Outlaw racer Alex Roy & roboticist Bryan Salesky sit down with Sam Abuelsamid — Navigant Research Analyst, Forbes writer and one of the world’s leading experts on the automotive & technology sectors — to discuss and debunk the top myths about autonomous vehicles. Will your teenager need a driver’s license in the future? Is Tesla winning the self-driving race? Will car accidents exist two decades from now? On this episode of No Parking, we cut through the industry hype and serve you the reality of autonomous vehicles straight up.
Alex Roy: Welcome to the No Parking Podcast. I’m Alex Roy, and I’m here with my wonderful co-host, Bryan Salesky.
Bryan Salesky: Hey Alex.
Alex Roy: This episode, we are going to talk to someone… Do you want to try to pronounce his name?
Bryan Salesky: Sam Abuelsamid.
Alex Roy: I think that’s accurate. Sam is one of the best journalists on the planet covering autonomous vehicles, which is why he’s the perfect guy with whom to discuss the topic of top 10 autonomous vehicle myths.
Bryan Salesky: Sam really does his research. He is… really, very few in the field that… He’s constantly traveling, he’s visiting the different teams and companies working on this sort of technology, and he goes deep, asks great questions. He’s one of the best.
Alex Roy: Doesn’t write clickbait, and also the author of the Navigant Research AV sector leaderboard. Let’s just roll right into it. This guy knows what he’s doing..
Alex Roy: So it’s always great to hang out with Sam Abuelsamid from-
Bryan Salesky: Did he say your name right?
Sam Abuelsamid: As well as anybody ever does.
Sam Abuelsamid: Something like that, yeah.
Alex Roy: Okay. Who writes the leaderboard that everyone cites in pretty much every article on the planet for which AV company is ahead or behind. People fight over it, it’s awesome.
Sam Abuelsamid: Yeah, I think “leaderboard” is maybe not necessarily the best title for it, but it’s legacy so we stick with it. It’s more like, “Who’s most likely to actually succeed?”
Alex Roy: Okay. I’ve given you some grief in the past, but actually, now that I understand your metrics, like what you actually use to rank the companies, I get it. so it’s only appropriate that Bryan Salesky and I come to discuss with you the Top 10 myths of the AV sector.
Sam Abuelsamid: There are no myths in the AV sector. Everything you hear in the media is absolutely true about this.
Alex Roy: Yeah, everything.
Sam Abuelsamid: There’s no question. If you hear a CEO of a company that’s developing this technology say, “We’re going to have a million of them in 2020,” clearly it must be true, right?
Alex Roy: Who would ever say that?
Sam Abuelsamid: You’re not allowed to lie about stuff like that, are you?
Alex Roy: No one ever lies in this or any other sector. Let’s roll right into it guys. I went on Twitter and I posted a request for top 10 myths foisted on us about AVs in the last five years. I got hundreds of responses, I thought I’d just read them out, round robin.
Bryan Salesky: Do I have to participate in this?
Sam Abuelsamid: Oh, yeah.
Alex Roy: Absolutely.
Sam Abuelsamid: Oh, yeah. I wouldn’t be here if it…
Alex Roy: But the thing is, Bryan, you’ve been so quiet, Argo‘s been so quiet for so long, it was hard to hang any myths on you, but we can try. All right, let’s just get into this. All right, first group is timing. Myth number one, yes or no.
Sam Abuelsamid: Yep, someday, maybe.
Alex Roy: CASE will happen by 2020.
Sam Abuelsamid: No.
Alex Roy: You know what CASE is?
Bryan Salesky: For our listeners, could you please define what CASE is.
Sam Abuelsamid: Connected, Autonomous, Shared and Electric. Or Electrified.
Bryan Salesky: UG.
Alex Roy: It’s what people say when they’ve been paid $10,000 to come speak in order to sound smart about what they think is going to happen.
Sam Abuelsamid: They pay 10 grand?
Alex Roy: $50,000. Have you seen some of these consultants?
Bryan Salesky: I’ve never once used “CASE.” I guess I’ve got to start.
Alex Roy: So just for the record, the CEO of one of the largest global AV companies does not know what the CASE acronym is.
Bryan Salesky: I didn’t say I didn’t know what it was, I said I don’t use it.
Alex Roy: You don’t use it. Oh, well, who would? So, just so we’re clear here, autonomous vehicles have to be connected, yes?
Bryan Salesky: They don’t have to be, no.
Sam Abuelsamid: It depends on how you’re using them.
Bryan Salesky: They’d be a lot smarter if it was.
Alex Roy: Do they need to be shared?
Sam Abuelsamid: If you only expect it to be autonomous while you are actually in the vehicle, and you tell it where to go, then yeah, you can get by without being connected.
Bryan Salesky: That implies connection.
Sam Abuelsamid: Yeah.
Alex Roy: Do they need to be shared?
Bryan Salesky: It doesn’t have to be.
Alex Roy: Optimally.
Bryan Salesky: But it would sure be a lot more cost feasible.
Alex Roy: Do they need to be electric?
Sam Abuelsamid: They need to be electrified in order to meet the power demands of these systems, because all those sensors, computers and all that stuff consume a lot of power.
Alex Roy: You couldn’t just throw a stack on a 1960s Restomod vehicle?
Bryan Salesky: Unlikely.
Alex Roy: Okay.
Sam Abuelsamid: But this depends.
Bryan Salesky: But depends though.
Sam Abuelsamid: If you put a big, honking alternator on it.
Bryan Salesky: Some of those are… have you seen the electric hot rods? There’s some pretty things out there.
Sam Abuelsamid: Yeah. This is an increasing segment of doing Restomods by electrifying them. Jaguar, they showed last year a concept version of an Old E-type that they converted to electric.
Alex Roy: The Eagle?
Bryan Salesky: Oh, no. It’s an official Jaguar car.
Sam Abuelsamid: Yeah, and then there was so much demand for it that they said, “Yeah, okay. We’ll build these things for customers if they want them.”
Alex Roy: Yesterday, a historic vehicle association announced that they are opposed to electrification of vintage cars.
Bryan Salesky: Who was that?
Alex Roy: I don’t know.
Sam Abuelsamid: I forget the exact name of the organization.
Alex Roy: Somebody said it, some bunch of jerks. Anyway, let’s move on. All right, so number two, myth of AV sector, that Tesla has it, and by it, fully self driving, today.
Sam Abuelsamid: No.
Alex Roy: Well, obviously not. I have one I could tell you it’s not, it’s dangerous. I love it, it’s just not autonomous.
Sam Abuelsamid: Well, let’s go back. Is the myth that they are selling it or that they actually have it available as a real thing?
Alex Roy: That you asked the question proves the point that it is a myth. All right, Bryan, you want to say anything?
Bryan Salesky: No.
Alex Roy: Okay, next, number three. I quote: “three months, maybe, six months, definitely.”
Bryan Salesky: It’s obviously a myth.
Sam Abuelsamid: It depends what you’re starting point is. Depends what T-zero is.
Bryan Salesky: There’s nothing deliverable in six months that is autonomous.
Sam Abuelsamid: At somewhere along an infinite timescale, there will be a T-Zero where three months, maybe, six months, definitely will be true.
Bryan Salesky: That’s true.
Sam Abuelsamid: As of today?
Bryan Salesky: No.
Sam Abuelsamid: It’s not true.
Bryan Salesky: No. In a few months, we could deploy at a very small, intimate, small section.
Alex Roy: When you say intimate, like a parking lot?
Bryan Salesky: No, I mean a street.
Alex Roy: A couple of blocks.
Bryan Salesky: I don’t know, a route.
Alex Roy: Okay, so you could probably shuttle employees around this campus?
Bryan Salesky: That’s right. You could stay in-lane, you could take all right turns, and so on. There’s ways to gain this in order to quote-unquote, be first, but that doesn’t mean-
Alex Roy: Not a business.
Bryan Salesky: That’s what mean. When we say scalable, we don’t mean we’re going to go deploy millions of vehicles tomorrow, what we mean is that we can actually build a meaningful business around it.
Alex Roy: This guy is so clear, it makes me sick.
Sam Abuelsamid: But the thing is, the guy who said that, a certain Mr. Elon Musk, who has made that claim at several different points over the last several years, what he has talked about is a Level 5 system, that you can just have it drive itself anywhere, anytime. That not only is not going to be in six months, that’s probably not going to be in six years.
Alex Roy: Alright, next myth-
Bryan Salesky: Can I interject?
Alex Roy: Of course.
Bryan Salesky: What’s your take, just a quick tangent here, on levels? Not to go down the rabbit hole again, but even today there was somebody who’s pretty, I would consider, well versed in the industry, and they were using Level 5 incorrectly. I just want to beat my head against the wall.
Alex Roy: In fact, someone who… very high ranking at one company is now very high ranking in another, but I’m not going to name him because I’m a nice guy.
Bryan Salesky: I’m not sure I know if it’s-
Alex Roy: No, it is.
Bryan Salesky: Sam, what do we do about this levels mess?
Sam Abuelsamid: Short of just dispensing with it entirely, I’m not sure that there’s a solution.
Bryan Salesky: Because the experts are not even getting it in some degree.
Sam Abuelsamid: I think SAE created that taxonomy, what, five years ago by engineers, for engineers. It was never meant to be consumer-facing language. It was never meant to be for marketing purposes. But even within that context, the original, and even up till now, it’s gotten slightly better, but the basic premise of these levels is fundamentally flawed because it’s not discrete steps, it’s not six discrete steps from zero to five. It’s a continuum. Because not everybody is doing the same thing, there’s an infinite number of possible configurations that could be considered Level 2, or Level 3 or Level 4. The only two that are end steps are Level 0, where you have nothing, and Level 5, where it’s capable of driving itself anywhere, anytime, any place. Anything in between is a continuum, and saying Level 4 does not… it doesn’t specify any particular capability or functionality, other than that it can operate without human input.
Bryan Salesky: Yeah. Thinking out loud here, I wonder if the answer is instead of trying to create prescriptive levels, that you actually have a set of predefined questions or something. That if you say you have an automated technology of some sort, you need to provide the answers to these kind of questions.
Alex Roy: So, “Is there a human in the loop? Yes, no.” Is that the first question?
Bryan Salesky: Yeah, exactly. It’d be on that order so that you… If you ask the average person, “Do you understand the operational design domain in which your automated vehicle is designed to operate within?” I don’t think anyone would have a clue how to get started. So the other thing we want to do is make sure that these questions are actually informative to a lay person.
Alex Roy: If someone asked me to explain What an ODD was, I would say, “Well, have you ever hung out with a baby?” because a baby’s operational design domain is very limited, and yet it is quite autonomous inside its domain.
Bryan Salesky: Right. Yeah. So anyway, I think there’s something there, but maybe that’s a whole other episode.
Sam Abuelsamid: Well, I think an example of why this whole premise of these levels is fundamentally flawed is a couple years ago, I was talking with an engineer from Nissan, who was in charge of their ADAS programs, and he’s on the committee that developed that technology. We were driving around and he told me right up, he said, “Level 4 is so broad. As an example of what a Level 4 autonomous vehicle is, the tram that runs up and down the McNamara Terminal at Detroit Metro Airport is a Level 4 autonomous vehicle. It’s design domain happens to be the track that it rides on, but there’s no human operator. If something goes wrong, it just stops wherever it is. It’s an extremely constrained domain, but that’s what it is.”
Bryan Salesky: Absolutely accurate.
Sam Abuelsamid: But a Level 4 can also be a Ford Fusion that drives itself around Miami within a geographic area, or based on environmental factors.
Alex Roy: Who would build such a thing. Guys, let’s get a move on.
Bryan Salesky: All right. Sorry.
Alex Roy: We have 40 more of these to get through.
Bryan Salesky: Yeah, let’s do it.
Sam Abuelsamid: 40?
Alex Roy: Yeah. Okay. Another alleged myth, and this is a fun one, especially for Sam and his Navigant leaderboard. Myth, or fact or fiction: that Tesla is dead last in self-driving tech.
Bryan Salesky: I saw Tesla on your board, it didn’t look good.
Sam Abuelsamid: Among the companies that I chose to rank, yes, I ranked them last because I think that the technology that they’re claiming will give them Level 5 capability is, I believe, fundamentally incapable of doing it, and their approach to the problem, I think it has some fundamental flaws to it. The way we do the leaderboards looks at more than just the technology but also their whole business model and whether the company’s even self sustaining. We factor all those things in, and-
Bryan Salesky: So you include go-to-market strategy.
Sam Abuelsamid: Yeah, go-to-market strategy.
Bryan Salesky: Support.
Sam Abuelsamid: Are they financially viable?
Bryan Salesky: Right. Service support of a Robo-taxi fleet is considered in that.
Sam Abuelsamid: Yeah, absolutely.
Alex Roy: Okay. Well, Bryan, does that sum it up.
Bryan Salesky: You don’t want to talk about your competition?
Alex Roy: But they’re not competition.
Bryan Salesky: Are you consider them competition?
Alex Roy: No.
Bryan Salesky: Me neither. Not for the type of technology that Argo is building.
Alex Roy: Mm-hmm (affirmative). It’s funny because when people ask me, “How can you have an opinion? You’re totally biased,” I’m like, “Listen. If you’d listened to what I’ve been saying for years… You don’t know me.”
Alex Roy: All right, let’s move on. Next one: your teenager will never need to learn to drive, because self-driving cars are almost here. Bryan.
Bryan Salesky: False. Unless your teenager is traveling-
Alex Roy: Through time?
Bryan Salesky: Unless your teenager is constricted to a geofence.
Alex Roy: Uh-huh. So literally chained?
Bryan Salesky: Not chained. If they’re in a city where some service is provided-
Alex Roy: Like Singapore.
Bryan Salesky: … and they’re able to get a ride in an autonomous vehicle, great, but they’re not-
Alex Roy: What if your kids are growing up in New York? In New York City.
Bryan Salesky: Right, but I assume most kids are going to be a little more portable than that, and want to be able to function in other areas.
Alex Roy: The thing about an AV in New York is no matter how great they are, if you’re heading north-south, the subway is really the thing to use.
Sam Abuelsamid: Well, that’s the thing. New York, you don’t have AVs today, but a pretty significant portion of the population of New York City doesn’t know how to drive, because they don’t need to.
Alex Roy: I live in New York, I would love there to be AV service that linked areas of Brooklyn to mass transit, because Red Hook and parts of Williamsburg, there are whole neighborhoods filled with amazing real estate that are starting to go up in value, but that are not served by transit. If I could link those neighborhoods to the F train or the L train, then whole parts of New York would be completely rejuvenated. It’d be awesome. I’d buy a house in Bed-Stuy or all the way out, and I would just take an AV to the subway. Please, Bryan, can you get on that?
Bryan Salesky: Sure, we’re working on it. What did you think of the little experiment in Manhattan where they closed down… what street was it?
Alex Roy: 14th Street.
Bryan Salesky: Yeah.
Alex Roy: The east-west busway.
Bryan Salesky: Did you hear about this Sam?
Sam Abuelsamid: Oh, yeah. I didn’t delve into the details of it.
Bryan Salesky: The reception was… They loved it. It’s awesome.
Sam Abuelsamid: Yeah. A lot of cities around the world have been doing things like that, taking areas and closing them off to individual-vehicle traffic.
Bryan Salesky: Yeah, I think it was met with quite a bit of skepticism at the idea, but then, once folks experienced it, they loved it. How long did it take you to get from one end to the other?
Alex Roy: Like 12 minutes, which is unheard of, otherwise it would take you an hour and a half.
Bryan Salesky: Right. Unbelievable.
Alex Roy: But the people who were against it were this peculiar cabal of friendless, crank crackpots. I didn’t even understand. I don’t know anyone who’s against it. Let’s keep going guys. How about this one. Myth, fact or fiction: that it’s available, autonomy, for purchase now or right around the corner.
Bryan Salesky: We’ve already covered this.
Sam Abuelsamid: Absolutely fiction.
Alex Roy: Fiction.
Bryan Salesky: Fiction, total fiction.
Alex Roy: That it’ll be here in five years?
Bryan Salesky: What?
Alex Roy: That’s the thing.
Bryan Salesky: That what will? Anything?
Alex Roy: I think we’ve already done this one, right?
Sam Abuelsamid: In five years, you may be able to buy some high-end vehicles that have some… what could qualify as Level 4-
Bryan Salesky: 3.
Sam Abuelsamid: … capability in highway driving.
Bryan Salesky: 3.
Sam Abuelsamid: Well, no. If it’s designed to actually be able to get itself into a minimum-risk condition without… If it doesn’t require a human to take over, if it can handle getting itself to a safe place if needed, then that becomes Level 4.
Bryan Salesky: Isn’t that really 3?
Sam Abuelsamid: No, because Level 3 implies that a human has to take over control.
Bryan Salesky: With a transition period.
Sam Abuelsamid: Right.
Bryan Salesky: When it’s told to.
Sam Abuelsamid: If it doesn’t need a human to step in, then it’s Level 4.
Bryan Salesky: But at some point—.
Sam Abuelsamid: That’s your operating domain.
Alex Roy: Listen. If the highway is the domain, but at the end of the domain, it says, “You’re taking over when you get out this off ramp,” isn’t the car globally defined as Level 3? This is the mystery, isn’t it.
Sam Abuelsamid: Yeah, this is why the levels are so flawed.
Bryan Salesky: Okay. I can’t believe we just put the two of you in a conundrum over the levels. This is great. Yes, technically it would be Level 3.
Alex Roy: All right, this next one is targeted at Bryan Salesky. This one’s going to make him crazy. Okay. “Progress towards autonomy will be exponential.” Only someone who claims that computer vision plus radar is sufficient, and that a fleet of cars, would say that. Is anything ever exponential?
Bryan Salesky: Well, I don’t know. If you step back and squint at it, there certainly has been a deceleration in-
Sam Abuelsamid: Depends which part of the curve you’re looking at.
Bryan Salesky: That’s exactly right. Yeah. Well said, Sam.
Alex Roy: So, that could be true?
Bryan Salesky: Again, it depends on which part of the curve you’re looking at, and how far you’re zooming out or in.
Sam Abuelsamid: It’s the classic case of, “Tell me which side of the argument you’re on, and I’ll give the statistics to proves that you’re right.” It’s all in the context, if you look at a certain part of the development curve, you can find a segment in there that is an exponential curve going up. You can look at other parts of the same curve where you see exponential development going down.
Bryan Salesky: Let me make this a little more tangible. We’re in an era where storage is cheap, there’s tons of processing power that is also coming down in price, we have better and better sensors, we have the ability to… Cloud processing is much more accessible to companies than ever before. You add all this together and it means that we can make dramatic progress much easier over a much smaller period of time than ever before.
Bryan Salesky: If you look at just this last, I don’t know, 5- or 10-year period, it might look exponential, it’s pretty serious in terms of the rate of improvement compared to where it was, say, 20 years ago. But again, I’m sure folks have more history than I do from, I don’t know, seeing computer… the evolution of operating systems or something coming out of the ’70s. I’m sure they would see, depending on which part of the curve for that, you would see whether it’s accelerating or not. I think there’s no question that, in recent time, there’s been a significant acceleration because of all those threads coming together.
Sam Abuelsamid: One of the things I learned early in my engineering career, one of the classic idioms of the field, is that every project, the first 90% of the project takes 10% of the time, and the last 10% takes 90% of the time. The exponential part typically only comes in that first 90%, but it’s that last 10%, that’s where the really hard problems lie. When we drive around, as humans, most of the time we’re absolutely fine, most of us don’t get into crashes all the time. We accumulate a lot of miles every year without crashing, but it’s those corner cases, that’s where the hard problem is for humans. It’s also where the hard problem is for automation. They may represent only a small fraction of the total operating time, but they are the toughest ones to figure out.
Alex Roy: That was fun! Okay. Next group, ownership myths: that no one will own a personal vehicle.
Bryan Salesky: False. I love to drive. People who want to drive should still be able to buy a car.
Sam Abuelsamid: Was it no one will own a personal vehicle, or no one will own an autonomous vehicle?
Alex Roy: No one will own a personal vehicle.
Sam Abuelsamid: Yeah, that’s false.
Bryan Salesky: Totally false.
Sam Abuelsamid: They’re going to pull my Miata out of my cold, dead hands.
Bryan Salesky: I think there’s a lot of people that would speak to that.
Sam Abuelsamid: It may be wrapped around me, but—
Bryan Salesky: So it turns out that you’ll probably still need somebody to bring their tool truck to your house to fix things. I think a lot of people, when they hear “car,” they just think of the passenger vehicle, personally-owned vehicle. Let’s not forget the commercial aspect to this as well. Car ownership’s not going away, ever, or at least anytime soon.
Sam Abuelsamid: Not in it’s entirety.
Bryan Salesky: Not in my lifetime.
Sam Abuelsamid: Yeah. For a lot of people, this we talked about earlier: in New York today, a large proportion of the population of cities like New York don’t drive because they don’t need to, they have access to other modes of transportation that are more convenient and more cost effective. That will always be the case, just as it will always be the case… or at least for a very, very long time, it will be the case that there will be a lot of people that own their own vehicles and operate their own vehicles.
Alex Roy: I agree, because I want to.
Sam Abuelsamid: Well, as the founder of the Human Driving Association, presumably you will.
Alex Roy: Of course I agree. Okay. Next, that people outside of megacities will voluntarily give up ownership of private vehicles. That’s really the same myth.
Bryan Salesky: It’s very specific.
Alex Roy: Yeah. All right, we’ll move on. Okay. Safety. 94% of accidents are caused by humans, computers aren’t human, therefore there won’t be accidents. Will there be accidents?
Sam Abuelsamid: Humans are a factor in 94% of crashes.
Bryan Salesky: Factor.
Sam Abuelsamid: Crashes. Not accidents, crashes.
Bryan Salesky: There are no accidents.
Sam Abuelsamid: But there is rarely ever just a single factor that leads to a crash.
Alex Roy: And the other factors, I know what they are, but I want to hear it from one of you gentlemen. What are the other factors that go into a crash besides the human in that 94%?
Sam Abuelsamid: The environmental conditions, the weather, the road conditions, visibility, other-
Bryan Salesky: Other drivers.
Sam Abuelsamid: Other vehicles around you. Yeah. Pedestrians or animals.
Alex Roy: All right. So, in a world of increasing automation and autonomous vehicles, there are still going to be crashes. Correct?
Sam Abuelsamid: Yeah.
Alex Roy: Because it’s a complex world.
Sam Abuelsamid: Well, yeah. For one thing, nobody’s ever created completely bug-free software.
Alex Roy: Right. Or humans.
Sam Abuelsamid: Or humans. All sensors have limitations, all actuators have limitations, all computers have limitations. Yes, there will still be things that happen.
Alex Roy: Okay, Bryan, anything you’d add to that?
Bryan Salesky: I think it’s great.
Alex Roy: Alright. Myth: no more accidents, ever. We’ve done that. Oh, let’s see. That they will substantially increase efficiency of commutes and decrease congestion, fact or fiction?
Bryan Salesky: The self-driving system alone cannot solve congestion. Fiction. An intelligent, networked fleet of vehicles that are being directed through some sort of system that knows how to load-balance the road network, that can help congestion.
Sam Abuelsamid: And efficiency as well.
Bryan Salesky: And efficiency.
Sam Abuelsamid: By filtering out a lot of the spurious inputs that you have from human drivers… you hit the accelerator, you may over…you may give it a little too much at first, and then back off—
Bryan Salesky: Which even equates to better fuel economy or range. So, absolutely.
Sam Abuelsamid: Yeah. Just by making it smoother and managing that congestion, you can reduce energy consumption, as well.
Bryan Salesky: Absolutely.
Alex Roy: All right. “AVs will help the environment.”
Sam Abuelsamid: I mean to the same degree we just said, it can help the environment if they’re properly managed and controlled. But it could also be much worse, because as we said earlier, AVs also require a lot of energy to operate.
Bryan Salesky: This, again, goes to a single AV, probably not. But if we switch to a shared fleet of vehicles, we shouldn’t need as many vehicles in these cities. We shouldn’t have vehicles circling over and over again looking for parking spots. There are efficiencies to be gained there that I think do benefit the environment, but it’s going to take more than just one.
Sam Abuelsamid: It’s been estimated that about 20% of the energy used for ground transportation in cities is just spent during that time spent looking for parking.
Bryan Salesky: Exactly.
Alex Roy: Okay. “That AVs will be affordable,” but isn’t that the point of AVs? They’ll be more affordable, like a bus. More affordable-
Sam Abuelsamid: It depends on the context. If you’re talking about individual-use AVs, where you own it like you do today, probably not, at least not for a long time. But if you’re talking about shared-
Bryan Salesky: You’re spreading the cost over the lifetime of the vehicle.
Sam Abuelsamid: Fractional ownership, yeah.
Alex Roy: Okay. Let’s move on to the trucking myths. Will trucking jobs vanish? Fact or fiction? Does autonomy just eliminate those jobs, yes or no?
Sam Abuelsamid: Not entirely, and certainly not in the near-term.
Bryan Salesky: Yeah. I think that’s the key is with all of this, obviously automation changes the nature of the jobs and the roles, the skills that we need, but these are over very long time horizons. This is not some imminent thing.
Alex Roy: Every time I’ve spoken to someone in trucking, whether they were in [inaudible 00:26:31] trucking companies or startups, they all say the same thing. They all say there’s a shortage of truck drivers, and therefore, if-
Sam Abuelsamid: Especially for long-haul trucking.
Alex Roy: Yeah. So autonomous tech, if that could help round out that shortage, that the trucking industry could in fact grow human jobs overall, and improve efficiency of trucking operations.
Sam Abuelsamid: Yeah, by utilizing human drivers where it makes the most sense, for those long-haul routes where you’re just running hours on end down an interstate. Part of the reason why there’s a shortage is because most people don’t like to do those jobs. It’s a tough job.
Bryan Salesky: Yeah. It’s a reallocation of where the workforce is putting their energy. That’s the way I look at it. Look, I was talking to somebody recently, they said, “Boy, for a self driving car company, you sure do have a lot of people.” We need testers, mappers and people that maintain the vehicles and so on. This is, I think, something that people don’t broadly understand is that these are job creators in a huge sense. Then, as they start to scale out, yes, you need fewer and fewer drivers, but they’re getting employed to do other things.
Alex Roy: My observation is that many of the people coming in, especially in operations, test specialists, people on the ground who are opening new cities and markets, are coming out of jobs that were not very technical, and moving into very interesting and technical positions. It’s going to take decades to scale out to ubiquity.
Bryan Salesky: Absolutely. The nature of the skills required is going to change and we need to be doing things to make sure that we’re training the workforce appropriately over time.
Alex Roy: All right. Here we go. Now we’re getting some controversy. Myth, fact or fiction: AVs need regulation.
Sam Abuelsamid: I believe that at some point in the not-too-distant future we’ll need some performance standards for AVs. Just as we have licensing standards for human drivers today, we need the technological equivalent of that for AVs in order to verify that they are in fact safe enough for use on public roads.
Alex Roy: Do you have an opinion on this?
Bryan Salesky: I agree with Sam. I think that’s necessary. I also think that a unified framework would be certainly helpful so that when we cross the border from one city or town to the next, we don’t have a separate set of rules that need to be adhered to. I think there’s just some practical things that need to be put in place to enable the deployment of the technology.
Alex Roy: Okay. Business. Here we go. Fact or fiction: tech companies will put Detroit out of business.
Bryan Salesky: Absolutely not. Fiction. Detroit’s going to thrive on this. They’re going to be known as the mobility capital of the world.
Sam Abuelsamid: I agree.
Bryan Salesky: 313.
Sam Abuelsamid: Let me just add to that. Tech companies and the auto industry have actually been collaborating for decades.
Bryan Salesky: Absolutely.
Sam Abuelsamid: Since the 1970s, when we first-
Bryan Salesky: This whole Silicon Valley versus Detroit just drives me nuts. It’s complete nonsense. Sorry, Sam. Go ahead. Carry on.
Alex Roy: Easy for you to say as the company who’s thriving off of these partnerships, because the next myth, you’re going to absolutely die laughing.
Bryan Salesky: Oh no, here we go. Sorry, Sam. Go ahead.
Sam Abuelsamid: Since the 1970s, when we implemented fuel economy and emissions regulations, and then safety regulations, we’ve been putting more and more electronics in vehicles. The industry could not be where it is today were it not for the tech industry.
Alex Roy: Well, you can tell that to the people who designed the Buick Reatta Infotainment system, which is not so good.
Bryan Salesky: Let me tell you, Detroit is one of the most resilient cities around, period. There’s more grit and determination in this city than most others. You can’t put them down, it’s not happening.
Alex Roy: Woo! Go Bryan.
Bryan Salesky: It ain’t happening.
Alex Roy: Okay, moving on. Fact or fiction: it makes sense to build the autonomous vehicle stack AND the car from scratch.
Bryan Salesky: Oh my goodness.
Alex Roy: These aren’t my suggestions.
Bryan Salesky: I’m not even taking this one.
Alex Roy: Somebody put it in my Twitter.
Sam Abuelsamid: Again, it depends on the use case, but in general, the more integration you can do, the better a system is going to be.
Alex Roy: Yeah, but building it, designing it, engineering it, and building it and manufacturing it from scratch yourself, everything, is that a good idea?
Sam Abuelsamid: Well, if what you’re talking about… and it doesn’t necessarily need to all be done in house, but-
Alex Roy: And operate the platform. Everything.
Sam Abuelsamid: Again, it doesn’t necessarily have to be.
Alex Roy: Okay.
Sam Abuelsamid: I think if you’re going to do, for example, something like a Robo-taxi or an autonomous delivery vehicle, designing that vehicle purpose built to operate autonomously and to be able to handle all of the use cases that it needs to, because it’s not just as simple as slapping a box of sensors onto the roof of the vehicle on a computer in the trunk. That doesn’t help you get a pizza from the car to the doorstep.
Bryan Salesky: Look, even a modern automobile company does not design, build, manufacturer, all on their own, their vehicles. There’s an enormous supply base necessary. You have to partner to be successful.
Sam Abuelsamid: Yeah. When I say integration, that doesn’t necessarily mean vertical integration within one company, but it means that everything, all the pieces have to be integrated and work together seamlessly.
Bryan Salesky: You have to partner to be successful, period.
Alex Roy: Okay, put that on your next T-shirt. Okay, here we go. Now the technology group of myths. Fact or fiction: “computing efficiency, in terms of MIPS/watts, exists to run a car autonomously.” You want to translate that? What do you think this guys?
Bryan Salesky: True.
Alex Roy: It does exist?
Bryan Salesky: Yeah.
Alex Roy: Alright, so it’s not a myth?
Bryan Salesky: No.
Alex Roy: Okay, great.
Bryan Salesky: [inaudible].
Alex Roy: You would know.
Bryan Salesky: You threw me a curve ball. We were climbing through all these “no, no, no, fiction.”
Alex Roy: Okay. Well, hey, listen, we went from a 100 suggestions down to 40-
Bryan Salesky: But there are caveats.
Alex Roy: Yeah, you want to tell us what they are?
Bryan Salesky: I’m biting my tongue on levels again.
Alex Roy: But it’s why you want to use an internal combustion or a hybrid today, right? Because the power…
Bryan Salesky: Absolutely. As Sam said earlier, it needs to be electrified in order to power these things that can be in excess of a kilowatt.
Alex Roy: As opposed to steam engines.
Bryan Salesky: Right.
Sam Abuelsamid: Well, often the steam engine is what powering a generator.
Alex Roy: So a pure electric vehicle today is probably going to be suboptimal for the long work hours required to operate a-
Bryan Salesky: It depends on the battery package and the range, and if it was designed with the intent of this huge load that gets put on it by having a self driving system equipped on the vehicle. I would say that most electric vehicles that are on the market today would be challenged, range-wise, to be able to do a full day’s work with a self-driving stack on them without recharging.
Sam Abuelsamid: Yeah. If you designed a purpose-built EV for this kind of application where-
Bryan Salesky: If.
Sam Abuelsamid: For example, instead of relying on charging several times a day, you instead went to battery swap, and you’re operating a fleet, now you could come in and swap the battery in a couple of minutes, the vehicle goes back on the road, battery goes into a rack where it’s decoupled from the use of the vehicle, you can slower charge it, and because you’ve got a fleet of identical vehicles, they all have the same battery.
Alex Roy: If only someone had tried to build such a business in Israel some years ago, and learn that this is problematic.
Sam Abuelsamid: It’s problematic for the consumer application, but not for fleets. Fleets are a very different use case. When you’re talking about high-utilization fleets, now it’s an entirely different ballgame. If you’ve got 1000 AVs with swappable batteries operating in a place like San Francisco or Miami, now all of a sudden it starts to make economic sense because the area that they’re operating in is limited. You’re not trying to deal with individual-use case where people want to be able to drive anywhere, anytime they want.
Alex Roy: It sounds like you concur with that, Bryan.
Bryan Salesky: Yeah.
Alex Roy: He’s got the concurrence face. Okay. Next: AVs will learn from human drivers, fact or fiction? Well, this is really targeted at Tesla, isn’t it, because that’s kind of a Tesla selling point. Every car on the road teaches their system. Every human operator of a Tesla teaches the system something. Does that make sense?
Sam Abuelsamid: Technically, I think every AV system is learning from human drivers.
Alex Roy: Through observation of third parties.
Sam Abuelsamid: Yeah, you’re training it based on the way your test drivers are operating with vehicles, and…
Bryan Salesky: I think we have to unpack this a little bit. It isn’t necessarily learning from the input into the steering, gas, brake. It’s learning more from the observations of the movements of all the different relevant objects in the world, and it’s using that to build up experience about how they’re likely to move in the future. That’s helping to train a whole set of algorithms that helps us understand the world through our sensors. There is a class of problems where it is beneficial to see how would a human react to certain situations, and to build that into the motion planner. Some of that is called imitation learning, and that can be helpful for certain specific maneuvers, but it’s not something that is applied, typically, broadly speaking. Sorry, was that too… Did I get technical?
Alex Roy: That’s a wonderful answer.
Bryan Salesky: All right.
Alex Roy: All right. How about this one: fact or fiction? This is very easy. Platooning.
Sam Abuelsamid: It’s a fact that you can platoon.
Bryan Salesky: Yes, it can be done.
Alex Roy: I’ve seen platooning in a CGI video. It looked great.
Sam Abuelsamid: There have been tests that have been run. There’s companies that are trying to commercialize it. For a highway operation, there are some real, potential fuel-saving benefits to doing it.
Bryan Salesky: I think it’s very real, and I think-
Sam Abuelsamid: But there’s also a lot of challenges in implementing it.
Bryan Salesky: That’s right. I think businesses that run huge fleets, and where they have a pretty big density of trucks on a particular stretch of road, can definitely see a lot of savings. I just rattled off a bunch of ifs, right?
Alex Roy: A lot of ifs.
Bryan Salesky: It’s one of those things that works for some businesses that truly have large-scale operations on dedicated routes. It probably doesn’t work so well for… I shouldn’t say “doesn’t work.” It would work, but it would require many of the vehicles in the fleet to actually have the technology so that you can pair with that car and go into some sort of platoon mode. That suggests that it needs to be rolled out in a pretty wide basis. So yeah, this is one of those things in theory, yes, but in practice, bringing it to the market, I think is very challenging.
Alex Roy: Okay. Fact or fiction: the difference between driver assistance and AVs is incremental.
Sam Abuelsamid: Depends on how big an increment you’re talking about.
Alex Roy: The whole thing.
Bryan Salesky: How big are your increments?
Sam Abuelsamid: Let’s skip this one.
Bryan Salesky: Arggggg.
Sam Abuelsamid: How big are your slices?
Alex Roy: All right. Fact or fiction: Lidar is a fundamental part of an autonomous car.
Sam Abuelsamid: Yes. I agree that it is one of many fundamental parts of an AV.
Bryan Salesky: Yes. Everything is fundamental when it’s safety critical.
Alex Roy: Okay. Well, that was easy. We’ve only got a few left. This has been a lot of fun. What, Bryan? You don’t enjoy this? Come on.
Bryan Salesky: No, it’s great. I loved it. This is fun.
Alex Roy: If we could just send this episode to every person who’s got a list of questions and be like, “Okay, here we go.”
Bryan Salesky: That’s why we are doing this. That’s a good idea.
Alex Roy: Okay. Here we go. “Privacy will not be a problem in AVs.”
Sam Abuelsamid: If you’re talking about shared AVs?
Alex Roy: Well, I get in the back of an AV… Have you seen Taxi Confidential? That’s a show, Taxicab Confessions.
Sam Abuelsamid: Yeah.
Bryan Salesky: Maybe.
Alex Roy: Will there be a camera-
Bryan Salesky: That’s not Cash Cab, is it? That’s different.
Alex Roy: Yeah, this Cashless Cab. That’d be a great name for the platform, huh? Cashless Cab.
Bryan Salesky: Not necessarily. I just saw a big thing on the controversy over [inaudible 00:39:47] from state to state, actually. That’s a very serious topic, Alex Roy.
Alex Roy: Well, it sounds like there could be a whole ‘nother episode for that one. You want to say something, Sam?
Sam Abuelsamid: Yeah. I think that it is something that is going to be a major concern for people, because I think, especially when you start talking about things like Robo-taxis, people have not necessarily considered the fact that there are going to be sensors inside these vehicles monitoring who’s in the vehicle, how many people are there, have they left anything behind, have they thrown up in the vehicle? So there will be no privacy in these vehicle.
Alex Roy: It’s funny because, if you think about it, if you get into a new car right now with a human driver, you also have no privacy. So I think this is all big—
Sam Abuelsamid: Well, if I leave this phone sitting on my kitchen counter, go get into my Miata and go for a drive, nobody knows where I am or what I’m doing.
Alex Roy: Well, you could also just keep your Miata, take an AV from time to time, and then drive your Miata.
Sam Abuelsamid: Well, that’s what I do. Well, except for the AV part.
Alex Roy: This guy is smart, this guy, Sam.
Bryan Salesky: Why do you like the Miata?
Alex Roy: If you have to ask, Bryan, like jazz, you’ll never know.
Bryan Salesky: I love asking the Miata owners because the tech crowd…there’s a large number of people who love these—
Alex Roy: And the Honda S2000s, too.
Bryan Salesky: Yeah, that’s correct. Absolutely.
Sam Abuelsamid: Because it’s small, lightweight, it’s very nimble. Modern performance cars are amazing in terms of what they can achieve, the acceleration, the handling, but their thresholds are so high that you can’t approach those thresholds and you can’t exercise your own skills at anything resembling a legal speed. There’s another classic idiom in the auto world, it’s better to drive a slow car fast than a fast car slow, and the Miata is the ultimate example of that.
Bryan Salesky: It’s the fastest slow car?
Sam Abuelsamid: It’s the best. Yeah.
Bryan Salesky: Do you have a Miata, Alex?
Alex Roy: No, but I’ve driven one and I loved it.
Bryan Salesky: Really?
Alex Roy: Years ago I went to F1 school, and I was not very good.
Bryan Salesky: That’s shocking.
Sam Abuelsamid: I can’t wait to hear the episode about your Argo driver training.
Bryan Salesky: Yeah, it’s coming.
Alex Roy: Then I drove a Miata for a bit, I love that, and I met some folks in Monaco. I was in Monaco, I met this woman, and she’s bragging to me about her driving skill. She asked, “What’s your favorite car?” I said, “Well, I really like Miatas.” She made this face of disgust, and that became my litmus test for whether or not I trust people’s judgment on cars.
Bryan Salesky: It was probably the wrong town. Is Monaco everything it’s cracked up to be?
Alex Roy: It is exactly what you think it is.
Bryan Salesky: You know what you’re getting. Okay.
Alex Roy: If that’s what you want, you’re going to love it, but if you hate that, you don’t ever need to go. By the way, the most popular vehicle in Monaco, with people I know who do like Monaco, cars and coffee, who collect cars, is a Renault Twizy. It’s the tandem electric little thing, because it’s the only way to get through traffic and there’s dedicated Twizy spots. Everyone with money in Monaco has Renault Twizy, they love it, and they think that everyone else is just an idiot.
Bryan Salesky: So, I’m addicted to these shows on MotorTrend, the MotorTrend Channel. I highly recommend it. There’s a show, Bitchin’ Rides or something, the guy had a… I know! All the shows are named like that, near as I could tell. But he had a giant, giant bus-looking thing, and it said General Motors Parade of Progress. It had a jet engine in it kinda thing. Do you know what I’m talking about?
Sam Abuelsamid: Yeah. It’s from the 1930s.
Bryan Salesky: Yeah, what is this thing? But they redid it. It’s beautiful.
Sam Abuelsamid: Yeah. They built a few of those in the 1930s. It was part of the GM Motorama shows that they used to do.
Bryan Salesky: Is that where it came out of? Okay.
Sam Abuelsamid: Yeah. They had the side flip down, there was a big display thing inside of it.
Bryan Salesky: This was out of the World’s Fair?
Sam Abuelsamid: Yeah.
Bryan Salesky: Futurama or… No, that’s a cartoon.
Alex Roy: I like what I’m hearing.
Sam Abuelsamid: I think the first one was from one of the World’s Fairs, and then they used them for a number of years.
Bryan Salesky: Right. Okay.
Alex Roy: Okay. We’ve only got a couple left.
Bryan Salesky: I love talking about car stuff.
Alex Roy: Well, we can keep…
Bryan Salesky: All right, never mind. Go ahead.
Alex Roy: All right. Fact or fiction: autonomous urban air travel for the masses is coming.
Bryan Salesky: That’s true. eVTOL.
Sam Abuelsamid: Yes.
Alex Roy: Well, okay. Well, that’s—.
Bryan Salesky: The electric vertical takeoff and landing—
Sam Abuelsamid: Keep in mind, you did say for the masses.
Alex Roy: For the masses.
Sam Abuelsamid: I think that is a fiction.
Bryan Salesky: That’s the issue. The masses part.
Sam Abuelsamid: One of my favorite recent stories that I read, when Uber launched their service between Manhattan and JFK, their helicopter service, the New York Post sent out two of their reporters from their Midtown office to see who could get to JFK faster. One reporter took the subway, and the other one took lifts and Ubers, or took an Uber to get to the heliport then to take the helicopter and so on. Basically, the person who took the subways got there three minutes faster.
Bryan Salesky: That’s a great story.
Sam Abuelsamid: And it costs like $6 instead of costing $300.
Alex Roy: I got to confess something guys, an earlier incarnation of Alex-
Bryan Salesky: You owned a helicopter?
Alex Roy: In 2002 or ‘1, maybe, they had a thing called New York helicopter, and it was like $199 to get to JFK, and I was going to miss my flight. So I went down there and took it, and I did everything, like, “I’ve got this all calculated. I’m going to make it.” I got there, and I didn’t realize that it drops you off at Terminal 4 at JFK, and to get from Terminal 4 to JetBlue took about 40 minutes and I missed my flight!
Bryan Salesky: Ah, mobility.
Alex Roy: Okay. All right. Here we go. Fact or fiction: the trolley problem should be used to discuss autonomous vehicle ethics.
Bryan Salesky: False.
Sam Abuelsamid: False.
Alex Roy: Okay. Do you want to talk about it or is just a blanket no?
Sam Abuelsamid: Well, I think, in general, the challenge of the ethics-
Alex Roy: Is to say no.
Sam Abuelsamid: … of programming ethics into a vehicle is that is one of those really minuscule edge cases. The probability of something like that happening, compared to all the other things that can go wrong, is so far… that’s your lowest priority.
Bryan Salesky: There tends to have been a number of things that have likely failed prior to even getting entrapped in such theoretical decision dilemmas.
Sam Abuelsamid: My latest column that I submitted a couple weeks ago for Autonomous Vehicle Engineering Magazine, SAE’s magazine, is about ethics. The ethical problem we need to be dealing with is not the trolley problem, it’s about the ethics of how we go about deploying and testing these vehicles [crosstalk 00:46:38] the public.
Bryan Salesky: That was a good article. I saw that.
Alex Roy: You and Bryan should go on a date sometime. You guys really got a lot to talk about. Okay, last one. Fact or fiction: “Our platform is ready and better than all the others, and we can demonstrate it if you give us $200 million.” So I think what they’re really suggesting here is that $200 million-
Bryan Salesky: Did this come on your Twitter feed?
Alex Roy: Yeah.
Bryan Salesky: Oh, okay.
Alex Roy: It was in quotes, like they’re doing a fake quotation. I guess what they’re asking is whether $200 million will ever be the difference between completing-
Sam Abuelsamid: Well, what you just read was, “We can demonstrate it if you give us $200 million.” Demonstrate it and prove it to us first, then we’ll give you the 200 million.
Alex Roy: There are people out there who will do it for $2 million.
Sam Abuelsamid: Yeah.
Alex Roy: All right. It looks like we’ve got our 10. Actually, it’s 11, and here they are. I’m going to wrap it up. CASE will be here by 2020: myth. Tesla’s in the lead: myth. Your teenager will never need to drive: nope. No one will ever own a personal vehicle: not true. Crashes will be 100% eliminated: not true. Trucking jobs will vanish: not true. AVs don’t need some type of regulation: we agree that’s not true. Tech companies will put Detroit out of business: not true. It makes sense to design and build everything: not true. Flying taxis for the masses are coming, and the trolley problem is of value. That’s 11.
Sam Abuelsamid: Is there anything that is true about any of these?
Alex Roy: They’re hard.
Sam Abuelsamid: It’s a tough problem.
Bryan Salesky: Which I keep getting… I get shit for saying that now.
Alex Roy: But it’s true.
Sam Abuelsamid: You’ve been saying that since the beginning.
Bryan Salesky: It’s just the honest answer. But that doesn’t mean they’re not going to get deployed or have a huge impact, it’s just going to take a while.
Alex Roy: Do either of you have any other myths you’d like to add to this list of all-time craziest myths? Because that pretty much… that’s a good list.
Sam Abuelsamid: I think we covered a lot of ground there. I think we’re good.
Bryan Salesky: I think we scorched the earth on this one, Alex.
Alex Roy: Awesome. Well, that was really delightful. Sam, thank you so much for coming.
Sam Abuelsamid: Thank you for inviting me. It’s always a pleasure to talk to you guys.
Bryan Salesky: Same here, Sam. Same here.
Alex Roy: I really enjoyed doing that with Sam.
Bryan Salesky: I did too. I think we tackled so many different myths. The key thing here in that whole dialogue is we aren’t talking about 50 years from now, 20, 30 years from now, we’re just talking about the next few years and trying to set it into perspective of what is likely to happen and what’s unlikely to happen. We’re not trying to say that some of these things never become a reality, it’s a time-phased thing.
Bryan Salesky: I would love to have it, and it will happen eventually. It’s a timing thing.
Alex Roy: Antigravity Jetsons car.
Bryan Salesky: Also would absolutely love to have that.
Alex Roy: Yeah, I would love to have that too. In fact, we should watch The Jetsons and do Mystery AV Theater 3000.
Bryan Salesky: That’s a great idea. I love it.
Alex Roy: All right. Well, if you want to learn more about the No Parking Podcast, you can check us out at noparkingpodcast.com. Please sign up for our newsletter. We’ll be keeping you updated as to upcoming guests, events and special things we’re doing. You can follow us on Twitter @noparkingpod.
Alex Roy: Brian, you don’t do social media, do you?
Bryan Salesky: I’m real busy.
Alex Roy: All right, you can follow me on Twitter @AlexRoy144. Of course, if you’d like to be a guest on the No Parking Podcast or suggest to us any ideas for things we should cover, please contact us, really me, at firstname.lastname@example.org. See you next week.