In our final episode of the year, we bring back automotive analyst Sam Abuelsamid for a year-end, mythbusting wrap-up with Bryan and Alex. Was 2020 ever going to be the “year of self-driving?” What should we make of industry consolidation? Are we past ride-hail, or could it be reborn? No Parking trades takes with Sam on this year’s biggest news and hottest trends, plus makes a few predictions for what to expect in 2021.

Listen On
Apple
Google
Spotify
iHeart Radio
Share

View Episode Transcript

Episode Transcript

Alex Roy

Hey everyone. This is No Parking, the podcast that cuts through the hype around self-driving technology and artificial intelligence. I’m Alex Roy with my cohost roboticist, Bryan Salesky. Hey Bryan.

Bryan Salesky

Hi Alex. How are you doing today?

Alex Roy

Well, I’m totally excited because today we’re going to do a 2020 wrap-up, talk about trends and myths, and make some predictions.

Now, I love predictions because one of my favorite things growing up were all those really cheesy Nostradamus documentaries and movies, and one of them had this amazing actor, Tchéky Karyo. He was the spymaster in the original La Femme Nikita. And this guy plays Nostradamus, I think it was in like the early nineties. And in that movie, I mean it’s not a good movie, he spends the whole movie making these horrible predictions, these visions of just terrible disasters. But the very end is a beautiful shot where he has a vision of spacecraft arriving at an alien planet and in the windows you can see human beings. And he starts crying because he knows that, despite all the things that are bad that happened, even in his visions, happier days await us. I love that optimism and that’s why I love trying to predict things, because if you take a step back and you have some common sense, you know, that no matter what happens on a long enough timeline, people figure things out. Interestingly, Abraham Lincoln said, “The best way to predict the future is to build it.”

So to talk about 2020 trends, myths and predictions, Bryan and I are happy to have our friend, Sam Abuelsamid, back on the show. His official title is Principal Research Analyst for E-Mobility at Guidehouse Insights, but he’s also an engineer, automotive journalist, and one of the most respected analysts in transportation, and I don’t often have nice things to say about analysts.

Sam, welcome back to the No Parking Podcast.

Sam Abuelsamid

Hey, it’s always great to be back with you and Bryan.

Bryan Salesky

I’m really excited for this, Alex. That was a lovely introduction. Here’s my view of it. This is Myth Busters Round Two. This is an annual tradition now. Let’s face it. This is like turkey. This is like a, you know, hot chocolate in the winter. This is just, this is what we do. So I’m really excited for this.

Alex Roy

It’s also my favorite episode of the year. So let’s dive in. One of the most, I guess, biggest trends every year, going back many years is the statement that “2020 was going to be the year of self-driving.”

Bryan Salesky

Was this really a thing Sam? Like, was this really? Is it, do you remember that? Was 2020 supposed to be the year?

Sam Abuelsamid

I seem to recall Travis Kalanick being on a stage back in about 2015 or ’16 saying, “By 2018, we’re going to have automated vehicles out there running on our platform… Going to get rid of the human drivers. These things are an existential threat to our business, so we’ve got to do it.” It worked out so well for Uber, hasn’t it? I mean, they’ve been out there for two years, right?

Bryan Salesky

Exactly. And then Mark Fields came out with a 2021 projection and everybody said, “Oh, he’s so slow, they’re so behind, they’re going to get crushed.” And here we are.

Alex Roy

Do you guys remember that there was a headline, I think it was in 2016, that Uber had placed an order for a 100,000 autonomous Mercedes S-Classes?

Sam Abuelsamid

It was actually a little bit different from that they had placed an order with Volvo for, I think it was about 24,000 vehicles, but they also did a separate deal with Mercedes, where Mercedes was going to bring their autonomous S-Classes to the Uber platform. So Uber wasn’t actually buying those. They were just going to have a Mercedes run their vehicles using the Uber platform.

Bryan Salesky

You do have to say though, Alex, they do have good taste.

Alex Roy

Of course they have good taste. What possible deal could have been behind that statement because was there ever a program to make those vehicles?

Bryan Salesky

What possible data did they have that said that was actually a good idea given where the technology was at the time? I mean that’s where the PR drives the strategy, drives the order volumes, which is totally disconnected from, sort of, reality.

Sam Abuelsamid

At the time, you know, I think Mercedes was, they definitely saw themselves having level four capable S-Classes available in a 2020/2021 timeframe. And they said, “Yeah, sure. We’ll put some of these vehicles on your platform.” I mean, this was four or five years ago now almost, that they made that deal. And, you know, it was a perfectly plausible thing to have, to deploy them wherever you could. And the Uber platform is the biggest ride-hailing platform made perfect sense to do that. I don’t know that there was ever, I think it was more like a memorandum of understanding rather than any kind of real contract to do it because nobody actually had any of this stuff. None of this stuff actually existed yet.

Alex Roy

Is it even possible to make the statement that any year is the year of self-driving? It just doesn’t seem like a statement that makes sense.

Sam Abuelsamid

Well, I think you can certainly make the statement. Whether you’ll be accurate or not is an entirely different matter. Many people have made those statements over the years, myself included. Although, I don’t think I’ve ever gone quite to that extent that any particular year would be the year of self-driving.

Alex Roy

What’s the year of refrigeration? What’s the year of aviation? Like what year was that?

Bryan Salesky

But how about the decade of self-driving? I think we are going to see more and more pilots and we’re going to see more and more companies eventually get access to the self-driving platform and it’s going to start to transform businesses. But what we’re talking about is a decade type of timeframe, not a year.

Sam Abuelsamid

There’s not going to be this, you know, suddenly, over one weekend all the human-driven vehicles disappear and are replaced by autonomous vehicles. It’s something that, as you said Bryan, is going to take time. We’re going to see gradual rollouts where the technology can work. It’s kind of a strategy that I’ve heard a few people talk about, yourself included, but it’s not going to just magically appear via an OTA update on everybody’s cars.

Alex Roy

“COVID-19 is the reason for all the delays.” And this is in the context of deployment of anything. A lot of companies have said COVID-19 is the reason, is that the reason?

Sam Abuelsamid

That’s an absolutely ridiculous statement. Let’s put it this way. Travis Kalanick, as I mentioned before, said 2018 was going to be the year. There was no COVID-19 in 2018. We knew a year ago, long before there were any known cases of COVID outside of Wuhan, that 2020 was not going to be the year. I think it may have slowed things down a little bit for some programs this year, but this was never going to be the year anyway. And thus, that couldn’t be the reason why it’s not here this year.

Bryan Salesky

Those of us who are actually building real prototypes, right? I mean, certainly the supply base has been set backwards because of factories being shut down, because in some cases, priorities get shifted around, right. But it isn’t the exclusive reason.

Alex Roy

One other thing that’s happened this year is there’s been a huge amount of industry consolidation. Some people predicted it. Some people said there’s going to be a Cambrian explosion of autonomous vehicle developers, but it seems pretty obvious the number of companies in the sector is going to go down. Sam, can you walk us through some of these deals and explain what you think is likely to happen next year?

Sam Abuelsamid

We’ve seen a number of deals and again, this consolidation is something that has been ongoing for a while now. We’ve seen shifting partnerships, there’s been a lot of partnerships over the last five or six years, and those partnerships have been changing over time. For Argo, for example, you got a big investment from Volkswagen that closed earlier this year. They had previously been originally doing work on their own, then working with Aurora now with Argo. Most recently, Aurora just last week, or maybe it was this week, it’s hard to keep track anymore, agreed to buy Uber ATG from Uber. So that’s Uber’s automated driving team. And earlier this year we saw Starsky Robotics go out of business. Zoox got swallowed up by Amazon. The reality is that this is a really hard thing to do. It takes a lot of resources to do it. And ultimately, there’s not going to be a huge amount of differentiation across automated driving systems. Once they’re working, it becomes more of a commodity. And so you’re going to have relatively few suppliers that are producing this and providing it to whoever’s actually going to be deploying it. So you’re going to have companies like Argo and Waymo and a handful of others. And most of the other smaller players that have come into this with big dreams have just not been able to get the resources they need to bring it to fruition. And they’re going to either go out of business like Starsky did, or get swallowed up, as Zoox did. They’re just going to end up combining forces and taking the best of what each has to offer and putting it together to try and make good systems.

Alex Roy

It seems as if there is a ratio of autonomous vehicle full stack developers – the companies that make everything – it’s somewhere between, one-to-one like a car company and an AV stack developer, or maybe even two to one. Does that mean that we’re going to have approximately as many autonomous vehicle companies as car companies in the world?

Sam Abuelsamid

No, I think it’s actually going to be fewer than that. I think a better analogy would be, if you look at the current supply base for the auto industry where we have tier one suppliers – those are the bigger suppliers that pull together pieces from tier two, tier three suppliers – integrate that. The tier ones are the suppliers that actually provide stuff directly to the automakers. So that’s companies like Bosch and Conti and Magna. What they do is they produce a lot of components themselves, but they also bring in components from a lot of other smaller suppliers – tier two, tier three suppliers – up the chain, integrate that and provide full subsystems to the automakers. So they may provide stability control systems or lighting systems or electrical architectures to an automaker. I think it’s going to be a very similar kind of thing with the AV companies who are going to have a handful of AV companies that are integrating, they’re pulling together the complete system, the complete subsystem, they’re may be taking some sensors from various providers, integrating that with a compute stack and with software, and then providing that to an automaker to install on a vehicle.

Alex Roy

So the statement that, “Automakers are going to be the losers in autonomous vehicles,” because there have been people there who have said, “Detroit is done.” What do you think of those that statement? It seems to be accurate?

Sam Abuelsamid

I don’t think so. I mean, ultimately somebody still has to build the vehicle. They’ve got to pull all those pieces together from across that supply chain and put it together into a vehicle that either gets deployed in a mobility service or transportation service or sold to consumers. I think some of the automakers that we see today as independent companies, 10 years from now, maybe even five years from now, probably won’t be around as independent companies. We’re seeing a lot of the same consolidation at that top level of the OEMs. You’ve got Fiat-Chrysler and PSA in the process of merging right now. And I think we’re going to see more of that consolidation there as well. But in general, there’s still going to be a base of automakers that are working with the next level down of suppliers, whether it’s the component suppliers or the AV stack suppliers and various other vendors, and all the way down the value chain.

Bryan Salesky

I think you could take a bit of a financial view through all this as well, and look at where the capital’s going, where the capital’s coming from. At the end of the day, we’ve said, this is a capital intensive thing. The car makers need to put in billions of dollars just to create a tailored vehicle, the folks that want to move people and goods, and that sort of own the customer they’ve put in billions to acquire those customers and to create a demand flywheel, if you want to call it that. And then the SDS providers of course need billions to actually build – hire the teams and build out the technology, and build out the prototypes to test the systems and so on, and get it fielded.

Alex Roy

Bryan, could you explain what the SDS is? Because some people may not know the terminology.

Bryan Salesky

Self-driving system, the corpus of sensors and hardware that make a car drive itself. So if you look at all of those, you add all that up. That’s a lot of money. That’s a lot of investment, right? I think, increasingly, no one company can really, in the early stages you can, but as you start to explore and understand what it takes to really grow and scale something like this, this is huge. We’re talking about transforming significant industries that are very well-established and sort of switching over from one goods or people movement model to another. That takes a lot of capital, right? And so a lot of the consolidation is really getting driven by those capital flows, right? Some companies will have an easier time of fundraising than not for whatever reason, based on how well they’ve partnered. Some companies are going to have their backs against the wall and get acquired. Others are going to need to make acquisitions in order to fundraise on the back of those acquisitions, right? Those acquisitions put things together so that you can fundraise. Um, but at the end of the day, if you follow, if you follow the capital flows and you follow where the money’s going, you start to understand better how this whole thing starts to get stitched together. And at the end of the day, the people who are going to be successful, or the companies that will be successful, will be the ones that somehow put all the pieces together and moved the capital in such a way that the three dimensions are put together – the vehicle manufacturing, the self-driving system and the partnerships to actually put all that technology to work in a real network.

Alex Roy

And that’s why my friends at Zoox, who were trying to do literally everything themselves, were acquired by Amazon. It’s very tough to design the car, launch the service, build the car, have a factory. You just can’t do it all.

Bryan Salesky 

Even companies that can practically print money, right? When you look at that, it just suggests that they’re starting to come to the realization, all of the players, that you have to have a really good, really solid partnering strategy. You’re not going to go this alone.

Alex Roy

One of the things that’s happened during COVID has been statements that, “Ride-hail is dead. Delivery is the future… No, actually wait, hold on, trucking is the future.” So many companies have pivoted. Sam, can you deconstruct for us the statement that ride-hail is dead and X is the future, and how much of that is true?

Sam Abuelsamid

Ride-hail has absolutely taken a huge hit this year for a number of reasons. First and foremost, I think most people don’t want to be in a small confined space, like a car with people they don’t know, that they don’t live with. People are also moving around a lot less this year, with such a large proportion of the population suddenly working remotely, certainly not traveling very much. There has been a lot less demand for ride-hail. In the depths of this thing in the spring, the demand for ride-hail was down something upwards of 90%, at least here in the U.S. and comparable in other parts. And even now, it’s recovered a little bit, but it’s still down more than 75% from where it was in 2019. I wouldn’t say that ride-hailing is dead. It’s basically more on pause this year, just because of the nature of what’s happened. It’s going to take probably quite a long time before it recovers to where it was in 2019.

Sam Abuelsamid

Even once we start to get things back on track, hopefully with a vaccine next year, even once we get to that point, we’re seeing some structural changes in the economy. A lot of people who previously commuted to offices are going to be working from home, or at least working in a hybrid model where they’re maybe not going to work five days a week, maybe going into an office one or two days a week. So there’s going to be a lot less demand there. It’s going to take a long time for travel to recover. So it’s going to take a long time for ride-hailing to grow back to where it was and then start to pick up again. Is delivery or trucking the future? Well it’s part of the future, there’s no silver bullet for any of this stuff. All of these things are pieces of a larger ecosystem, which is why, what I do now is actually what we call the lead research on the e-mobility ecosystem, not just any one piece of it, but how all these pieces come together. As we’ve been staying at home, we’ve been relying more on deliveries from Amazon and from local grocery stores and restaurants. So we’ve got first mile, last mile deliveries, you’ve got long haul trucking. All of that stuff has grown this year actually, while other parts of the economy have shrunk. And so we need to address all of these things, all transportation is ultimately going to be able to take advantage of the kind of work that Bryan and his team are doing. They’re all going to benefit from it in the long run.

Bryan Salesky

As travel starts to recover, Sam, we know we’ll eventually get back to where we we have the same problems that we did before around congestion. And when we talked to mobility folks in the policy space, of course they encourage higher density vehicles where you can put as many people into a rolling platform as possible to make better use of the roadway. Now of course, that goes completely against the age that we’re in now in a COVID era where you want to distance, and you don’t necessarily want to be in a confined space with a bunch of other people. I think air travel is sort of an exception where those who need to travel by air, they’re making the risk/benefit. They’re maybe going to get tested, there’s maybe policy around that, but at least from a road travel perspective, I actually think that the form factor of these vehicles may start to change. You want to brainstorm on that a little bit with me? How do you think it might change in order to marry what we think consumers are going to want in a COVID era, plus the fact that congestion is still going to be an issue. How do we resolve that?

Sam Abuelsamid

I think the form factors are going to evolve. We’ve certainly seen some shift this year back towards personally owned vehicles, personal use vehicles. And a bunch of surveys have been done, people are more interested now, in this pandemic and post pandemic era,` of wanting to rely more on personal vehicles than on shared vehicles. I think that that’s not necessarily going to be a permanent shift there. I think we will start to see a shift back, but to do that is going to take some, some changes.

Bryan Salesky

So the question is, do you have a bunch of like really tiny pods that make better use of the space, or maybe you just do nothing and you have inefficient use of the space and you deal with the congestion? I don’t know, but I sort of wonder if it moves toward, okay fine, we’ll put out a van type of form factor, but you know what, we’re going to start to have compartments and all sorts of other things to isolate you.

Sam Abuelsamid

Pods are definitely part of that puzzle. One of the things that will help is that structural change away from as much commuting as we’ve done over the last several decades. Having more remote work, that’s going to take some of the pressure off the road infrastructure. If you’ve got fewer people commuting into jobs in a city or in a business park then that leaves a little more room for some of the other vehicles. So you can have a combination of these things, and one of the things I’ve been writing about this year is this idea of there’s an opportunity right now, we’ve got this period of time where we can start to make some policy changes, while we’ve had this reduction in traffic and vehicle miles traveled, to rethink how we’re allocating road space to make it available for other use cases, to get more shared or not necessarily larger pooled vehicles, but have vehicles that are shared mobility vehicles, smaller vehicles, but that are utilizing lanes that are dedicated for them. So they can have better traffic flow there. Setting aside more space for micromobility applications, for bikes and scooters and things like that. A lot of cities are looking at this and trying to come up with solutions, take advantage of this moment in time to make some of those changes.

Bryan Salesky

We’re talking about the form factors though, I got to ask, why is it that all AVs of the future, Sam, need to look like toasters?

Sam Abuelsamid

There’s no reason why they have to look that way. If you’re thinking of it from purely from a form following function perspective, then that’s kind of the shape that’s going to evolve.

Bryan Salesky

So you’re building it from the inside out to encapsulate that?

Sam Abuelsamid

Your focus now is on the space for the riders in that vehicle. So you’re building around that space and then wrapping the smallest possible package, shrink wrapping a package around that volume that you’re dedicating to your passengers. But I think it will evolve over time, and we’ll start to see more flare coming back into those designs.

Alex Roy

So if the upside is form follows function, what are the downsides of a toaster? I can think of a few.

Sam Abuelsamid

Other than the pure boredom of looking at these things, I don’t, from a functional standpoint, I don’t know that there’s any particular downside to it, but walking down the street and seeing streams of these things, driving themselves down the road, it would certainly get rather monotonous.

Alex Roy

These things are going to have to take us to the airport eventually. And over time it becomes really inefficient to have a brick moving against the air at 45 or 55 mph or whatever speed it is. And that consumes energy.

Bryan Salesky

Aerodynamics.

Sam Abuelsamid

As we start to expand the operational envelope for these things to include highway driving higher speeds, I think we will start to see vehicles, obviously, are more aerodynamically refined. Certainly some of the concepts we’ve seen over the years of automated vehicles like the Mercedes F 015, a few years ago, I mean that certainly wasn’t a toaster on wheels.

Alex Roy

Do self-driving vehicles have to be electric?

Sam Abuelsamid

I would say they have to be electrified. They don’t necessarily have to be purely battery electric vehicles. These systems consume a fair amount of electrical power and so you need something beyond the traditional 12-volt electrical system just to supply them. So you’re going to see electrification and some companies are like Ford for example, are going down the hybrid path with their vehicle. Others are doing battery electric, some are doing plug-in hybrid. So I think we’re going to see a mix, but the common factor is going to be electrification in order to provide both the efficiency and the emissions reductions that you want, but also to provide the electrical power that’s needed.

Alex Roy

Am I correct in my assumption that there’s tension between the people who want to believe that everything should be not only autonomous, but electric. That in order to do that, you would have to build a lot of charging infrastructure in and around the area where you want to deploy to feasibly have an electric autonomous fleet. And that’s the problem. That infrastructure just doesn’t exist.

Sam Abuelsamid

Certainly, the desire is to shift towards all-electric, to go to a zero emissions environment. That’s the end goal for all this. And yes, it will take a fair amount of electric charging infrastructure to support that. But even if we don’t do automated driving, even if we don’t do self-driving, you know, the goal is to get to electrification anyway. So you’re still going to have to have a lot more charging infrastructure than what we have today if we’re going to ultimately convert our entire vehicle fleet to electric. I think that’s something that’s and it will continue to happen and expand just to support the overall goal of reducing emissions.

Alex Roy

A lot of people making this statement, “Self-driving trucks are easier than self-driving cars.” Can you deconstruct that statement?

Sam Abuelsamid

It’s not necessarily an easier problem. It’s a different problem. There are some aspects of it that are easier. There’s some aspects of it that are harder. If you’re thinking about a tractor trailer, these things weigh 40 tons at 65 mph, if they have to come to a complete stop, they take a lot longer to stop than a 3000 pound car does. And so that means that you’ve gotta be able to see further down the road in order to respond sooner to what’s happening. So you need sensors that can detect longer range. But on the flip side, if what you’re focusing on is highway driving, you’re also eliminating some problems that you have with urban driving. You don’t have to deal with intersections and traffic signals and pedestrians, at least not to the same degree, you don’t very often find pedestrians crossing the interstate. So it’s a different problem, but not necessarily an easier problem. From a business perspective, it does have some potential advantages in terms of taking drivers, if you can have automated trucks where you can have higher utilization of those trucks because there are limits on how many hours a day a truck driver can be behind the wheel, it does have some business advantages. If comparing delivering goods, whether it’s a local delivery or a long haul truck, versus carrying passengers, the demands from whatever’s being carried in that vehicle are going to be a little different. Packages don’t really care quite as much about the ride comfort or ride quality as passengers do. So you may be able to safely deploy something to deliver packages a little bit sooner than you could reasonably do for passengers. The unit economics are going to be different because you don’t have to deal with having sensors in the vehicle to detect who’s in there, detect if they’ve gotten sick, if they need medical attention or they’re smoking or whatever in the vehicle. So it simplifies some things makes other things more complicated.

Bryan Salesky

Yeah, I agree with what you said, Sam. I think, given the stopping distances are so much longer, it really does change a bit how you model and reason about the world. Because if you, I mean, let’s face it truck drivers today, from an efficiency standpoint, they will sort of overdrive their human perception. In other words, they will get closer to things then they could stop if everything in the world came to an immediate stop. But, in order to do that, when they bring in those following distances, they’re making some really, really tough judgment calls that for a computer to make those judgment calls is very difficult. And I think that’s where a lot of the problem space exists. It’s dangerous to say that pedestrians don’t exist on the roads where big trucks operate on. I think you’re right, on divided highways that are in very rural parts of the country, that’s largely true. You could drive for a thousand miles and not encounter a pedestrian. But when we’re talking about trucks operating on boulevards or highways where there’s a large amount of on-off traffic, absolutely bikes will use those thoroughfares, you’ll see pedestrians on those thoroughfares. And so I just think this is where you got to now get in on the business side and really look at the use case. Are you doing hub-to-hub, exit-to-exit autonomy, and there’s people on those end points to then take the truck the rest of the way and take it to its final last miles to the destination? Otherwise, it’s in a very rural part of the country where you’re on a divided highway and it’s basically sort of an established trucking route. To me that’s very different than if you’re driving a big rig from one end of Miami to the other. And this is something where the devil’s kind of in the details, I think to really understand what the performance and the ultimate specs need to be for the self-driving system.

Alex Roy

There has been a trend recently of declaring driverless deployment. We have done driverless in this place, as if that is a sign of victory, like a major milestone. Could you unpack some of these driverless announcements and what real value there is behind what they claim to do?

Sam Abuelsamid

I mean, anytime you make a change like that, taking somebody out from behind the steering wheel, that is certainly a milestone, whether it’s a significant milestone, is something entirely different. The folks at Cruise announced that they had begun testing some of their vehicles in San Francisco without a safety driver. Um, of course there are lots of caveats, you know, there, there’s still a safety operator in the passenger seat with their hand by the kill switch to keep an eye on things and be ready to bring everything to a stop if anything appears to be going wrong. They’re also operating only in one particular neighborhood of San Francisco rather than the entire city. They’re also, at least the videos that they posted, show the car running around at night, presumably late at night or very early morning when there’s no moving traffic around. So in the video we’ve seen so far, the only thing in motion is that test vehicle, everything else is parked. There’s no pedestrians, and they do have a teleoperation system as a backup. So they are testing vehicles without safety drivers in there.

You know, Cruise had originally announced a couple of years ago their intention to have commercial operations by the end of 2019. Obviously they didn’t hit that milestone. What I’ve been told is that part of the funding agreements from their investors like SoftBank and others, included hitting various milestones to get their tranches of funding. And since they missed that milestone of commercial operation in 2019, they had to set new milestones and apparently doing testing without a safety driver was one of their new milestones. And so they had to do this and announce this in order to get a next batch of funding from their investors.

Bryan Salesky 

I mean, there’s been plenty of teams that have had to do this and it’s just to me, so long as you’re on public roads, it doesn’t make sense, right. If you’re on your own test track, do what you want.

Sam Abuelsamid

Was it back in 2009 or 2010? They had those 10 milestones that the team was given in order to get a certain, certain level of bonuses from Google. And they hit those milestones. But just because they hit those 10 milestones doesn’t mean that they were anywhere near ready for deployment.

Alex Roy

What about this statement? This is a trend out of the same audience, that level four self-driving is not actually self-driving, it’s partial automation because it’s fenced that only operates in one area.

Sam Abuelsamid

I disagree with that. I think if it doesn’t require a human to supervise it or take it over at any point in time, if the vehicle can operate even without a person in the vehicle and doesn’t require one hundred percent supervision, you know, even if it’s limited in its operating domain, it’s still self-driving within that domain.

Alex Roy

I like to say that if I could sleep in the back or if I could sleep at home while it goes somewhere, it’s self-driving. That’s my test. Very simple. There are a lot of sensor companies out there, and it seems like in 2020, there are even more sensor companies announcing themselves, as if their sensor was the breakthrough. Does any of that make sense? Do we need more sensors or do sensors have to change or evolve before we deploy autonomous vehicles?

Bryan Salesky

I think that sensor technology, thankfully, is evolving and that’s a good thing for the industry. I mean, we need longer and longer range, higher and higher resolution, and for these sensors to be robust and all sorts of different climates and environments,. That’s a good thing, right? I think the myriad of startups that are out there in the space, haven’t really gotten to the more difficult parts. They’re still working on very much achieving the basic specs required to enable full self-driving. And they operate in different wavelengths, some of them. Some of them operate on different principles, whether they’re photon counting or other types of methods of operation. And they all have different strengths and weaknesses. And I think we probably have a few more years before a clear winner starts to emerge, and in fact, a winner may not emerge given that some of the most advanced sensors are actually being developed now by the AV companies, and there’ll be a competitive advantage. They won’t sell them to other customers. So I think the sensor market’s one to watch. I think it’s really fascinating. Those who do launch a fleet of self-driving vehicles with their own sensing technology, it will be interesting to see, like what has to happen in the market for them to then decide when to sell their capability versus to sit on it and use it as a competitive advantage. I think it’s one of the fascinating stories to watch out for in the coming years.

Sam Abuelsamid

There’s absolutely no silver bullet sensor, or silver bullet technology out there that’s gonna make this suddenly have the whole world be self-driving.

Bryan Salesky

They all have strengths and weaknesses that’s for sure. There’s no one thing that will solve the problem.

Sam Abuelsamid

You know, I think there’s a lot of interesting stuff that’s coming out there, whether it’s thermal imaging, imaging, radar, some of the, some of the newer types of LiDARs . There’s a number of things going on here, not only do the sensors have to get better in their performance and their capability, but they also have to get less expensive. A lot of these sensors are still very costly. When we hear about a company coming up with some new new sensor type or new approach to the problem, I think that’s a good thing. And, we need to keep encouraging them to make those advances because it benefits everybody. And you know, which one or which combination of sensors will ultimately be the right one, it’s too early to tell. The only thing we can say for sure is that it almost certainly will be a combination of sensors. It won’t be any one single sensor. You’re going to have multiple different modalities that are part of the solution.

Bryan Salesky

And we see now, Sam, that companies that even had the most amount of religion around any one type of sensor are starting to embrace the fact that they have to go to multiple modalities.

Alex Roy

Will self-driving be winner-take-all?

Bryan Salesky

Three trillion miles in the U.S. alone, different types of miles, different conditions, different platforms, different use cases, different customers. There’s no way it’s winner-take-all; it’s too much capital to go after.

Sam Abuelsamid

It’s not realistic to expect it to be winner take all, nor should we want it to be winner take all. This is again, going back to that ecosystem idea, this is part of an ecosystem of mobility. We need diversity in that ecosystem, just like any natural ecosystem where you want diversity in there to make it robust and make it resilient. There are going to be problems with everybody’s system at point in time, and you don’t want too much reliance on any one solution that if something goes wrong or if it goes down, uh, you know, that it brings down the entire system and brings everything to a stop or worse leads to some sort of catastrophe. So I think you do want diversity in the system and we should be discouraging any attempts at a monopoly or winner take all solutions.

Alex Roy

Well, that was a great episode. Thanks so much guys. And for the record, I believe that when the vaccine arrives, people are going to get jam packed into every moving vehicle they can, head into cities and it’s going to be the roaring twenties all over again. Sam Abuelsamid is principal research analyst for e-mobility, at Guidehouse Insights. You can follow him on Twitter @SamAbuelsamid. And Bryan Salesky, of course is the voice of reason and the CEO and founder of Argo AI. Thanks so much to both of you.

Bryan Salesky

Thanks Sam, that was great.

Sam Abuelsamid

Thanks for having me on again. Always great to talk to you.

Alex Roy

If you love what you heard, please hit us up on Twitter @NoParkingPod. I’m everywhere on social media @AlexRoy144. Please share No Parking with a friend, like us, subscribe, give us a great review on whatever platform you listen to podcasts on. This show is managed by the Civic Entertainment Group and Megan Harris is our producer. Until next time, I’m Alex Roy. This is the No Parking Podcast.