Ed Niedermeyer and the Future of Self-Driving

In this week's episode of Industry Focus: Energy, Fools Nick Sciple and John Rosevear bring in the Autonocast's Ed Niedermeyer for a fascinating talk about autonomous driving -- including the hype cycle, software and hardware issues, legislation, venture capital, and a hundred other of the thousand things the industry encompasses. Ed explains where the industry seems to be today, some exciting recent autonomous news, some of the biggest sentiment changes we've seen play out in the last few years, why self-driving is so much more complicated than so many anticipated, and much more.

Plus, on the back half of the show, Ed talks some about his book -- Ludicrous: The Unvarnished Story of Tesla Motors. Tune in to hear some of that unvarnished story, from flaws in the company's culture to the serious safety concerns and more.

To catch full episodes of all The Motley Fool's free podcasts, check out our podcast center. To get started investing, check out our quick-start guide to investing in stocks. A full transcript follows the video.

This video was recorded on Jan. 30, 2020.

Nick Sciple: Welcome to Industry Focus! I'm Nick Sciple. Last week, John Rosevear and I sat down with Ed Niedermeyer, auto journalist, host of the Autonocast, and author of Ludicrous: The Unvarnished Story of Tesla Motors. I personally started listening to the Autonocast about a year and a half ago, and it's been invaluable to me in learning about the technology behind and the practical considerations that drive the self-driving car industry. In our conversation, we cover a wide range of topics, from how the hype cycle has driven attitudes toward the self-driving industry, the recent reveal of the Cruise Origin ridesharing vehicle with no steering wheel or pedals, and whether Uber really should be developing its own self-driving stack. I hope you'll enjoy our conversation.

Ed Niedermeyer, thanks for joining us!

Ed Niedermeyer: Thanks for having me!

Sciple: Ed, you're someone who focuses heavily on the autonomy space. This industry, I know from my own personal experiences, as I've tried to learn about it, it's very complex. And it's really easy to have perceptions of the industry that maybe don't fit what the state of the art is today. So when you look at the industry today, what do you think is the number one thing people get wrong about it is?

Niedermeyer: Man. Well, it's really hard. As you say, autonomy is a super complex issue. Actually, even just tracking perceptions of autonomy itself -- you could break just public perception of autonomy out into its own area of study, and it would be fascinating and complex and super hard to nail down. Even though you don't have the proprietary trade secrets and stuff that you have on the technology side or the business side, for that matter. So trying to characterize any one public perception is just difficult.

I would say, what we've seen happen is this wave of hype that happened, from 2013 through about 2018 or so. And in the last couple years, we've seen this move into what we call the trough of disillusionment. And I think in both of those cases, what we've just seen is that perceptions have always gone to the extremes of what's actually going on. Like, the signal has been pretty consistent. And the noise has just gone up to these extreme highs of overhype and extreme lows now, I think, of pessimism. So, that's really where the perception issue is, I think.

John Rosevear: Yeah, it's classic Gartner hype cycle. I mean, we've seen this before with other emerging technologies, but this is so big, and such promises were made for it and being touted three, four years ago. Here we are now, it's like, "Oh, this is never going to happen." But that's not what the signal's telling us, either.

Sciple: What do you think has been behind this difference in perception versus what's going on in the industry? I know we've had a number of operators that really were very optimistic about the space, as you mentioned, back earlier in the decade. And over that time, we've seen a number of deadlines come and leave. What's been the divergence between what the industry thought that they could deliver on these deadlines and where we're looking at today?

Niedermeyer: I think you have a mix of factors, right? So, definitely, in any technology sector, optimism is almost a precondition. You have to be optimistic to try and develop new technologies and build new business models. So, some of this is just a product of that natural optimism. I definitely think that there's a dynamic in venture capital. It was a little bit of a surprise to me, learning about venture capital, how trend-driven it is. The mythology around venture capitalists is independent, you know, the lone man surveying the horizon from this craggy clifftop. But they're very herd-like, really.

And what you have when you get these trends is these rushes in a certain direction, and then these not ideal incentives for founders and people who do understand the technology to overplay it, to overhype it, and especially when you start competing to get the biggest valuation, the biggest round, or whatever. And so, yeah, I think a combination of natural optimism and not ideal dynamics in venture capital have led to a lot of overhyping, a lot of over promising. And this trough of disillusionment, to some extent, is just the natural price that you pay for that.

Sciple: Absolutely. As we've seen the trough of disillusionment play out, and we've seen access to capital in this space start to slow down, we've seen some consolidation in the space as these main autonomy players have coalesced around major OEMs. We've got Ford and Argo, GM, Cruise, and Honda all working together. Do you see that as a trend that's going to continue over time? That we look five, 10 years down the line, we see these major autonomy operators partnered up with traditional auto OEMs, and those being the folks who drive this industry?

Niedermeyer: Yeah, absolutely. It's been really fascinating to watch, because the narrative in the early, more hype-filled days was that Waymo and maybe a handful of others that were the leaders were going to be able to reduce the automakers to sort the Foxconn of cars. 

Rosevear: Yeah, hardware makers. And Android or Waymo was where the money was going to be. That was the story.

Niedermeyer: Which is strange, right? Because, it turned out, lo and behold, these are very tightly integrated systems. And when you have extremely complex technological devices, whatever they are, the more tightly integrated they are, the better. And so, I think some of this is just a natural transition to, as you get closer to the problem, and you get closer to an understanding of what this product is going to eventually look like, the more you appreciate the fact that no one part of this is just that much more important than the rest of it. It's just about integrating a bunch of really complex technologies into the best possible package. And frankly, that's what carmakers do, right?

Rosevear: Yeah. To go to your point about the power of VCs, there was a certain current from the VC that all of the good stuff is going to be invented south of San Francisco, the dumb people in Detroit just can't do this. I think everybody's kind of come to an understanding that that's not actually the reality, that these are more partners rather than Detroit banging at the hardware for the great software that will collect all the revenue and everything. And that has been interesting to watch, too. I think the example of Tesla has been instructive for a whole lot of people. Like wow, making cars is really hard! [laughs] I think a lot of people have had that awakening in the last three years or so.

Sciple: Yeah. Ed, as we talk about how the technology has developed, we've gotten more realistic expectations of how these things are going to play out over time and how these companies need to work together. We've seen, John mentioned Tesla's Autopilot, the AS system that's on the road today. We're starting to see other automakers start to roll those autonomy features out. When you look at that sector of autonomy separate and apart from that level four-plus full self-driving, how do you see those features starting to roll out here in the next few years?

Niedermeyer: That's been a really interesting thing to watch. And to echo what John was saying, clearly a lot of people in all aspects of this space, if you look at automated driving, or really mobility technology more broadly, everyone's been watching Tesla. And though the pros tend to outweigh the cons in public perception, I think certainly for the people who are running these companies, they're paying as much attention to the challenges that Tesla's had as to the pros, and they're learning, yeah, some of this stuff is really hard, it just doesn't make sense for us to reinvent the wheel that way when we have these willing partners who are not only bringing know-how, but also, by the way capital, which we need to survive because we don't have viable business models right now. 

Rosevear: Also employees. 

Niedermeyer: Yeah, totally. And that's where the driver assistance system kind of actually comes in. Because these companies that are developing, and a lot of it on the sensor side, they're finding, "Oh, wow, the market that we thought was going to be here for robot taxis in the shorter term is just not materializing." We're seeing a lot of companies pivot into a dash driver assistant, which Tesla has done a lot to popularize. I think it's also done a lot to raise questions about, given their specific implementation. But I think one of the influences that Tesla will have in a lasting way is that that technology will become very common. and what the auto industry is good at is commodifying technology.

Sciple: Sure. You mentioned the sensor suppliers, which is an area I wanted to talk about briefly. Over the summer, there were suggestions that Velodyne Lidar was getting ready to IPO. We've seen the costs of those systems really come down significantly over the past couple years. When you look at the sensor market and the companies playing in that space, what should we know about that as investors and folks following the space?

Niedermeyer: I would say sensors, every time I talk to venture capitalists and investors, I always ask the question, what sector of automated driving in particular is the most over-invested? And the answer is literally always sensors. And if you get them to drill down, it's almost always LIDAR. There are tens of LIDAR companies. There's no way the market can support that. We've been waiting for consolidation, frankly, over the last year. There's got to be some this year. We have seen some companies fold up. One company returned money to investors. I think there will be more forced down that route.

Sciple: And is that just the commoditization of the industry? Or that self-driving has not ramped up as quickly as folks expected, and there's over capacity? What's been driving that?

Niedermeyer: A little bit of everything. Definitely the fact that the robo taxi market hasn't developed. There are [...] businesses that you can use as a stock app, but there's only so many. There's this belief in Silicon Valley that if you have the best tech, the world beats a path to your door. And what they're finding, what the smart companies have really embraced is that there's a whole skill set around incorporating yourself into the automotive supply chain. You can't just arrogance your way into that. You have to work the system, and the companies that have embraced that are getting the most contracts. And the ones that don't aren't going to make it.

Sciple: Yeah. Let's move on and talk about some of the companies that are operating in this space. As we record here on January 24th, just this past week, the big reveal was the Cruise Origin vehicle rollout. You were there in person, Ed. What should we be taking away from this rollout?

Niedermeyer: Yeah, that was really interesting. I'm actually working on a piece about that right now. I think the event itself did not do a great job of explaining what I think is important about that vehicle. I think one of the most important things that came out of that event that really resonated with what I saw in the vehicle itself was, Dan Ammon, actually, the CEO of Cruise, answered a question that I see people get stumped by a lot in this space, which is, what does an AV offer to a consumer that a human driver in an Uber or Lyft doesn't?

This is actually a quite a difficult question to answer. A lot of people will assume privacy. No. It's going to be full of cameras. Not that. Cost? Well, not necessarily. Although I do think that the Cruise Origin vehicle is clearly cost engineered in a way that I think is really interesting and really important. But the answer that he gave was not "it's going to be cheaper than Uber and Lyft." It was, it's going to be more consistent.

So the perspective that I think is crystallizing -- again, I feel like I'm doing a lot of work here; they're not necessarily communicating this specific message very well, if, in fact, this is the deal. But basically, I think they're positioning that vehicle and their service to be the McDonald's of mobility, where it's that consistency. You always know you're going to get the exact same vehicle, the exact same driver. Even if you're paying the exact same price, or even maybe a slight premium, that's going to be worth it to a lot of people. That's how McDonald's went from a mom-and-pop little shop to a big -- of course, franchising was part of that. There's a lot.

But that consistency is what McDonald's brand is built on, and that doesn't exist in ride-hailing today. And that is actually a really fascinating opportunity. And if you look at how the vehicle itself is engineered, I think it also fits into that kind of McDonald's value, not fancy, not luxurious, not space age, just good value and practical and easy to clean. And I think those things matter. And I think that really, what that vehicle shows is that Cruise is thinking a lot about a business.

The problem is, from a communication side, it's very hard for all of the companies in this space to pivot from the soaring, "change the world, get rid of road deaths" rhetoric that they led with on their communications, and go from that to "we're just building a business that's going to work." That's a tough transition to make. So, I understand why they're having a hard time doing that. But, that's actually the question that needs to be answered right now. Can you actually build a business that's going to work? And that Cruise Origin vehicle really improve my optimism that Cruise can, in fact, make that happen.

Sciple: Ed, when I look at this vehicle -- maybe tell me if this this comparison isn't fair -- it didn't look all that different from what May Mobility has been trying to do with these fixed shuttle services. How different or similar is this Origin vehicle to what May Mobility has been trying to do?

Niedermeyer: It's funny, I see that perception a lot from people who weren't there. Honestly, if I hadn't been allowed to go on stage and sit in the vehicle, I might make that comparison as well. It's not, it's a different vehicle than that. First of all, it's not a low-speed vehicle. There are two regulatory categories, low-speed and everything else. It's going to fit in the everything else category, which, of course, is a challenge right now because we don't have rules for that. Cruise, and others, everyone is having to talk to the regulators about that. And that's just not having a steering wheel and some of these other Federal Motor Vehicle Safety Standards. 

The vehicles that I would compare it to most are actually the London Electric Vehicle company, which is owned by Geely, a Chinese company. They sort of reinvented the classic black London cab. It's this very bulbous, spacious -- big, flexible interior; big, flexible opening doors, and bench seats that face each other.

I think Toyota's JPN taxis for the new Tokyo Center taxi is similar to that as well. Those are the vehicles this reminds me of. It's a robo taxi. And honestly, if I've been in charge of comms at Cruise, I would have probably presented this instead of "beyond the car," which is, again, aspirational and ambitious and all that, but vague; I would have said, "This is a robo taxi for the real world," because that's really what it is. It's the first robo taxi I've seen where I'm like, OK, I can see a fleet of these operating as a viable business.

And frankly, in this business, that's as big to me as even being able to go on a fully driverless right in a Waymo in Arizona, which was also a big moment for me personally, and I think for the space. But this is almost as important to me, because the question of can you operate these things as a viable business is every bit as big as can you make cars that drive themselves.

Sciple: Let's move on to Waymo. You were the first journalist that was able to ride in one of these without a driver operating the vehicle. What did you learn from that ride?

Niedermeyer: I learned that the psychological impact of not having anybody -- of putting your life in the hands of technology is much greater than I expected. I've been in quite a few autonomous vehicles, always with safety drivers. And I sort of thought I'd seen it all. I felt kind of jaded by the time I did this, frankly. And the psychological impact of realizing, "Wow, I'm sitting here putting my life in the hands of this robot, and there's no human there ready to take over in case something goes wrong," that was more psychologically impactful than I thought.

And it really reinforced something that I've been thinking for a while, which is that trust is really the most important thing in this business. And that really runs counter to what we saw in the early hype years of autonomous vehicles, where it was this race, and people were doing these stunts and these kind of shady demos and stuff, and cutting corners, even, as we saw with Uber in Arizona. I think when you realize what an impact it is for consumers to really understand they're putting their lives in the hands of a robot, as that sinks in, more and more companies have to take it slow, not take as many risks, be more responsible, because the trust aspect is so critical to this business.

Sciple: Another thing I've heard you talk about when it comes to Waymo and the autonomous ride is, when you look at, snap your fingers, assume this technology exists here today, that the driver is able to move around in the geofenced area -- there's all these other things that rideshare drivers and those sorts of things do, whether that's trying to find your driver on the street and them being able to flag you down, and that sort of thing. How should we think about those sorts of problems as we look at the future of autonomy?

Niedermeyer: I think the story of autonomy in a way has been, like a lot of things, this progress from thinking of it as a hard problem, but sort of like one hard problem, which is teaching the cars how to drive. And as you get away from that R&D type of approach to it and toward more of an actual product development approach, you start to realize, when you take a human out of the equation, you're not just taking a human out of the driving equation; you're taking a human out of a whole customer experience, the customer service experience. Waymo's doing a lot of research into that. It's been interesting, too, to see the Cruise Origin. Actually, I was just as Zoox today, it's another company specifically in the robo taxi space. They all have very different approaches to this consumer product experience question. I don't think we have all the answers to that. 

So what's interesting is, as the technology matures, and as that becomes, we can build more confidence -- so, yeah, we're going to be able to teach cars how to drive, even with some limitations, we're still, I think, as a space, the sector is still beginning to understand that customer interactions can be very complex and unpredictable as well, just like driving can. And they kind of have to now focus a little bit more on that, and how do you solve some of the more complicated consumer interaction problems without a human in the loop?

Rosevear: This has been interesting for me to watch, because it's like the automakers know more about this than the tech guys. I mean, for Ford and Domino's. Everybody's laughing at that because there's a human hidden in the fake self-driving pizza delivery thing. But they were doing that to gather this kind of information. How does a self-driving pizza vehicle deliver a pizza, with nobody to get out and bring it to the person? Slightly different space, but these kinds of questions, I think, are much more important than the technology side realized until fairly recently. And it's like, "Whoa, OK, wait, we have to think through a taxi service." Yeah, I think people like Uber were maybe onto this more than some of the pure technology, pure software shops. But yeah, it's been really interesting to watch them creep toward these kinds of questions.

Sciple: Yeah. On Uber, that's another one, when you look at their autonomy aspirations over the past year, kind of hard to get a grasp on where they stand. They had the crash back in 2018. And then in 2019, it was revealed that if they want to keep their current autonomy strategy, they're going to need to pay a license fee to Waymo going forward. So, if you look at Uber's self-driving business as it exists today, where does it stand?

Niedermeyer: I actually don't really understand why Uber still trying to develop their own self-driving stack. It doesn't make sense to me in a number of ways. First of all, they've really been the poster child for what not to do on the trust issue -- which, again, is becoming more and more important all the time. They do have a good understanding of the consumer experience thing, which I think is valuable. But they're also a public company. They can't just keep raising money forever to funnel into a thing that is sucking down hundreds of millions of dollars a quarter.

And then finally, I think on a strategic level, the thing that I have not ever really understood about Uber is, how do they make that transition from being an asset-lite company that doesn't own their fleet to now, with robo cars, you're not just owning a fleet, you're operating a fleet. That's a whole business right there, with a lot of expertise and knowledge and all this kind of stuff that they can't just go out and necessarily acquire. 

I think they need to keep a toe in the technology, because I think ultimately, for better or for worse, they're going to be one of the gatekeepers, in terms of saying, like, "This vehicle is safe enough to let on our platform, versus, it's not." And they may in fact be making those decisions before regulators are, if we don't see some action by regulators soon. So, they need to keep a toe in the technology. Developing their own stack, though ... again, maybe I'm missing something, I'm not seeing the emphasis there.

Sciple: Right. And when you look at what Lyft's strategy has been, it's been to partner with other autonomous operators. The big one there is Aaptiv. That's been running their autonomous operation in Las Vegas for the past several years. Just announced a joint venture with Hyundai. When you look at Aaptiv, that's the only tier-one supplier that's really playing significantly in this space -- would that be a correct characterization?

Niedermeyer: It depends on what you mean by seriously. Almost every tier-one supplier has some strategy around autonomous drive. But as a full stack developer, yeah. And by the way, Aaptiv is one of the successor companies of Delphi. It was split off from the core traditional automotive supplier business, and sort of the tech part of the business. So they're a little more focused on that. And so, maybe they arguably have to do that, right, more than a Bosch -- although, Bosch is very involved with Mercedes and Daimler. [...] says they're investing a lot of money into this technology.

Frankly, I think a lot of start-ups have been talking about disrupting OEMs for a long time; I think what's turning out is that really, the risk in some ways is much more to the tier-one and even tier-two, potentially, suppliers, the supply chain, than the OEMs themselves.

Sciple: Last thing talking about companies and sectors and autonomy. Another area that gets maybe not nearly as much attention as the robo taxi business are these autonomous delivery robots that are either on the sidewalk or in the bike lane. Are we going to see these before we see these full self-driving cars on the road? How should we think about this robot delivery business?

Niedermeyer: I've definitely not paid as much attention to the delivery robots as I have to the vehicles. I sort of have come into this from more of an auto industry background. So, I think the big appeal of the delivery boss is regulatory -- well, it's two things. On the tech side and the regulatory side, both of the challenges, which is where some of the biggest challenges in the space lie, are made easier by the fact that these things just move slower and have less mass. Safety issues arise from speed and mass, and these have less of both of those.

That said, I'm not entirely convinced that there won't be -- there already has been regulatory backlash against these things, certainly on sidewalks. There's a company whose name escapes me right now that is targeting the bike lane. That's, I think, interesting because it avoids certain kinds of backlash, but then again, the quote-unquote bike lobby has its own power, too. Anywhere you go, you're going to have challenges. So, I definitely don't think that these delivery bots are necessarily a shortcut, at least at scale. Just like robo taxis, I think we're going to see them in limited areas where the governments and local regulators are OK with them, where there's not a populist backlash against them, and, again, if they can make money.

Sciple: You mentioned regulatory barriers. That's something I want to talk about as well. What are the current regulatory barriers to self-driving?

Niedermeyer: This is a very big, complex issue. I think there's a couple things, though, that are worth noting. Essentially, the Federal Motor Vehicle Safety Standards are the core of automotive safety regulation that every vehicle on the road has to comply with, unless it's a low-speed vehicle, which is under 25 miles per hour. Those standards are quite old, and a lot of them assume a human driver. There's been talk for a long time about building new regulatory categories.

There are two approaches to it -- you either build a new regulatory category and say, "This is an autonomous vehicle, and here are the standards that make sense for that," or, the approach that seemed like it was the one that was happening is, you just get waivers as you need them from the very specific rules. So GM was one of the first to apply for a waiver. They said, "We want to have a Cruise Chevy Bolt on the road with no steering wheel. Can we have a waiver to do that?" Well, it's done. Has it been two years since they applied?

Rosevear: Over a year, yeah. They've got another one going on with the City of New York, of course, and that one is just sort of sitting there as well. Yeah, it has been a long time.

Niedermeyer: Yeah. So, what's interesting about this, though, is that the obstacle right now is about the physical aspects of the car. There's zero regulations about the software, which is interesting, because Elon Musk has been forwarding regulatory barriers as this thing that could prevent us from updating our cars to be full self-driving capable. That's completely not true because there's no regulation of the software. And their cars, the physical parts of the cars, are all FMVSS compliant, otherwise they wouldn't be able to sell them right now. And so, if you have an FMVSS compliant car and you're just pushing software to it, right now -- in this country anyway, in the United States -- there's zero regulatory barrier. So, ironically, the guy who's complaining it about the most has the least to complain about, which, of course, raises questions about why.

Sciple: Right. When you say FMVSS compliant, that really is just saying you need to have a steering wheel, pedals, that sort of thing, correct?

Niedermeyer: Yeah, and with certain dimensions, places, locations, blinkers, lots of minutiae. But, a lot of which revolves around human control schemes. Because clearly, from a regulatory perspective, when you have humans driving, you want that control scheme to be fairly standardized so that anyone can just get in the car and use it. And so, those are the regulations that are a challenge, getting rid of the human controls.

Sciple: Do you think we'll ever personally own self-driving vehicles?

Niedermeyer: This is a really interesting question. I think if you asked me a couple months ago -- obviously, never say never, but I would have said that's not really on the horizon. I had a really interesting conversation, myself and the other members of the Autonocast, with John Krafcik of Waymo, CEO of Waymo. And he's mentioned it before, that it's possible that you might have a privately owned vehicle that is fully autonomous within a certain geofence. He did mention that.

He gave us a little bit more detail on that, a little more of a sense, let's say, of what that might look like. I think the prospect he raises is an interesting one, which is essentially, maybe there's a one-year lease or a short-term lease, where you get this vehicle that is capable of being fully autonomous within an established Waymo fence, but it's yours, and after that lease term expires, it maybe goes into the fleet. And I think the key factor in that, in how they're thinking about that, is actually just a service. Essentially, he was like, "I can't imagine trying to keep LIDAR service for 10, 15 years."

And that's the thing. With a privately owned car, it's engineered to be about 10 years, or more, hopefully. Average car on the road is 11 years old, or at least it was; I think it's going up. And so, they can deal with the cost part of it, actually, by having a shorter-term lease, probably charging a hefty premium for it, and then putting it into a fleet of ride-hailing or delivery vehicles. I think that's how they make the economics of it work. So, my mind has expanded on this. And that's one of the great things about this space, is people are thinking very creatively about a lot of different ways to deploy this technology. You have to keep your mind open and be ready to learn all the time because things are not set in stone.

Rosevear: Another discussion a few years ago, I had this discussion with a senior board Ford like, that autonomy could be something you switch on in a normal passenger car. Like, you go to a dinner party and it's late, you had three glasses of wine, well, the car drives your home, that kind of thing. People are not thinking about that anymore because of the sensor issue, as you said. It's, is the car itself driving or is it not, for the most part, unless it's a very expensive car or a car that is extravagantly maintained, at least right now, with what people know how to do, it's not going to work that way.

Niedermeyer: No. And yet, at the same time, I think one of the reasons Tesla has been so successful with the oppression they've had is that it's very hard for people to think about their own life now and imagine a future where they just don't own a car, especially, again, in this country, where that's so fundamental. And what Tesla benefits from is, on the one hand, they're giving you this new, futuristic, cool thing; on the other hand, it's coming to you in a totally recognizable package, which is the privately owned premium sedan. That idea works for people as a story, because it bridges the gap between this future that's a little bit too different to just embrace and what they have right now. Now, just because it works as a story doesn't mean it makes sense beyond that narrative level. I think that's where Tesla gets interesting.

Rosevear: It's one of many places where Tesla gets interesting. [laughs]

Sciple: Yeah. Before we move on to Tesla, I do have two more questions I want to ask about the broader autonomy space, before we dive into Tesla and into the book and all that. You mentioned Waymo. We talked about the geofence currently in place, as far as operating driverless robo taxi services, it's really only in Chandler, Arizona right now, with Waymo. As you look at these businesses scaling and moving to additional geofences and that sort of thing, how do you see that playing out over time? And what kind of timeline is realistic to expect?

Niedermeyer: The timeline... man, I'm just going to take a pass on that because just way too hard. I do think, one of the major perception hangups in the general public understanding autonomous vehicles has been the paradigm of what a car is and how they operate it. That's so well established, over 100 years, and billions of dollars a year in marketing. So we think of autonomous vehicles as just being self-driving cars. And clearly, that's not the case.

And clearly, there are a lot of things that you need to do to enter a new market. Some of that are technological, like adapting to local driving conditions, building a new map, understanding where you are and what you need to avoid, those sorts of things. But a lot of it is economic as well. It's expensive to scale. And I think people also really underestimate, if you're operating a fleet, that takes real estate, it takes people, it takes equipment, it takes investment.

And so you're investing huge amounts of money in the technology itself. And then, to scale this -- it's not a typical software product. I think a lot of people thought of autonomy as a software product, and once you develop it, you just copy and paste it and it scales. That's not how this is going to work. It's going to be much more like scaling, in a way, almost, a traditional auto business, in that you have to invest in bricks and mortar and people and organizations and structures and local government relationships and all these things.

We're not going to see a typical tech company scale, where it's like, "OK, we've proven it works. It's going to be everywhere." It's going to be piece by piece, city by city. And as my co-host Alex Roy says, what's really going to be interesting is when you start to get multiple providers in the same city. That's when I think the space can be really fun to watch, because then you get some of that head-to-head competition, and you'll see what's working, what's not.

Rosevear: And you're not going to see tech margins, software margins, [laughs] when that happens, either. That's another issue. I mean, you've got to make money here. That's something that people like Dan Hammond are talking about a lot, but not everybody talks about as much as they should be talking about it. Once you've deployed all that hardware, once you've bought all that brick and mortar, once you've hired all those people, what's left of your margin? Yeah, you can get rides down to $0.98 a mile or whatever, but if it's costing you $1.03 a mile to deliver it, that doesn't work as a business. That's another place where it doesn't work as a business.

So what do you cut? What do you change? How do you do that? That's the kind of thing that will be sorted out with competition. A lot of the companies that look like sure bets two years from now may not look like sure bets four years from now.

Sciple: Absolutely. And that is endemic of the hype cycle, the peak of the hype cycle. And after that trough of disillusionment comes, if you look at the hype cycle, it's against the slope of enlightenment, where you have this slow progression over time of the technology taking place. Maybe it's not as flashy and exciting as we thought at the beginning, but it becomes part of our everyday lives and integrates with what we do. 

Ed, when you look at self-driving, we're now in this trough of disillusionment. But as you look out 10 years down the line, we're on this slope of enlightenment. What does that look like?

Niedermeyer: Interestingly, I think a lot of this hype cycle stuff really does exist in the realm of perception. I think this Cruise Origin reveal encapsulates the sort of complexity of where we are, because in a way, as a product, I feel like that's one of the first hints of how I think what the slope of enlightenment might look like. It's not trying to be fancy in space. It's not trying to impress the typical tech hype kind of thing. It's very simple, basic, engineered to a cost kind of a thing. 

Rosevear: It's a bus.

Niedermeyer: Yeah, it's like a taxi bus. It's a new thing. But it's pragmatic, and it's clearly built around operating a sustainable, viable business. On the other hand, the communication around the launch of that vehicle has been really mired. It's weirdly mired both in the AV sector and the auto industry, in a weird way. I don't mean to rag on Cruise's comms team so much, but clearly, there were challenges there.

So to the extent that peak hype and trough of disillusionment and therefore also the slope of enlightenment exist in the realm of perception, weirdly, communications are super, super important. I can't blame anyone in this space, because just like this technology is new, developing new technology, there are unknown unknowns, and you just get these really challenging things that come up and you can't anticipate; I think the same thing is true in communications and managing public perception. That's a lot of what we're seeing, is just finding out how to modulate those expectations, how to get people excited in your company, but not too excited.

So, weirdly -- maybe it's because, as a writer, I focus more on that perception side, but I think it really, really matters. And I think that companies are going to have to really invest a lot in that so that they can deal with that, because it's not enough to have the vehicle that looks like the slope of enlightenment; you have to get people to understand what that slope of enlightenment is, what is realistic, and how to think about that.

Sciple: Let's change gears now and talk about Ludicrous, your book on Tesla. We just got done talking about getting people really excited and jazzed up. I think Tesla has definitely succeeded at that. This was your first book. Why did you decide to write it?

Niedermeyer: Boy, so long ago now. Look, I think Tesla is the most important story in the world of cars. That's really the core of it. I had obviously been aware of them for a while, and I wrote some blog posts and stuff about them in the past, but I didn't really take them seriously until I got interested in the battery swap thing because I'd been really interested in the company that was doing that. And that was the first exposure I had to mobility tech as something distinct from the auto industry.

And so, with Tesla, it's important in multiple ways. Yes, it's important because they popularized electric cars in a way that hadn't been done before. But I think, more importantly, almost, to me, is that they popularized the auto industry more generally, in a way that it hadn't been before. That's great. For someone who writes about the auto industry and with my background, that's a wonderful thing. I'm very excited that more people are interested in that. 

Unfortunately, what's gotten that interest has been a narrative that very often diverges from reality. And so that was the opportunity I saw. For me, at its core, it was an opportunity to leverage the excitement that Tesla had created about the auto industry and use it to ensure that the right lessons are being learned.

And now, of course, the Tesla fans say, "Well, yeah, you think the right lessons are that Tesla sucks." One of the challenges with the book was absolutely getting away from this intensely polarized discussion about it. And frankly, it ended up being a very therapeutic experience, where I could really purge some of my emotional responses to the topic, which was good.

But really, the core of it was, this is an opportunity to get people who wouldn't have paid attention to auto industry before to understand why is it the way it is? Is it that these companies are all just dinosaurs who never change because they're dumb? Or have there been evolutionary forces that have shaped the auto industry into what it is today? And why has Tesla been able to be really, really successful at some things and really, really not successful at other things? What are the cultural differences between tech and the auto industry?

I think their story illustrates these really important things, and especially illustrates the most important lesson, I think, you can take from any story about this new world of mobility tech which we discussed just a little bit ago, which is, it really takes both of these, right? You need the technology, you need start-ups to innovate and come up with new stuff. But to turn that into a product that exists in the real world and can last for years or decades and millions of miles, and all the things that need to happen for these to become real products, you really need the auto industry, especially to make them at scale. 

And so the real core challenge that Tesla's story illuminates, I think, is the need to make these two very different cultures somehow find it an accommodation, and the ability to come together, because it's in both sides' self-interest.

Sciple: You talk about culture a fair bit in the book, comparing and contrasting Tesla's manufacturing culture to the Toyota Production System that really governs much of manufacturing across the rest of the other businesses that do manufacturing. How significant is that cultural difference? How does that shape Tesla?

Niedermeyer: They're almost polar opposites, right? You think about what a start-up is -- it's small, it's creative, it's a small group of talented people who operate in a loose, unstructured way in order to be as creative as possible and come up with creative solutions to tough challenges. Manufacturing culture is almost the exact opposite. It exists because scale demands it. So it's fundamentally scale-based rather than built around a small group of people. It's much more about fitting each individual thing into their role in the broader system than necessarily be as loose as possible so that you can come up with as creative stuff as possible.

And it also is about regimenting the pace, not just the relationships between the different parts of the structure, but making sure that every part of this massively scaled system is all moving at the same pace so that all the pieces interlock harmoniously. Which, again, with a start-up, it's more like, pick up a Red Bull and hackathon your way through this problem, then go sleep for a couple days, then come back and do it all again. It's very fits and starts.

So it's almost like, if you think about human personalities, it's the two polar extremes of the creative type and the company man sort of thing. It's more than that, but, can you imagine a couple made up of those two kinds of people? It certainly happens; opposites attract, opposites very much complement each other. And frankly, to me, one of the big turning points in Tesla's history, and really some of the tragic things is, they had partnerships with Toyota and Daimler, two of the most respected companies in the business.

And what my sources have told me about that is that it was basically both of those relations, particularly the Toyota one, were pretty much doomed from day one. It wasn't just Toyota's fault. There was definitely arrogance on both sides. There was a lack of appreciation of what the other side could bring to a relationship, and an unwillingness to make concessions in order to make that relationship work. And unfortunately, I haven't seen any progress, certainly from Tesla, toward wanting to have that kind of accommodation. 

What you have seen, however, is more recent start-ups who are clearly inspired by Tesla, clearly getting funding because of Tesla's success, who recognize, there's something that Tesla's done really well. But there are other things where there's a lot of own goals, and manufacturing is one of those things where, OK, why would we go through the multiple production hells that Tesla keeps putting itself through every time it comes out with a new product when you could partner with people who know this stuff works, and it just requires you to make some concessions and accommodations to the way they can be successful?

Rosevear: It's so different to talk to people from like Lucid or Rivian, and just to listen to really how much they've taken the Tesla lesson to heart. I mean, Rivian went out and cut a deal with Ford. I mean, they got Joe Hendrix on their board. And he's Ford's manufacturing guru. They really went that way. Lucid hired a bunch of Tesla veterans, and also a whole bunch of industry veterans, people from Mazda and a few other companies who really knew manufacturing. And they're like, "We're going to do this the right way, not some new way." And it was really interesting to hear that, in contrast with what we've heard from Tesla for seven years now.

Sciple: We spent the bulk of this conversation talking about autonomy more broadly. Let's talk about Tesla Autopilot. When you look at that technology, obviously, the big number everybody talks about is the million robo taxis by the end of 2020. When you look at the state of the art of that technology today, how do you handicap the likelihood of that playing out?

Niedermeyer: Look, I think Tesla's story is really interesting, because there's a lot of good and there's a lot of bad. I think if you look at the company holistically, certainly, that's the case. As an electric-vehicle manufacturer, that's absolutely the case. They've done some really, really good things, along with having some challenges that they've had to overcome, and maybe not even overcome fully.

But I think when you look at Tesla as an automated-driving company, I think the story is very different. I think it's very unambiguously not positive. Autopilot was essentially a system that Google had developed early in its self-driving car program, and they refused to put it on the market. They at one point considered putting a product just like Autopilot -- in fact, called Autopilot -- on the market, and refused to do it because of the safety concerns around predictable abuse that we're now seeing pop up in these NTSB investigations of the people who have died and run into ambulances and things like that.

And so not only do they take a product that Google certainly knew was unsafe, and say, "Well, if you're not going to put it on the market, we will," but also, multiple engineers have come forward and said to the media that, "We need to have better driver monitoring, whether that's a camera-based system or a capacitive touch thing to make sure your hands on the wheel, instead of having that torque-based wheel sensor." And clearly, I think history has done that.

And really, throughout Tesla's involvement in automated driving, I've always gotten the sense that it's like, they know that autonomous drive is coming; they also know that because they've got all this other stuff on their plate, they're maybe not necessarily going to be able to compete by doing things the way that the consensus, like the level-four robo taxi fleet thing that's really formed a consensus around how this technology can be deployed.

And so what they've done is, they've tried to pull together as much of it as they can into a consumer product. And so they're unique in the marketplace in that sense. Nobody else is out there saying, "Buy this car, and it will either be sort of self-driving or it will definitely be fully self driving at some point."

But I think it's important to understand the history. I think it's important to understand how limited their sensor stack is. To me, it's the hardware. They can update the software. And if you just believe that computer vision is going to be [...] reliable anytime soon, you probably believe that Tesla is going to be able to deliver on this promise. I do not think that that's the case. If you talk to basically anybody in what I would call the mainstream of the autonomous vehicle business, none of them think that that's even close to on the horizon.

And Tesla's history is full of them having to push the boundary on things in order to survive because the core of their business has always been so challenged. And I think this is where that pattern reaches a breaking point, because they have taken money at the peak of hype in late 2016. They started taking money based on a level-five capability that nobody else in the industry was even talking about, and with a sensor suite that's an 8-S, a driver assistance, not a full autonomy sensor suite. I don't think there's any way you can do it.

And frankly, I think this is the biggest question right now about Tesla, is how do they manage that? Because they do have loyal fans, but they've been getting a lot of money from people, and if this starts turning into lawsuits, let alone fraud charges or anything like that, I think it'll get really interesting, fast.

Sciple: Ed, you wrote the book. The book came out back in August. Obviously, a lot has happened with Tesla. A lot always happens with Tesla, but particularly in that period since the book came out. If you were writing the book again today, what of what's happened since then would you add to the book and why?

Niedermeyer: One of the last chapters, I ended with the Roadster and Semi reveal, and used that as a way to show that internally, the culture has spiraled a little bit out of control. And it's like, what Elon thinks is cool is going to happen whether or not an electric semi is a product category that makes sense, let alone whether or not there are synergies with what Tesla is doing in its core business. If Elon thinks it's cool, it happens.

And this is a very, very core issue with Tesla and with their culture. I wish that I could have used the Cybertruck reveal like to illustrate that, because I think it illustrates it even better than the Semi and Roadster reveal. The Cybertruck reveal really illustrates what a tight rope Tesla is on. To me, I mean, whatever you think Tesla is worth as an investor -- and I don't do that kind of financial analysis, I'm not an investor, I've never given investor advice. Tesla's story has really been the first time I've been exposed to the whole stock market aspect of all of this in a really intense, personal way.

And I think no matter what you think Tesla is worth, or how much money they're going to make, or any of that, you have to recognize just by looking at the history, that there has been so much risk, and there continues to be so much risk, and Elon Musk is a person who is intensely comfortable with crazily high levels of risk. And when the Cybertruck reveal happened, I had people asking me, "How could this happen?" And my answer was really, it's surprising that something like this hasn't happened already because there have been so many instances where they've been incredibly lucky to have not had some kind of real egg-in-the-face moments and real confidence-shaking things.

And, frankly, from my perspective, certainly, looking at the future of that company, since they're not trying to change how they operate and what their culture is like, I think that more Cybertruck-esque, and potentially even crazier, situations await us in the future.

Sciple: Yeah, Tesla's everybody's favorite company to follow. I know I'll be following it very closely going into the future. Ed, thanks so much for coming on the show, sharing your knowledge. If folks want to follow your work going forward, where can they find it?

Niedermeyer: I'm on Twitter at @tweetermeyer. I'm also focused full-time at this point on the Autonocast. I'm doing some freelance work, which you'll be able to find by following me on Twitter. Otherwise, autonocast.com, or subscribe to the Autonocast on your favorite podcast platform.

Sciple: As always, people on the program may own companies discussed on the show, and The Motley Fool may have formal recommendations for or against the stock discussed, so don't buy or sell anything based solely on what you hear. Thanks to Austin Morgan for his work behind the glass. For Ed Niedermeyer and John Rosevear, I'm Nick Sciple. Thanks for listening and Fool on!