure was deeply challenged by what the iPhone proposed. The iPhone proposes, okay, one phone, totally different components. Presume it lasts a day instead of two weeks. Presume it doesn't matter if it drops when you break it. Presume it doesn't care at all about bandwidth consumption or memory. Presume you're basically indifferent to the component cost because you're selling it for $600 instead of $150.
--:-- Steven Sinofsky: There were people at Nokia who said you will never be able to sell a phone for more than $150. People at Microsoft would say that too.
--:-- Benedict Evans: Exactly. This is more like the shift from ocean liners to aircraft. The difference would be, on the one hand, the shift from ocean liners to aircraft. On the other hand, the shift from propellers to jets. The shift from propellers to jets is basically all the same companies. The shift from ocean liners to aircraft, it's not the same companies. Although Cunard actually bought an airline company in the 50s because they thought maybe that was what they could do, but of course it didn't work.
--:-- Steven Sinofsky: That's also like pistons to jets and Rolls-Royce still being a leading jet manufacturer.
--:-- Benedict Evans: So you have that question of, is this some fundamental thing that they don't know how to do? Or is it they just haven't done it yet? I think this is a crucial misunderstanding people make, which is they've done it first. No one else is doing it today. Why is that? Is that because there's some fundamental structural reason they can't do it? Or is it because batteries are still $200 per kilowatt hour, and that's not cost competitive with gasoline, which incidentally is also a big reason why Tesla is still losing money? And they're waiting for the battery prices to come down, and then they're going to do it more slowly.
--:-- Steven Sinofsky: This is where it becomes very emotional about disruption. Either you're just a very big bear on how car companies behave, and that they're entrenched bureaucracies. It is important to put car companies in some broader context. These are 100-year-old companies that have survived many different waves of technology and many different changes. In all fairness to them, they invented modern management. Everything that's interesting about management sort of came out of GM.
--:-- Benedict Evans: Came out of GM.
--:-- Steven Sinofsky: As an aside, everybody listening to this, please go read "My Years at GM" by Alfred Sloan. It's an amazing book because you're going to recognize many things in that book that companies do today about how they manage brands, how they deal with distribution and networks and manufacturing and all of that kind of stuff, even if some of it appears dated because there are labor unions and things like that.
--:-- Benedict Evans: The interesting break point when one looks at cars is you kind of go up a level from the electric. A great vignette of this is Tesla discovered the Model 3 had a problem with the brakes. They pushed down an over-the-air firmware update that fixes the problem with the brakes. If you look at the way a conventional car you would buy today is put together, there are dozens and dozens of separate subsystems in there, all of which come from separate vendors. The ABS is a system. The backup camera is a system. The airbags are a system. They all come from separate vendors. They're all integrated, as we were saying earlier, by the car manufacturer. They want all of those systems to be commodities so they can get the best price on them. The only place, if they have a user interface – obviously some of them have no user interface – if they have a user interface, that manifests as a button on the dashboard. There's this old joke that you can see the org chart of a car company in the dashboard, and you can see that the HVAC people hate the steering wheel people or something. You have an org structure that's set up to deal with these parts as components and not integrate them at all. You want them to be not integrated, because then you can just swap out Bosch for Lucas and it doesn't matter.
--:-- Steven Sinofsky: You then look at the way Tesla has built their car, and it's one central computer running an operating system as opposed to a real-time operating system running like a real operating system on some Linux fork or something. The way this has been described is you go from basically complex cars with very simple software to actually very simple hardware but with complex software. You actually have a computer controlling the car.
--:-- Benedict Evans: Which is also an analogy of what happens with feature phones. There's the camera and the phone app and the SMS and they're not integrated except on the screen. Then you go to a smartphone where suddenly you've got a piece of software that's controlling all of these. That's very similar. The PC industry actually had this exact... The reason that none of the PC makers, other than Apple in a sense, are successful phone makers is because they were exactly like Detroit. They had a graphics group, they had a peripheral group, they had an IO group, they had a storage group.
--:-- Steven Sinofsky: I used to describe Dell as being a very specialized version of FedEx. They buy the parts and kind of put them together on the way to getting them to you. They're a logistics business as much as they are a technology business.
--:-- Benedict Evans: I'm slightly upset because there's lots of cool technology inside Dell as well. But basically, they're an assembly business, which is also what Detroit was and what Apple isn't. If you literally look at the headcount of those companies, the number of mechanical engineers relative to the number of supply chain managers, procurement people, and QA people, it's sort of out of whack relative to what you would see at Apple.
--:-- Steven Sinofsky: This gets you to an interesting point. We're kind of setting aside electric. It seems pretty clear electric is a commodity over time. You go up one level. This stuff is stuff that's a bit institutionally harder for car companies to think about because they've got a whole org chart that says, "I've got an ABS man and I've got a backup camera person and I've got a brake light person." Tesla doesn't have any backup light person. Tesla has a software team.
--:-- Benedict Evans: Speaking as a manager, this is a very real thing. You're building your new electric car at a big existing car company and it's going to have ABS brakes in it. So you're going to go to your brake expert. You're not going to go to the software team and say, "Make me some brakes." It's not that the software team makes a device driver for the brakes. It's that the brake people give you the brakes. The brake people are the brake people. They're going to look at this problem and go, "First job, go to Bosch and get the brakes that I'm going to use for this."
--:-- Steven Sinofsky: Some of this actually manifests itself in my Chevy Volt. It's very clear that they went, for climate, for the heater and the AC, they went to the existing heater and AC people and said, "I need a heater and an AC." One of the things that's super weird is it's not really integrated with the battery powertrain that's in the car, and it's basically an old-school kind of heater. The same with the dashboard. The Chevy Bolt dashboard looks like a GM dashboard from all the other cars.
--:-- Benedict Evans: You see this through the whole experience. That's an interesting locus for disruption because it's a lot easier to argue that this is difficult for car companies to adjust to than it is for electric per se. You see that manifest in things like the Tesla software update and in, to some extent, Tesla's ability to add new capabilities or new features to the car over the air, never mind autonomy, which we may come to later. You can do this, you can do that, it can do this cool thing, and you can just decide to add it. As opposed to your conventional car, I've got a BMW 3 Series, it will get new features. When I buy a new BMW, I will not get new features.
--:-- Steven Sinofsky: You don't expect them, you don't want them.
--:-- Benedict Evans: But that really does speak to it. It is quite conceivable that there is a bunch of stuff that Tesla as a company is going to do that is in fact very, very difficult for car companies to do. The question, I think, is how this gets to one of the four things we talked about earlier. How much does that change the competitive advantage of the car?
--:-- Steven Sinofsky: I think we had a conversation about this a few weeks ago where you were comparing this with laptops and phones. Does Tesla's approach produce a better car or does it produce a different car? This notion of better and different, this is why I think it's such an emotional debate for people when they sit in a Tesla versus sitting in a Chevy Bolt or sitting in a gas combustion car. The Tesla experience, it's a very different kind of car.
--:-- Benedict Evans: When people talk about the Tesla experience, a lot of what they're actually saying is stuff that's generic to electric. They say, "Oh my god, have you felt the acceleration?" You do understand in 10 years the crappiest GM car you can buy will have the same acceleration because that's just electric.
--:-- Steven Sinofsky: I came from a Prius to my Bolt and the thing that I'm like, "Oh my god, this is the fastest car I've ever owned." This is very similar. This is one of the things that happened with the original tablet PC that we made at Microsoft, which was around 2000. All of a sudden, we did these internal surveys. "Do you love your new tablet PC?" We had gotten like 100 units and deployed them for a test. Everybody was like, "This is the greatest PC I ever owned." All of a sudden we're like, "Oh my god, we're onto something. It's really big." Then we dug into the research a little bit and we realized that these new tablet PCs that we had just made, these one-offs, were actually made to be super, super good PCs. They actually weighed like three pounds and they were super thin, and they had really great screens on them because the pen screen was made really well. We realized nobody was actually using ink at all. They just loved the fact that, compared to their 7.5-pound ThinkPad, it was the smallest, lightest laptop they'd ever used.
--:-- Benedict Evans: I think the thing about the Tesla integration is, the first point is, a big part of the experience is the acceleration. All cars will have that. That's not Tesla, that's just electric. All cars will have no maintenance. They won't have oil changes. They will be quiet. This is just electric. This is not Tesla. In all fairness, I think most of them are going to end up with a similar miles per kilowatt hour kind of range because the physics is sort of shared. There's not all this leakage that you might experience with choices you make in horsepower and gas combustion engines. There'll be a variance of course, but even if you look today, they're really pretty clustered around the same sets of measurements.
--:-- Steven Sinofsky: So the electric stuff is a commodity. Then when you get to the integration of these components, you can argue it's going to be a lot more difficult for legacy car companies to do this. Structurally, by the org chart. They actually have actual reasons why it's difficult for them to do this.
--:-- Benedict Evans: What's not quite so clear to me is whether that translates into a reason why you would or wouldn't buy the device, why you wouldn't or wouldn't buy the car. I think the analogy, what you were talking about a couple of weeks ago, was the difference between an Apple laptop where there's no choice of any of the components and it's super, super optimized and the case is made out of machined aluminum to fit each component. The laptop is really, really thin and has really good battery life and so on, as opposed to a Dell laptop where you've got a choice of 45 different components and you can swap and you can have this or you can have that. That means there's more empty space inside the laptop because they've got room for the bits you didn't choose and it's got to fit four different components. It might have 10% worse battery life because it's not super optimized. The difference is, you also have the choice of all the components. I think that's the Tesla versus GM conversation. Is it super, super optimized and hyper designed around one specific configuration?
--:-- Benedict Evans: Or is it, okay, we're running five different cars on this line and we'll mix and match and we'll get this and we'll get that. I mean, an example I saw: Mercedes have just announced an electric SUV, and they're putting the electric motors in the front under the hood instead of down on the chassis level next to the axles, which is what Tesla is doing. If you're only making electric cars, it's better to do it the way Tesla is doing. Of course, Mercedes is making this on a line that's also making the 3 Series and the C-Class. Therefore, it's more efficient if they have some overlap in the mechanical processes. You can argue, well, maybe they'll lose 5% battery life by doing it like that, or the weight distribution won't be quite so good. On the other hand, they might save 10% on the cost of making the thing, which means it's $5,000 cheaper.
--:-- Steven Sinofsky: It's so important to really hammer this point home because this is disruption at a very micro level within an organization. When we were building Surface and ARM PCs at Microsoft, one of the things that happened is we showed up and said, look, when you use ARM chips, the graphics card is right next to the CPU and they're all part of the same thing. You can't buy an ARM chip from one vendor and a graphics chip from another and mix and match them. Most of the people who traditionally make PCs, I was looking across the table from the graphics person and from the CPU person, and they didn't know who... Okay, which is getting fun. They thought literally one of them was going to not get to do their job.
--:-- Then it went another level where they said, well, we actually need to leave room to be able to swap out a new CPU, because if we get a different one in the middle of the production run, we want to save all that upfront engineering cost on the chassis and on the assembly line. And we said, well, the thing is, they all just come soldered to a board at manufacturer time, so you can't switch them. The dimensions of the CPU could just move around a whole bunch. It'll all change. And they literally couldn't—they just weren't interested in making it because they didn't know what their job would be if they weren't optimizing that particular thing.
--:-- More importantly, they didn't know what their job would be if they couldn't change around the parts, because their whole economics of what they were building was based on optimizing the inbound supply chain for switching different things. Plus the tech enthusiast side of it—the purchasers who said, we need to have a bunch of graphics on this device, so we're going to up the graphics level, or we need this device to have longer battery life, so we're going to lower that. And the marketing people who want to have good, better, best for every single PC. They couldn't imagine just having good, better, best be defined like Apple does, by amount of storage or screen size. The Mac all-in-one desktop was always just small, medium, and large. They were basically the same except for the screen size, which turns out to be a very consumer-oriented way.
--:-- What's going to happen with cars is going to be very interesting because it's not just that offering. It's the whole purchase process, supply process, advertising, dealer compensation process, all of these things. There's a whole interesting question around what electric does to the car industry: you can imagine a $15,000 car that does 0-60 in three seconds. And it's super safe by default, and no maintenance?
--:-- Benedict Evans: That said, of course, a Porsche still drives... Well, an electric Porsche will drive an awful lot better than an electric Tesla, because just because it goes fast in a straight line, there's more to being a good car than that.
--:-- Steven Sinofsky: Right, right, right. But it does remove layers. You buy the bigger engine. I have a 7 or 8-year-old BMW that I bought secondhand, and I don't even know what the engine is. You look at the badges on the back and it says, is it the 328 or the 330 or the 335 or the 335i? That will not mean anything. There will be one gearbox. So those differentiations within the car will go away. There will not be different gearboxes.
--:-- Benedict Evans: So one last thing we have to talk about is the big one for the sense of "is software eating transportation": the very long-term vision of where autonomy fits into all of this. If we go back to our four layers—there is the electric, there is the kind of integration we've been musing about, there is the driver experience, and then there is the autonomous part. If you go through those, the electric is a commodity. The integration stuff is a bunch of interesting internal questions in the supply chain of the car manufacturers, but it's not terribly clear that translates into a different car or a car with strong competitive advantage. There's the dashboard experience, and then there's the autonomous part. And then there's the dealers and everything else.
--:-- I think the easiest place to locate disruption is in the dashboard, because all the things we've been talking about about the org chart is really hard for a traditional car company to say, we're not going to have any manual controls in the car except for a few sticks on the dashboard. BMW tried it with the 7 Series 20 years ago. But this is literally none—there were the sticks on the steering wheel and then there's a screen.
--:-- Steven Sinofsky: Yep.
--:-- Benedict Evans: I think there's a bunch of reasons why it's really hard for legacy car companies to do that. The question is, is this like iPhone hard, or is it channel hard, or is it physics hard? And does it make a fundamental difference to people's willingness to buy the car? I did a totally unscientific Twitter poll. My question was: if BMW, Mercedes, and Tesla are all selling a car with exactly the same drivetrain, the same acceleration, the same electric—everything is exactly the same—the only difference is that you have the big screen dashboard as opposed to the Mercedes or BMW dashboard. How excited would we be about Tesla as a company? Would this really be a $50 billion company?
--:-- Steven Sinofsky: Two things on that. One is that you have to factor into that choice all of the negative selling that will happen from car companies without that. They will talk about safety, driver distraction. They will literally go to the government and try to get dashboards like that banned. This is exactly what Detroit has been doing for decades over electric. The phrase "range anxiety" was not dreamed up by the physicists at GM. It was dreamed up by the marketing people selling against electric vehicles. The other half of that is just going to be the fans of gas combustion engines and the fans of existing companies saying, well, if company XYZ that I love doesn't have an all-in-one dash, that means that they're bad.
--:-- But the thing about it is that lining up brands like this—go back and read *My Years at GM*. Detroit mastered the art of selling the same thing to different people at different prices with slightly different design. We had a Camaro type LT. I remember when I was little, because it had this LT right on the door where I would open it, and I always asked my mom what does LT mean. My mom had no idea. She just knew it was a Camaro, which was not the same as a Firebird or as a Trans Am. Even though they all looked very similar—Burt Reynolds sort of drove the same car as we had, but it didn't have an Eagle on it. Mom's like, oh, the Eagle is really expensive. I'm like, for the sticker?
--:-- Benedict Evans: They gave him one.
--:-- Steven Sinofsky: Right, right. That's a different thing.
--:-- Benedict Evans: So, autonomy is the software player. It's not the dashboard. I think what we're getting at is the dashboard is fine. This is not the wealth of nations. This is not the $100 billion change. It's the autonomy. Because electric is a commodity. The integration is a commodity. The dashboard is fine. It's autonomy that is the question. And the disruption conversation in this is: Tesla is not competing against dumb Detroit guys. It's Silicon Valley Software Company versus 20 other Silicon Valley Software Companies, plus China.
--:-- Steven Sinofsky: Plus China.
--:-- Benedict Evans: Plus Silicon Valley Software Company that got bought by the dumb Detroit guys.
--:-- Steven Sinofsky: Sorry, Detroit, but you know what I mean. That's what they said when that happened.
--:-- Benedict Evans: Yeah, who—the Detroit guys or the Silicon Valley guys?
--:-- Steven Sinofsky: Either way.
--:-- Benedict Evans: So the question here is, clearly this is a fundamental, profound new technology. We can have a whole other conversation about how long away. We've done a lot of other podcasts about when will autonomy come, what will happen as a result. Cities and towns will change, and lives will change, and everything changes.
--:-- Steven Sinofsky: Exactly, everything changes.
--:-- Benedict Evans: But who is trying to build this right now? Google, all the big Chinese tech companies, Cruise, several dozen smaller companies trying to build component parts of this, all of which at some point will be available for sale to anybody who gets any of this working. Plus, people who make components all have projects going on. If you make LiDAR, you are affiliated with projects. There's LiDAR companies, mapping companies, radar companies, simulation companies. All the delivery companies. But also, there's people building all of those components. So you've got Waymo building their own LiDAR, building their own camera systems, building their own databases, building their own mapping systems and their own simulation tools. On the other hand, Voyage will buy the simulation tool from Applied Intuition and they'll buy the mapping from DeepMaps. The point is, there is a whole ecosystem that's trying to create autonomy. Within that ecosystem, Tesla is one company trying to build this as well.
--:-- From a disruption conversation, it's not that they're all disruptive or they're all innovative companies. They're all participating in this giant transformation.
--:-- Steven Sinofsky: Exactly. A whole new scenario all at the same time.
--:-- Benedict Evans: So there is no sort of "the new people are doing something that the old people won't want to do." There's 50 new people competing against each other. And depending on where they are, these people are all coming out of the same universities, studying the same kinds of machine learning, and then they're all ending up at these companies and they're all spinning off from those companies and they're all changing jobs. This is a whole community of knowledge that's being built at one time.
--:-- Steven Sinofsky: There's another point in here that has been implicit in several of the previous things we've talked about: you have the sense that there is an entire ecosystem. There was a whole ecosystem making gasoline car engine components and supply chain. Part of the reason that Apple ran into such difficulty in the 80s and 90s was Apple was trying to compete with the entire ecosystem. They weren't just trying to compete with Microsoft. They were also trying to compete with all the people who were selling components. They weren't trying to compete with Dell. They were trying to compete with Microsoft and Dell and the 300 companies.
--:-- Benedict Evans: And Intel.
--:-- Steven Sinofsky: And Intel and the 300 people. And Seagate and everybody who sold stuff to Dell. They were trying to out-compete the entire ecosystem. So you have to really understand what was going on. What this means is that Apple had to make one decision for every component that made a Mac, and that was the only one they could make. When they picked a hard drive, they had to pick the brand of hard drive, write all the firmware, integrate it into the operating system, and do all that. And then they were done. If the industry went a different direction, they would just get left behind. If Dell decided, hey, we've got a better return on that drive versus this side, they could just stop shipping Seagate and start shipping Maxtor.
--:-- It's super important because of the maturity of the industry. You actually needed that flexibility then because you just didn't know where things were going to go. Famously, Apple dragged out their AppleTalk for a very, very long time, even though networking had all moved on to TCP/IP. And then they were on SCSI. And then they were on Firewire. That whole era, they seemed almost like a generation behind. Sometimes they pick Firewire, and it just never made it to the PC ecosystem.
--:-- I hear that when I see Tesla making their own this or making their own that, and you think, okay, set aside the fact that you have a major cash flow problem, and why are you spending money to do this rather than just buying things?
--:-- Benedict Evans: Yeah, don't worry about that.
--:-- Steven Sinofsky: Never mind the cash flow question. Why are you competing? Do you want to compete with the entire ecosystem? Or should you be riding on top of that ecosystem and finding the unique thing that you alone can do? Should you bet that you'll be the only person doing X when there's a whole ecosystem that's trying to do X?
--:-- Benedict Evans: And also, this is where it has a lot of parallels to the early Mac. It's not just that they have to do all that, which is almost insurmountable, but they have to do it at a fairly low volume.
--:-- Steven Sinofsky: Yeah.
--:-- Benedict Evans: And the low volume in many of these things is what makes it really, really difficult. Because then you can't even get the attention of manufacturers to help you, even if they're sort of white-labeled parts of it.
--:-- Steven Sinofsky: Yeah, so this is—I've heard gossip about this in the car supply chain. Of course, a lot of the bits inside of Tesla were bought from the car supply chain. Most of the bits inside of Tesla come from it. They're not making their own. He did not do Henry Ford and go and buy tires starting from rubber.
--:-- Benedict Evans: It's a bit like that. They're not making their own motors to wind the seats backwards and forwards. And the problem is the volume is so low that they can't get the best deals from the best manufacturers, which is part of that whole ecosystem question. So we come back to the autonomy question again.
--:-- Let me think about another way of putting this. The bull case here would be: Tesla is competing with car companies that are doing software, so they'll win. And in autonomy, they're competing with software companies that are doing cars, so they'll win. The bear case is: no, Tesla is competing in cars with car companies that are doing cars, and they're competing with software companies that are doing software. So that gets you back to the autonomy disruption being only one of the strands through this. What's your competitive positioning? Is it that you've done something no one else can do? Or are you just trying to compete with the whole industry at doing something that industry knows how to do, and where do you want to put yourself within that?
--:-- Steven Sinofsky: I think this is part of the genius of the turn of the Tim Cook era at Apple: no, we're not going to make all the phones ourselves. Why would we make the phones themselves? No, we're not going to make the chips ourselves. No, we're not going to design this ourselves. We will pick a certain number of key points of leverage and make those ourselves, but we're not designing our gyroscope. These are incredible lessons from the Mac era, and they're putting them to work. This notion of a learning company is incredibly important.
--:-- One of the things I would like to raise is a great car company example. We talked about *My Years at GM* and the early days. Another famous GM experiment was the Saturn brand, an experiment in the 80s where GM was looking at Japan and they were losing everywhere. They couldn't make small cars. They couldn't make fuel efficiency. Labor costs were too high. The dealer experience was hard. Back in the 80s, when you wanted to buy a Toyota, you would just go in and they would say, do you want a red one, a black one, or a white one? If you went to buy an American car, it was all the stereotypes of the worst. They would sweat you in a small room forever. You would have to figure out option packages number T43 or QR7, and they overlapped, and it was a horrible experience.
--:-- So what GM did is they did Clay Christensen before the book existed: they started a whole brand. They hired all different people. They relaxed every constraint imaginable and said, go do it. Except, to your point, what they were trying to do was have this one badge of GM compete with all of Japan. It turns out it's very, very hard to do that. It ended up costing billions of dollars, and they shut the whole thing down as a failure. Books have been written—another great book is about the history of Saturn. Did they really fail because they couldn't make all the changes, or they couldn't recognize the changes, or they misunderstood what was really going on?
--:-- Benedict Evans: Yes. I suppose you could argue that the Saturn thesis would have been: we'll shut down all of GM and Saturn will become GM. At that point, maybe it would have worked. But as long as you're going to continue running it as a separate thing, well, what about the rest of GM? And that's where so much of this becomes very, very interesting. Ultimately, your framework for thinking about Tesla really raises so many interesting questions. I think it all comes back to what our founders always need to make sure they think about: it's never just the product. It's never just the price. It's never just the way you promote it and use channel management. It's never just about the pricing structure. You have to really consider all of these elements.
--:-- The thing I was thinking about recently was one of the accumulated learnings over my career: I'm always a sucker for a beautiful product. And the thing I've learned over time is, okay, yes, but what's the route to market? Yes, but what's your differentiation? Yes, but how are you getting the components? Yes, but what's your sales process? Yes, but. Yes, but. Our final thought on this is that the beautiful product can really get you in front of a customer, but it takes a lot of things to get you in front of all of the customers. Or you can get in front of all of the customers, but it's going to take a lot more to fully meet their needs in a differentiated way and get the price and the margins that you need. All of those things are really coming together. I think where we're seeing things now is that you have to start to consider all of those and not just anyone. And that's what's so interesting about this.
--:-- Steven Sinofsky: Right. Thank you.
--:-- Benedict Evans: Thank you.
--:-- Host: We hope you enjoyed this episode of the a16z podcast. If you'd like to learn more about all things tech, innovation, and how to build the next generation of companies, be sure to visit future.com, where you'll find essays, explainers, and more from those on the front lines of the future. You can also sign up for our newsletter, which curates the best of what's published on future and beyond. Just go to future.com forward slash newsletter. Thanks again for listening.
