ROADSIGNS EPISODE 56

WHAT WILL IT TAKE TO MAKE AUTONOMOUS TRUCKING A REALITY?

Have you been imagining a driverless future where the machines do all the work and humans watch? You might be in for a surprise. Even at Level 4 autonomy, self-driving trucks will have technical limitations. They will be designed for specific applications that are best suited to automation. They’ll also require industry buy-in and public acceptance. In this episode, we ask how technology developers are clearing those hurdles to make autonomous trucking a reality. And how might the trucking industry improve safety and efficiency by combining automation and human labor in the future? To find out, we bring in Boris Sofman, head of engineering for the autonomous trucking program at Waymo. Tune in and learn more about the road ahead for self-driving trucks.

Access the full transcript here.

FEATURED GUEST

Boris Sofman is the Head of Engineering for Trucking at Waymo, an autonomous driving technology company with a mission to make it safe and easy for people and things to get where they're going. Boris leads a team that builds on Waymo’s technical foundations and experience to pursue the tremendous opportunities in trucking.

Demarco Thomas

Boris Sofman

EP. 2

Brought to you by:

Demarco is a serial entrepreneur whom is a managing partner of Metromax Group. Metromax owns multiple transportation related businesses ranging from over the road transportation, business process outsourcing and last mile delivery. Collectively, the Metromax group employs over 180 team members globally.

HOME

Episode Transcript

Dan Ronan From Transport Topics in Washington, D.C., this is RoadSigns. Here is your host, Seth Clevenger.

Seth Clevenger Thank you for listening to RoadSigns, the podcast series from Transport Topics that explores the trends and technologies that are shaping the future of trucking. In this episode, we're going to take a fresh look at the current state of autonomous truck development in 2021. Increasingly advanced driver assistance technologies are already available on the commercial truck market, but at the same time, some technology developers are working hard to make fully autonomous trucking a reality. Today, several companies are testing self-driving trucks on public highways in the U.S., especially in the southwest. At least for now, though, these test trucks still have safety drivers behind the wheel as they pave the way for commercialization in the years ahead. But what obstacles remain on the road to autonomous trucking? And what will this technology mean for the future of the freight transportation industry and its workforce? To help us answer those questions, we're going to bring in a guest who is right in the middle of this emerging field of technology. With that, we're excited to welcome Boris Sofman, head of engineering for the autonomous trucking program at WAYMO. Thanks for joining us, Boris.

Boris Sofman Thanks, Seth. It’s a real pleasure to be on.

Seth Clevenger So WAYMO is best known for pioneering autonomous driving technology for passenger cars, going back to the Google self-driving car project. But here on RoadSigns, we're most interested in WAYMO's trucking division. And commercial trucks are a very different animal in several important ways. Of course, a Class 8 tractor trailer is a much larger vehicle that requires more time and distance to come to a stop, and it typically spends most of its time traveling on interstate highways rather than through urban and suburban roads. So as you ramp up your testing and development of autonomous Class 8 trucks, to what extent were you able to directly transfer the same sensors and software that you use in your self-driving cars? And what have you been doing differently on the trucking side?

Boris Sofman Yeah, it's a very good question. Probably one of the most exciting elements of this role and the whole opportunity is the degree of leverage we're able to get from WAYMO's 10 plus years of investment in the space. And so what was really visible from the very beginning, and part of the core thesis of tackling trucking in parallel to bringing up the consumer transportation side of the business, is just how much reusable underlying technology is possible and how a majority of what WAYMO has been pushing on from the beginning is not about solving for an autonomous car, but fundamental technologies for autonomous driving. And that actually covers everything from the hardware side, where you think about the sensors, the lidar, which is absolutely best in class, and cameras, radar, to the compute, the know-how of how to integrate that hardware into a platform. On the software side, you have A.I. and M.L. infrastructures, you have core perception systems, behavior prediction systems, planning systems, simulation, infrastructure, logging. And so in a lot of ways, what we're finding is that these building blocks become very versatile because the idea of understanding the world around you and the idea of predicting what other agents might do or what you need to do, it actually carries over quite well, even if there's unique differences around the domain, like freeways being very dominant for trucking or specifics around dealing with variable mass or articulated trailers or the customers themselves. So those need adoption, but the underlying foundations of how you think about tackling the challenge of automating a vehicle actually carries over quite well. And that was actually a thesis in the very beginning of the company. And so when trucking started to become an early R&D project back in 2017, like a small number of people, like five people, were able to bring it up and actually get it to start the test drive on a freeway. And it kind of validated the fact that a lot of this tech carries over. And I think it's one of the biggest advantages for WAYMO, because what we're seeing now is actually technology going both ways, where a lot of the building blocks that we push on from the trucking side actually progresses underlying core technology base and actually help the car side as well. And that's what we want to amplify going forward.

Seth Clevenger Well, that's fascinating how much you can carry over from all that work you've done over the years, working primarily with passenger cars. But let's talk a little bit more about the on road testing that you're doing today in your self-driving truck operations. And right now, your trucks are running on interstate highways in the southwest, particularly in Texas. So why did you see this as an ideal region for testing? And can you tell us more about your day-to-day operations as you really paved the way for a commercial launch of this technology?

Boris Sofman So eventually we obviously see trucking being something that generalizes to the entire United States and obviously globally as well. But we had to really focus and get a beachhead in a place that can help us prove out some of the most novel parts of technologies that are necessary to solve this problem and create the least additional headwinds along the way. And so the Southwest US is in general very favorable because of weather conditions, the geography itself, in the case of Texas, New Mexico and Arizona, there's very good kind of political factors that make this conducive to having a clear path to L4 trucks. And then on top of that, you look at some of these corridors and because of the the Port of L.A. becoming a big conduit for a lot of goods traveling across the United States and in Dallas and Houston being one of the biggest freight hubs in the U.S. as well, we can not only develop our Driver over very representative of initial routes, but also ones that become very important pieces of a broader network that starts to build value from the very beginning, given the volume that they serve. And so most of our testing is that I-10, I-20, I-45. We also test run our around our Mountain View headquarters in the Bay Area around 280 and 101. And so a lot of what we're doing day to day is collecting lots of miles. We're doing lots of tests. We're bringing up new pieces of technology. We're doing data collection and trying to discover what that long tail of challenges might look like, the sort of things that only appear when you drive hundreds of thousands or millions of miles in between events. And so we're testing those core capabilities and starting to think about what it's going to take to really validate this and get it to a level of robustness where we feel good about going driverless. And at the same time, we're building up our muscle on the commercialization side by engaging in fleet pilots and discussions with potential customers that will hopefully transition once we're able to go driverless. And then obviously working on our partnership in parallel with Daimler to actually build the platform itself.

Seth Clevenger So many moving parts and so much work that goes into it, you know, measuring, you know, miles in the millions as you prepare for this. And, you know, just turning my attention back to the ride hailing side of the business momentarily. Last fall WAYMO began offering fully autonomous ride hailing service to the general public in part of suburban Phoenix. So within that specific geographic area, customers today can call up a self-driving Chrysler Pacifica and take a ride with no driver on board. Now, when you look at the trucking side, your class 8 trucks still have a safety driver behind the wheel for backup performance as you continue testing and validation. But when do you anticipate you'll be ready to take that step and pull the safety driver out and begin operating unmanned class 8 trucks on public roads? I'm not necessarily asking for a specific date, but how will you know when you're ready to take that step?

Boris Sofman Yeah, very top of mind. I think one of the biggest advantages that we have at WAYMO is that we're the only company that's actually gone that distance and as you mentioned, has a service that you could land in the Phoenix area, download the WAYMO app and actually have a car that has no safety driver. That's something that nobody else has done. And the advantage of that is, is actually that we've learned the hard way how big that chasm is from a really, really cool looking demo to something that actually can go driverless. And that we have the confidence in to release with real customers on the road, knowing exactly in a quantifiable manner what how it performs from a safety bar. And so for us, we're able to take advantage of that. And so that know-how becomes obviously a huge advantage for trucking at WAYMO where we are able to work backwards. And for now, as you mentioned, we do have safety drivers and we have a clear timeline and path to get the commercialization that aligns the software progress with the hardware and when they need to align. But for the time being, we're leveraging drivers in order to be able to aggressively push on the capabilities and learn quickly without having to worry about safety concerns. And so this isn’t going to be something that happens overnight. We obviously care a lot about the safety and tech readiness over any specific point in time. And probably one of the biggest things that we would care about is that when we do go driverless, it's not a one-off demo that we then retreat from and use as a checkmark. But it's something where that marks the beginning of twenty four hours a day, seven days a week operations that we can then start to scale from one truck up to a large number of trucks. There are a few just fundamental physical barriers to going driverless today where the trucking industry is not quite as mature as the car industry in terms of having redundant platforms that are already built into the trucks themselves, and so being able to go fully autonomous with no human driver is actually something that we won't do and nobody else should do without a system that actually has all of this redundancy and safety features built in so that you have failure modes if your brakes fail, your steering fails. On the passenger side, we've had multiple generations of vehicles have done this. I think that experience is going to be very helpful for trucking. But the trucking side of the industry isn't as far along, and that's why the partnership with Daimler is actually so valuable. This will be a big step up from just the platform and mechanical standpoint where we'll be working together to ensure that the platform itself not only is capable of driverless, but also has a redundancy that will enable it. And so in a lot of ways, we're working simultaneously to mature the software capabilities and the autonomy stack while also ensuring that the hardware specifications in this project with Daimler are going to meet all of those requirements and enable us to start to ramp up production when the two collide.

Seth Clevenger Whenever we bring up the topic of autonomous trucks, I think naturally one of the first questions that comes up is what it all means for the future of truck driving jobs. I've been making the argument for years that professional truck drivers are going to be absolutely essential for the foreseeable future. Even when fully autonomous trucks hit the road, they're not going to be able to go everywhere and handle every load and every route. They'll be designed for specific applications. And I think they'll be deployed in ways that ultimately compliment drivers and may even make their jobs better by reducing length of haul and enabling more home time. And those are all my thoughts as a trade press journalist who has been watching this closely. But you, Boris, of course, are right at the center of it on the development side, so I'd like to hear your take on this question of what automation means for the future of the professional truck driver.

Boris Sofman Yeah, Seth, your intuition is actually spot on. This isn't something that's going to happen overnight. It's going to be a gradual progression. There's going to be areas of focus and an evolution. Every time you grow an order of magnitude and volume of trucks like, on the road, that's a giant amount of miles that you're going to absorb and learn from. And there's going to be things that we need to adapt to react to. Different regions in the US have different challenges, like weather, different freeway properties. We want to expand functionality not just on freeway, but surface streets. And so all these things take time. And I think most people underestimate the complexity of not just going to hit your first route, but how much effort it then takes to scale into a more generalizable and broader functionality. And we've definitely learned this on the consumer side with cars where we're we're working to build that generalizability in the technology, but that takes a lot of effort and a lot of time. And so as a company, obviously, the long-term vision is to really create this flexibility in the logistics network in the United States that is actually additive and not cannibalistic in terms of its economic impact. And so, when you think of trucking on its own, it's the centerpiece of our economy, but it actually has a pretty huge shortages of drivers where, you know, today I believe we have like a 60,000 driver shortage that's projected to widen to 160,000 throughout this decade. And the average age is quite high. And the replenishment of new people going into the space despite the growing demand is not great. So the great thing about autonomous driving is that the types of routes that it's most conducive for, are actually oftentimes the least pleasant ones to be a professional driver on. And we will always have compatible needs for either certain routes and also all of the kind of tangential industries around it that are going to benefit from the efficiency gains and flexibility gains that come up. And in fact, even our own drivers are oftentimes veterans with decades of experience. And so we're bringing on more and more people from the industry, whether it's technicians or drivers or trainers, testers, and trying to absorb their skill sets and benefit this progress. And so I think when you think about all of the.. it's almost similar to the Internet, where there's obviously a lot of impacts in terms of every business that translated to. But at the same time, the sheer versatility and efficiency that that enabled created vastly more overall economic positive impact, including jobs in exactly the industries that might have initially thought that there would be a disruption. And I think we're pretty confident that there's a clear complementary aspect here that not only helps directly address some of the job shortages, but actually creates a lot of positive secondary benefits that will start to more and more benefit various industries around trucking.

Seth Clevenger Well, thank you for that insight. And yeah, you're certainly right on right on the spot when you say that, you know, driver recruiting, driver retention is really an existential problem for the trucking industry. And, you know, we see this this on. Opportunity for automation to help bridge that gap and supplement the workforce and quite possibly make a lot of those jobs better in the future. And, you know, you look at some other industries that have been finding ways to combine automation with human labor as technology advances. And, you can look at manufacturing as an example. You go to a car plant or a truck plant. And of course, it's amazing the robotics that are in place that handle a lot of the heavy lifting. But, of course, you still see people there who were maybe handling some of the fine tuning and the more detail oriented work and of course, overseeing the process...

Boris Sofman Even like Amazon facilities. If you look at the automation that’s happened, a lot of times what happens is the density increases, but the net jobs does not go down. Oftentimes it actually goes up because it just shifts them into other areas that end up benefiting from that efficiency. And I think that sort of property I think is going to hold here as well.

Seth Clevenger Yeah, I mean, exactly. The warehousing example is a good one. And that sort of interaction of robotics and automation with human labor. Any more insights on how you think the trucking industry specifically can most effectively manage that evolution with combining automation and human labor in the years and decades ahead?

Boris Sofman Yeah, I think this kind of goes to our previous conversation where there's a lot of already a good trajectory here, where driving specialists are integrated into every aspect of operations. We're building out supplemental roles like dispatchers, technicians, customer support. And, of course, that even if the route itself becomes automated, there's a huge amount of complexity and work required, and also as part of the core job from truck drivers, beyond just the driving part. And so you think about loading, unloading, inspecting repairs, the records keeping of everything that's happening, the mechanical work that comes with a more complicated system, the customer support aspects. And so we expect that this will actually end up just evolving how people integrate into this chain. But we're we've actually been very pleasantly surprised that by the more we learn about this, the more the actual capability of driving the truck with different constraints and different degrees of efficiency, how much that actually opens up the ability to then wrap around that and reallocate how people end up complementing these systems. And so we've been doing a lot of like really close case studies with partners and potential customers to really understand this, and the vast majority of cases, it's actually very additive the moment you're able to break through some of the barriers that existed in the past.

Seth Clevenger And let's talk a little bit more about potential deployment models for autonomous trucks. Some of the strategies that have been discussed involve transfer hubs where you'll have autonomous trucks that would pick up and drop off trailers and then leave the first and final mile to manually driven trucks. Or if you can automate the first and final mile, you could potentially deploy a self-driving truck in a more traditional dock to dock operation that's the standard today. But how exactly do you envision deploying autonomous trucks at WAYMO?

Boris Sofman So one of our biggest advantages at WAYMO is we obviously have a huge amount of expertize built up on surface street navigation, which, again, trucking is going to add unique challenges to it, especially from kind of the motion planning occlusions like all these really interesting technical challenges. But there's a giant head start on how do you navigate complex interactions with pedestrians, vehicles, intersections and things like that. And so for us, that's going to be a huge advantage. That allows us, far more efficiently than we think anybody else, to be able to stitch together not just the freeway capabilities, but also end-to-end surface street capabilities. Now, we're trying to be pragmatic and make sure we don't boil the ocean and try to tackle everything at once. And so initially, we're going to explore a mix of these types of models where we are looking at transfer hubs, we are looking at surface street functionality. We're going to have a parallel road map that where different degrees of capability are phased in over time. And again, this is just software. So, you know, somebody that ends up partnering with us, you would see the capabilities expand over time and fundamentally change the way these trucks can get used. But eventually, I think we're in a really great position to try to complete these networks in a way that probably would be very hard for other companies to do, but initially we want to start simple because getting to market as fast as possible and learning probably has the most value. And then we obviously would very, very quickly expand both routes and the coverage of surface streets along the way.

Seth Clevenger And then when you look at the trucking industry itself, of course, it’s very fragmented. You have small and medium sized companies that handle much, really majority, of the work and the operations within the industry are also quite diverse. You know, you have companies specializing in longhaul, some of them are more focused on regional and local routes, and companies are hauling different categories of freight from dry van truckload to refrigerated, flatbed, less than truckload. You have liquid bulk, dry bulk, hazardous materials. When you look at the industry, what types of fleet operations do you believe are best suited to autonomous trucks in the near future?

Boris Sofman It's a very good question. So obviously a few of these have unique challenges tied to them, if you're transporting hazardous materials or liquids or things like that. And so, you know, we might intentionally kind of offset when we kind of green light those types of cargos. But in general, what we're looking for is to create a pretty huge amount of flexibility in how people can use these trucks within a fairly wide range of conditions where different customers have different types of loads. And the main focus for us is to make sure that we're focusing on kind of high value types of routes as an initial capability. So particularly line- and longhaul routes and in terms of fleets and how we use them, we want to open up a pretty wide range of a variable load permutations that we can fully support and make it as versatile of a product as possible for our customers. And so from our standpoint, we're creating a product. That product, just like any other product, digital or a physical product, is going to expand in capability over time. But we want this to be something that's easy to use, whether you're a tiny family owned business that just needs a kind of a handful of trucks or you're a giant Fortune 500 company starting to really automate your logistics in a broader part of the US. If we do a really good job of this, it should be a product that's equally usable and acceptable by both and a lot of our kind of studies and evolution on the kind of product and commercialization side are actually really trying to keep this in mind so we're not overfitting to any single type of user.

Seth Clevenger And I also want to pick up on some of your comments about the integration of this technology into modern trucks, and you referenced WAYMO's partnership with Daimler Trucks that was announced in October, with the plan being to produce a self-driving Freightliner Cascadia model piloted by the WAYMO Driver. But as you mentioned, to make that happen, you're going to need to add new components to the vehicle, redundant steering and braking systems, for example. So take us through the process, because it's not going to be as simple as just adding your sensors and software to an existing truck.

Boris Sofman Yeah, absolutely. This is one of the unique extra challenges for trucking that right now makes it makes a larger body of work necessary on a platform side than than, for example, if you want to be another entrant into the autonomous car space. And so this is a really deep integration. Like you said, it's not just adding software, is it adding sensors to a truck. It's a really, really deep integration of our self-driving system, which includes sensors, compute, all the electronics, very, very precise positioning of those sensors that are meticulously designed to optimize for fields of view, occlusions, coverage, redundancy, things like that. And so all of that has to be very meticulously co-managed. And luckily, we've had the experience of doing this with a number of car partnerships, which is carried over really well into how we manage the Daimler relationship. And right now, there is no truck in the market where we can order something that is complete, even from a redundancy of safety feature standpoint. And so that's something that we're partnering from the very beginning with them, with Tier Ones that will be integrating, steering, braking and other parts, and then obviously creating a feedback loop so that we can validate how these systems start to work through prototypes on the software side and ensure that we're not missing anything so that as we lock it in for broader production, we're happy with the results. And so we're actually pretty excited because on the other end of this, we should come out with an L4 capable truck with what we feel is the best OEM partner that we could have picked on the market that has been meticulously co-designed with the software strategy and aligns really well with exactly what we need. And just like just like you see in a lot of industries, like, for example, Apple with the iPhone, like a lot of these really, really new industries being vertically integrated and really having a deep and precise control over both the hardware and software part is actually quite critical, especially in an early phase before technology starts to get kind of standardized. We're very, very, very far from that. And so having really deep control and influence over how we design the platform actually makes the software challenge much more tractable.

Seth Clevenger Yeah, well, I'm personally looking forward to a I guess I would call it a test drive, but I guess it would be a test ride, when the day comes and when that all exists and there's a truck that's ready for commercialization. So I'm looking forward to that on a personal note.

Boris Sofman Yeah, us too!

Seth Clevenger But for autonomous trucking to really take off, it's going to require buy-in not just from the industry itself, but also it's going to depend on public acceptance. Regulators and legislators will need to be on board. The motoring public will need to be comfortable with it. So what steps can a company like WAYMO take to build trust and make people more comfortable with the technology that you're currently developing?

Boris Sofman Yes, this is really top of mind. So the very first thing is transparency and accountability. So, you know, in terms of a cultural value, safety is engrained is the top of the list at WAYMO. You've probably seen as a good example from the car side where recently we released the entire safety methodology and process that we use to validate our cars to go driverless, completely driverless, in the Phoenix area. And we're the very, very first company to have released that sort of a framework that describes it. And we're very, very proud of it, because this is something that we hope starts to kind of educate people and build that trust that the rigor we put into validating the system is beyond anything that most people would ever imagine. The sheer number of miles that goes into the statistical assessment of every single type of scenario you could possibly imagine, from pedestrians to motorcyclists to traffic lights, failure modes. It's about as thorough as I could have ever expected, like as I started to learn about what would happen outside. And and so our hope is that this demonstration of that rigor and the transparency in which we approach it actually starts to build trust in the public and press and regulators. And in fact, there's also another kind of really important factor that we're thinking very deeply about is that we want to raise the bar so that everybody is held to this level of rigor, because our biggest fear is that somebody cuts a corner and actually has an accident that actually sets the whole field back, particularly in something like trucking, where the safety case is such a higher bar on a freeway with a fully loaded, 80,000-pound truck, you have no room to try to wing it and try to see what happens. And so from an education standpoint, we're increasing more and more our degree of transparency. I think the success on the car side and the continued scaling of the car fleet will probably help us on a truck side where there's a halo effect from that trust and that success rate that carries over. We've already shown, you know, really, really meaningful numbers of miles and have had a really stellar safety record on that front. And then obviously we want to work closely with the community and then partnerships like the one with you to help educate the public. And then, of course, one of the best things you can do is actually experience the technology firsthand and take a ride. There's nothing quite like sitting in the back of an autonomous truck going down a highway. It's kind of crazy sitting with a car like you see a car, like you're sitting in the backseat. There's nobody in the driver's seat, the steering wheels turning on its own in a, you know, unprotected, like, fast paced road. And even when you know that it's really safe, you still feel like, it's easy to take for granted, just like how meaningful that is. And I hear you going to be joining our virtual demo, and so things like that are hopefully like a good step in the right direction. When you start to see the progress of the industry and you see that you can follow along in our mindset and the journey that we take in order to get the full driverless where it's not just bottoms up working towards where we think we need to go, but it's working back from what we know is the right safety bar. And doing that sort of bidirectional search actually makes for a better development process and a better product.

Seth Clevenger Well, thank you for that. That's very helpful. And yeah, I think it's just in some cases, to your point, a matter of experiencing it firsthand and it builds that confidence. And on the in the passenger car side, I think consumers are getting more and more used to some safety features that start to incorporate lower levels of automation. And perhaps that helps further the conversation as well. And speaking of conversations, you know, there's been this ongoing discussion within the trucking industry about the role of automated driving technology, really ever since we saw some of the early self-driving truck demonstrations several years ago. And over time, that conversation has become more nuanced. I think the industry as a whole has learned a lot along the way. But I also want to give you a chance, Boris, to address any lingering misconceptions about the technology. Is there anything that you would specifically like to address that you'd like to clear up about the future of automated driving?

Boris Sofman Yeah, I think this is something that actually applies to both trucking and the car side. I think the world and I mean consumers, investors, even companies themselves, I think still underestimate just how massive the surface area is of this problem and just how complex it really is. And I think it's deceptive oftentimes to look at a really cool demo or other types of statistics or a video of a flawless 30-minute run and extrapolate from that and say, wow, they're really close to driverless, because if you just look at the statistics, flawless run will happen all the time if you have if you like, a very basic level of robustness. And so those are not like the right metrics. What really matters is how do you quantifiably how we think about how you handle the one-in-a-million-mile rare scenario and these corner cases that end up being the long tail that keeps you from going from a really great demo to fully driverless. And I think WAYMO really learned that lesson on the car side where, you know, there were incredible demos like seven, eight years ago, and it took a lot longer to actually get to a service that we could be running day in and day out and just almost take it for granted where it's just happening behind the scenes. And there's customers right now taking autonomous drives in the Phoenix area. We're trying to really internalize that on trucking, where even more so on trucking, you can show progress reasonably quickly. But the rigors that are necessary to deal with these less visible and sometimes less glamorous situations like long-tail construction zones or dealing with every permutation of failures from your sensing to mechanics. All of those things are the things that really are the gate to go driverless. And I think that it would be great for the world to really appreciate that, that, you know, you can't look at the cosmetics and extrapolate directly from it because this is really, really hard, and I think there's a lot of respect for it. And then even when we do reach driverless or, whenever any company reaches driverless, it's not something that just happens overnight. There's a huge amount of focus on a particular well defined domain and that path to like really expanding into broader functionality. It'll take time. And just like any other massive and transformative technology, it takes decades to evolve and really kind of grow into its full form. And that also is something that's exciting because it means that there's always new challenges on the horizon. But it also means that, like anybody who says that overnight, we're going to solve all of autonomous driving, it's just a fundamental impossibility, and so maybe that's a good lesson for just autonomous driving in general, not even just trucking.

Seth Clevenger Yeah, that's very helpful. Before I let you go, Boris, I'd like to open it up for you to share any more final thoughts you may have about the work that you're doing at WAYMO and your vision for the future of the trucking industry.

Boris Sofman Yeah, I absolutely love the work here. It's like a collection of some of the most talented people I've ever crossed paths with. And it almost feels like it's the most cutting edge R&D across every segment of this problem, which is just so massive that you almost have to rethink everything from validation to sensing, to the autonomy stack to operations to everything else. And so it's been super cool. It's autonomous driving that got me into robotics in the first place, so this is kind of a dream to be able to kind of see the landscape kind of enable us to not have to squint too hard to actually see how this makes it to commercialization and starts to have a big impact. Trucking is just one of the most exciting industries, because if you extrapolate forward from a first phase where you're able to solve existing routes and kind of introduce an alternative within the confines of the current logistics industry that already is meaningful and transformative and addresses a trillion dollar industry, what becomes really, really exciting is if you think about and you fast forward when this becomes far more prevalent and you compare it to what's happened with like really, really important technologies like the computer or the Internet, it doesn't just transplant certain segments of a sector with a little bit higher efficiency or cost dynamics. It actually completely enables you to rethink how you approach this really critical part of the economy and almost kind of spawns entire industries and enables opportunities that didn't exist before. And so that's where it gets really, really exciting, where once you think about what happens at scale and what becomes possible, it's actually just fascinating to think about how the entirety of logistics can potentially start to morph to embrace this and completely unlock elements of the economy that just were never possible before. So I think it's sometimes fun to be able to zoom out and think about it that way. And we're far, far, far away from that. But I think that's what makes it so exciting and continually transformative over both the near-term and long term.

Seth Clevenger Yeah. I mean, I couldn't agree more on all this happening in trucking and freight transportation. There's is often overlooked and perhaps underappreciated industry, but it is so essential and I’m still waiting for my first boring day of work covering it. There's so much happening. And a lot of that has to do with the technology that companies like WAYMO are developing.

Boris Sofman Yeah, we joke. It's still not boring. Still not boring. So, yeah, Seth we got to get you out as soon as the world's in better shape to get a real ride. It'll be fun to talk to you and actually get you on a truck.

Seth Clevenger Yeah, I appreciate that. And you know, this has been a great conversation, but I see what we're short on time, so I think that's a great stopping point. We'll leave it there. Thanks again, Boris, for joining the podcast and sharing your insights. We really appreciate you taking the time.

Boris Sofman My pleasure. Thank you.

Seth Clevenger Before we close, let's take a moment to revisit our original questions, what will it take to make autonomous trucking a reality? And what will this technology mean for the future of the industry and its workforce? It's clear that automating highway driving is no simple task, but the technical challenges are only one piece of the puzzle. As we've discussed during this episode, a successful rollout of autonomous trucks will also require buy in from the trucking industry. Government regulators and the motoring public developers of this technology will need to continue building that public trust and confidence as they pave the way for commercialization. While the exact timing remains elusive, the trucking industry should expect to see various forms of self-driving technology reach the market in the coming years, but it will scale up gradually. Fully autonomous trucks capable of unmanned operation will be designed for specific applications. They won't be able to go everywhere or haul every load. And that means the trucking industry and society in general will continue to depend on well-trained, professional truck drivers for the foreseeable future. Ultimately, trucking fleets will most likely deploy self-driving trucks in a manner that complements human drivers and may even make their jobs better. Automating some longhaul routes, for example, could create more regional and local driving jobs that provide more home time for drivers. As technology continues to advance, trucking, like many other industries, will develop new ways to combine human labor with automation to improve safety and efficiency. If you've enjoyed this episode of RoadSigns, please let others know. Rate and review us on Apple podcasts and Spotify. If my questions have sparked questions of your own, share them with me and the RoadSigns team you can email us at share@ttnews.com. We'll read them and respond daily. And of course, we'll be back in two weeks with a fresh episode of Roadsigns. Until then, I'm Seth Clevenger. Thank you for listening.