Human vs. Machine: Self-Driving Trucks

in a self driving truck. Ready. Yeah, I was sitting in the back of a cabin on a bed, actually in a freight Liner Cascadia. That's like the biggest, baddest truck you can buy. We're on Florida's Highway twenty seven with the Everglades on either side. Truckers call this part of Florida Alligator Alley. I was riding along with a team from star Ski Robotics, and then the driver, Jeff Runions, took his hands off the steering wheel and slid his feet off the petals. Oh my god, this sounds terrifying.

And you're on a bed. I mean there are there even seatbelts no seatbelts, and Brad, I don't know if you realize how big and heavy these trucks are, and they go fast, and all of a sudden he takes his feet off the petals and then a gust of wind hit the truck and the computer sort of lurched us towards the left wane. He's trying to fight. I was freaking out, but everybody else in the truck was

acting pretty much normal. Maybe they knew something about the air bags in the back of the vehicle that you didn't. There weren't any air bags, but there were algorithms, and those algorithms need a little bit of time to understand the driving conditions on a given day and to figure out just how much to sort of push the steering wheel to fight against the wind. You are in the morning,

when before you have your bath. Star Ski is still fine tuning its technology to make sure it works in all scenarios, and if they succeed for truckers, which is one of the most common jobs in the United States, that could mean huge changes. Back. Hi, I'm Brad Stone and I'm Max Chapkin, And this week on Decrypted, I

ride a robot truck from Fort Lauderdale to Tallahassee. From small startups to Silicon Valley's biggest tech companies in the world's largest automakers, there's a race to build a software that could power a future of cards without drivers. Long haul trucking will likely be one of the first areas to feel this profound shake up. This technology promises to make our road safer and our goods cheaper, but it's

a transition that will have consequences. Trucking is worth seven hundred billion dollars in the U. S Alone right, and it employs a lot of people. More than three million Americans drive trucks for a living, and another four million support the trucking economy and jobs like waitressing and roadside cafes and being a gas station attendant. It's also a job that lower income Americans depend on. What happens to these workers is a question that needs to be answered.

This episode kicks off a series of pieces about technology that threatens to replace jobs. It's one of the most important tech stories of our time. If you have a story to share with us about your job and the oncoming wave of automation, record a voice message and send it to decrypted at Bloomberg dot net. So Max, Right now, self driving car research is absolutely dominated by big name comp and is Google, Uber, Ford and and GM, pretty

much every Detroit automaker, the Chinese companies. So how did you find the small start up Starsky Robotics and why does the startup even matter? Right now? What's interesting? I got tipped off by one of their investors. They were backed by y Combinator, so shortly before demo day, I got tipped off, and people suggest I talked to him. But I think it's important to remember that a lot of these big companies working on driver list tech were

really small, just very recently. The other thing that that drew me this is this guy was a little different. So I went to South Florida to see Starsky's trucks in action. My name is Stephan Seltzox Marker, and I'm the CEO and co founder of star Skirobotics, self driving truck company. Stephen is Starsky CEO. I found him interesting partly because of what he's not like. So I didn't go to Stanford. I didn't participate with two thousand four

Dropper Grand Challenge. I'm not a Google exer. I don't really have any of the credentials of someone who's supposed to be good at this space. What interested me about Starsky was the sort of difference there, and also the fact that they had this different approach, which is that they're planning to use remote control drivers, basically people in

call centers who are overseeing trucks. So if you're a truck driver, you don't necessarily have to spend three weeks in a truck, sleeping in your cab, eating truck stop food, talking on a CB radio. You just go to an office, you sit in a desk, you drive a truck for eight hours, and then you go home and see your family. I'm imagining this new new job is a little bit like playing Mario Kart for a living. It actually sounds kind of fun. The other thing is that Starsky, unlike

other truck companies, is actually hauling cargo. So while they're collecting data and trying to log all these research midles, like all the other companies, they've also got like a load in the back of the truck that they're getting paid, you know, a modest amount of money for right, And we should know that many of these other companies have deployed lots of capital to to this kind of research. Starsky is only raised around five million dollars. Is the

strategy of using real cargo presenting a different set of challenges. Though, Yeah, it sounds easy to run a trucking company, but it's not necessarily their Their first truck, which they nicknamed Rosebud, broke down even before they could get it out of San Francisco. And what was wrong with Rosebud, like, well, the bad alternator or something. So have you ever bought a truck before? Neither? And I that's that's the core problem with Rosebud. I bought. I bought a truck off

an auction site. Starsky, as we mentioned, is just one of many companies working on this. The Silicon Valley companies of course, Detroit and and the big companies like bay Do in China that think they can make an impact, and Brad, don't forget the big trucking companies Volvo, Daimler, all the big names and trucking are also getting it in the space. So where is Starsky relative to these giants in terms of progress? So what stars He's going

for is what's known as level three autonomy. And let me just explain quickly Level one autonomy is cruise control, your basic thing that is in lots of cars. Level two is cruise control and what's known as lane keeping, where the car will sort of steer for you. Okay, so take your hands off the wheel, but for god's sake, don't take a nap, right And Level three, what we're talking about with Starsky, means that the driver needs to be there in case of an emergency, but the driver

can basically not pay attention during other times. Level four means driver can be totally out of picture, could be unconscious, could be texting whatever. And that's where the big companies, the Googles and forwards are focused on, honestly industry, yeah, whereas we want to be like a very real business. Len to be real business. Now, when Stephan says he wants to be a real business, he means that Starsky's plan is to have the driver only intervened in an emergency.

But unlike say Tesla, which is also going for this same idea with a driver in the driver's seat, Starsky's idea is that the driver is going to be someplace else. We're building hours, so it's not needed person behind the wheel Because the point for US is to not have a person behind the wheel. So this is kind of

like the military's use of drone pilots. Yeah, exactly. So Starsky when they raised their five million dollar funding round, they did so thanks to a video that they made where they were remote control driving a truck in a parking lot and it was making a sort of pseudo delivery. Their plan, and they want to do this by the end of the year, is to have a real delivery with real cargo where they pick up at the port, drive the whole distance, and then drop off, all using

teleoperation using a remote control driver. And would that be the first such achievement of its kind as far as I know, Yes, And and Starsky is actually closer to this milestone than you might think because trucks, unlike say, cars, spend most of their time on highways, which is of course the easiest place to do autonomous driving, right, no pedestrians, no bikes, no traffic lights. Yeah, and and trucks are big and heavy, which means that it's a lot easier

to get the sensors on them. And there's also this safety component. Trucks drive a ton of miles and they're involved in a lot of fatal accidents. So in theory, you could eliminate a lot of those if you'd let a computer do most of the driving. But of course there's a dark aligning to this potentially silver cloud. Trucking is a huge sector of the economy, employing millions of people. Having a computer drive the truck raises the specter of job losses on an epic scale. So it makes me

want to run for the hills. It makes me want to hug my kids. I hope they have a future. That's Ed Witkin, the president of the Transportation Trades Department for the a f l C i OH. He's basically a big union muckety muck in the world of trucking. Um unions including the a f l C i O and the Teamsters, which are the big well known trucking union, are warning that this could be really, really bad for their members, which isn't surprising, right. It taps into an

almost primal fear. Computer come, they take our jobs, right, But trucking is weird because it's one of those jobs that a lot of people don't really want. This is me, I don't take take companies because truck and burgers sixties seventy hours. Some of the like going on players of working eighty hours just to trying to pay that drop. I had a big, stretched out truck, big thing. But you know I live in this drug staying it for a couple of days, and you'll know what I'm found about.

That's Jeff Runyons again. The truck driver was behind the wheel on the drive in Florida. He told me what being a long haul trucker was like, plus to eat terrible. Yeah, you know, you get truck stout. I used to go to memphitt Of all the time. That's why I was two fifty. Stopped at the drug stops. You got green fright to make in the cafes. You go to any truck stop and you'll see everything's fatty. Yeah, yeah, yeah. I think companies treat you like they say drivers or

dom a dozen't so. Despite the glorification of the occupation in the seventies and movies like Smoking in the Band at bad Food, poor treatment from your employer doesn't sound great.

And the wages are bad too. I mean, people think of trucking as being this this kind of good paying job, but the truth is the average trucker makes around forty dollars a year, and that easily includes you know, sixty seventy hours of work every week, so in the end, you're not making that much better than minimum wage, right, And that explains why the industry sees such high turnover. There's a massive labor shortage. And this is what you hear if you talk to people who are who are

involved in autonomous trucking. The American Trucking Association, the big trade group, says that around fifty thou trucking jobs are unfilled at the moment, so there is appetite for more and more truckers. That's my colleague josh Idolson, who covers labor relations for Bloomberg. The volume of stuff to move around the country keeps getting bigger, and the number of people coming in to the industry to drive and staying in it is not keeping up, in large part because

so many people leave. Turnover is so high because being a trucker is pretty much terrible, but in some of the poorest parts of the country, it's also one of the only jobs available. One way to look at it is these are some of the most cherished, mediocre jobs in the United States. What about military service? Max is that comparable. It's yeah, Actually, it's it's very comparable. Josh says that, um, it's the it's people who are who might enlist in military service that the trucking companies are

competing for. Maybe if they could get the federal law adjusted so that people who are younger than twenty one can do interstate trucking, maybe they could get people before they would otherwise decide to go to the military, for example. And this is why Starsky and the other Silicon Valley companies are so interested in the sector. This is a rare industry where computers could really help what the technology is and what it will mean for trucking jobs. That's

coming up right after this word from our sponsor. You know where you want to be. Red hat has the broad portfolio of open source technologies to get you there. Meet your evolving business challenges head on with secure solutions for the enterprise, including Linux platforms and containers, hybrid cloud infrastructure, application integration and development, operations management, and beyond. Visit red

hat dot com slash open tech to learn more. Red hat the open technology to help you realize your vision before the break. Max, you're explaining that even though the trucking industry employs millions of people. It's actually facing a big shortage of workers, and automating the driving process could directly address this. So let's explain what exactly automation involves. Okay, so there are two basic components that the sensors that

are collecting data and the algorithms. So the sensors they tell the computer where the vehicle is, what's around it, and then the algorithm crunched the data to figure out what it should tell the vehicle to do. Companies that are aiming for the most advanced autonomy use something called LIGHTER, which stands for Light Detection and Ranging. Of course, it's somewhat controversial because Uber and Google are now fighting over some intellectual property around lighter. Star Ski like Tesla uses

raiders and camera. Yeah, it's a lower tech approach and Stephen laid it out for me during our ride. So like what we have here in the front are like to automotive grade really high quality cameras yea with high body lens feed until our software got it. So while we drove, Jeff the driver basically kept his hands off the wheel and and Kevin Keo, a star Ski engineer, kept an eye on the systems using a laptop and

a tablet that was built into the dashboard. So at the moment and we have we have mop based roots, we do, but at the moment it relies heavily on the landmarks in front. Got it getting some information from the GPS, but currently and it's getting most of its information from the laemark so similar it is like teslaut. So the car is basically watching the road through cameras and sensors and trying to stay in the lane the way a real driver would. So Max what happens if

the lane markings are faded or simply aren't there. In general, interstate highways are pretty good, and Starski is making life a little easier for itself by saying that the driver, not the computer, should handle those edge cases. That's that's part of the same reason why Google has done like three million miles and is it ready to play anything. It is because like they're trying to look for all of those weird curves, all of like the weird times

where U where whatever weird thing happens. Now Google wants it's a computer to basically handle everything. Yeah, Starsky's thought is the safety driver will be sitting in a call center somewhere monitoring maybe five or ten trucks on alarm will go off if something weird happens, and then he'll immediately seize the wheel and take control. Let's hope the wireless vous is reliable. How how does that work? Starsky doesn't want to say exactly how they're transmitting data, but

basically they're controlling the truck wirelessly by remote control. Right now, the system looks a bit like you what you'd have if you were really into video games. There's a steering wheel and then three screens, and the driver just sits there and sort of drives like in a video game. But the idea is that most of the time the computer is driving itself. Now, let's go back to the beginning. You're in one of these trucks in Florida. You're on a bed in the back, and and the wind is

blowing causing the truck to sway and the computer to overcorrect. Right, we were hauling a lot of weight on this particular day. It was It was a shipping container which is been out of stainless steel. It's heavier than the standard illuminum boxes that Starsky had been testing before, and that proved slightly problematic. Here's gonna be something new right here. This, Yeah, now this curve during the week we were taking a

point container, we'll see a new experiment. Basically, we're getting pushed by the wind and the computer a bit like a human driver was struggling to figure out exactly just how hard how to respond, just how hard to turn that wheel? Max. You deserve hazard pay for this? Why why were they doing this test in Florida. So a bunch of states are trying to regulate driverless cars, but

Florida is really really ridiculously relaxed about this industry. You don't need a special permit or any special insurance to to drive a driverless car in Florida, and you don't even need a human driver behind the wheel as long as somebody somewhere is in control of the car. Now here's the most interesting part. It gives us a peek at the future. Maybe robots won't take all our jobs, but it's likely that they replace many jobs, including what we think of as blue collar jobs that involve manage

a computer. So I think like humans and technology are probably going to be better than either want to. Yeah, probably forever. That's Stephan again, the stars Key CEO. Stars Keep brings together these two unlikely groups rod En m Gene seven dollar lotts, Yeah, mixed with truck drivers who some of them live in trailers, many of them like they never went to college. Self driving engineers like Kevin Kyo, the engineer who works for Starsky, can make millions, even

tens of millions at companies like Google. Truckers are basically treated as disposable by their employers. Were we were straddling people who thought that Bernie didn't go far enough and people who didn't think that Trump went far enough. Stephan argues that this is going to be really good for truckers, which I also heard from Jeff who was our driver for the day. Some people who are negative, Yeah, and then some are really interested in attack behind him. Yeah.

And then you tell them about how you're gonna make it so they have forty hour weeks that seven the hours the weekend have, you know, instead of me and gone all the time. They can make same amount of money. But Jeff, a star Sky driver, isn't exactly an objective source here, is he? No, And it's a real threat that many drivers will get left behind. That's what one driver, Tom George told my colleague Josh. Even where they're not a driver shortage, there would certainly be technology trying to

to edge into uh, supplanting a driver. I've been seeing uh trucking, uh just truck manufacturing developing into a you know what I always called trying to engineer the driver out of the truck. So Max, it's hard to know what the long term impact of artificial intelligence is going to mean for all of us, not just truckers, right right, and and there are there are a few different schools

of thought, which I'll break down for you quickly. The first is basically that advances in robotics artificial intelligence are going to make us more productive, make us superhumans, and and those of us who lose our jobs going to get even better, more creative, more fulfilling jobs. The second idea school of thought is that technology is going to kill so many jobs that's going to create this permanent underclass where where you have a bunch of people who

can't find anything. And the third school of thought says this is all hype, don't pay any attention to it. Yeah, I've also noticed the strain of optimism and Silicon Valley that says the job destruction will be real, but this this is a natural cycle of the economy, and that it will evolve a new work that we haven't yet imagined will kind of take care of the disruption, right. And I think it's important to remember that that for most of the country, this sort of Silicon value way

of thinking is not how they see it. They see this as basically science fiction. Just the other day, Transportation Secretary Lane Chow told reporters that explaining self driving technology of the public would be one of the biggest hurdles in terms of getting people to adopt it. And even truckers like Tom George, who we just heard from, are convinced they'll be able to drive trucks better than a

computer for some time to come. That's gonna be really, really hard for automation to take over because there's so much interaction that the driver has to be a part of that can't be automated by a computer. So who's going to notice that a strap is loose on your load while you're driving down the road. I don't know that the computer is going to be able to first bull notice it. And second of all, how are they going to just do it? What are they gonna do?

And Tom's not wrong. It's true that there are thousands of rare cases, those edge cases, those emergencies where computers don't work very well and they may not work well for a long time. Yeah, but that doesn't mean it's going to take a long time to get to that hybridge stage. And Starsky is a version of this with remote control. But there are others. Auto and another driverless trucking startup called Embark talk about having drivers in the

cabs but allowing them to sleep. Peloton, which is another Silicon Valley trucking startup, has a convoy style approach where you could imagine a big line of trucks with one driver just in the refront. But the bottom line is that all of these companies, all of them are trying to make trucking require fewer workers. So for now, what's next for Starsky? So what they're trying to do is have this big test on the highway where they're going to try to drive a whole uh delivery without having

a human in the car. And I think probably they'll they'll try to raise some more money. Meanwhile, I guess things are looking less optimistic for drivers who don't have a Silicon Valley startup to get involved with. It's not just about drivers and people that operate trains and buses. It's also about people that do sort of ordinary jobs that you and I encounter during the day. That was

Ed our union leader again. Um Ed's bigger idea was that at a certain point the government needs to take some responsibility for job losses and and making sure that this scary underclass scenario never plays out. And the other thing worth mentioning is that ordinary jobs quote unquote might also include a bunch of the jobs that we think of as being good lawyers, some kinds of doctors and Brad I hate to say it, but maybe journalists. What

about podcast hosts? Or they say always forever? But Max, using your experience at the back of this truck, do you really believe this is going to happen in our lifetime? Do you really see those engineers feeling comfortable stepping away from the from the vehicle? And then, you know, can the same AI techniques apply to these other occupations where you know you do need human intuition, human input. At least for now, I don't think it's gonna be such

a stark change. And I mean there's with trucking, there is this labor shortage, so there's sort of a built in cushion, and a lot of people said things like anyone who's a truck driver in their sort of career is not going to be seriously affected by this. I'm not sure how the true that is. And I think the other thing that we have to keep in mind is a lot of time, these automations don't happen in

one big chunk. They happen in sort of small ways that that affect people's jobs, causing them to make less money. I mean, one type of automation that's pretty underrated is the advent of automatic transmission, which coincides with the declining wages of trucking. Now that also there are other causes of that, but but one thing it did is make driving a truck easier, which meant that trucking companies could

use lower skilled workers. It kind of goes along with this notion that AI is this magical thing that doesn't exist, and yet when it does exist, we don't call it AI anymore. We just sort of assume assume that it's there. It's it's machinery. I like the Starsky approach, this melding of AI, telematics and and then humans because you know that this is the model that we've seen in medicine, in law, really in journalism technology not replacing humans but

making them more efficient. Yeah, it strikes me as as as fundamentally like realistic, but also it has a certain humanity to it. And what I liked about this company, and I think what the companies that succeed in in this world are going to have at least some empathy for the workers who are being affected by by these changes. And those that are just thinking about this as pure technology, I think are going to run into trouble because because

they're not being realistic about how the world actually works. Yeah, the star Ski approach does see him like it's something that is a little bit more practical sooner. But of course I didn't ask you the biggest question. The most important question is what's what the name? Is? This a reference to the popular seventies TV show It is so Stephen told me that he was actually looking for CB Radio Lingo, so like ten four good Buddy type type

of things. And I guess Starsky and Hutch was slang CB slang in the seventies for a team of drivers, so he was thinking that it was going to be like you have a call center for drivers, So it's sort of like a star Ski and Hutch, or maybe a lot of star Skis and a lot of Hutches. Good buddy, and that's it for this week's episode of Decrypted. Thanks for listening. We want to hear your stories. Tell us about your job and whether you're worried about technology

and automation coming for you. Recorded voice message and send it to decrypt it at Bloomberg dot net. Also, I'm on Twitter at Chapkin and I'm at brad Stone. If you haven't already subscribed to our show wherever you get your podcasts. While you're there, please leave us a rating in the review. It really helps more listeners find our show. And by the way, we're crying a new thing on LinkedIn where we give you a behind the scenes look at our episodes each week and the discussion on some

of the thornier problems we'll discuss on the show. LinkedIn also has their own tech show called Work in Progress, where they talk about the future of work. One of their recent episodes is about universal basic income, which is also something we've been looking into here at Bloomberg. This episode was produced by Piagkari Aki Ito, Liz Smit, and Magnus Hendrickson. Thanks to Nico Grant and Isabel Gottlieb for

all their work on this episode. My Business Week story was co written by jos Idolsen and edited by Jim Alee and Nick Summers. You can read it on bloomberg dot com slash business Week or on the brand new Business Week app Alec Mchabe is the head of Bloomberg podcast We'll See You next Week over Now. Decrypted is brought to you by red Hat, whose broad portfolio of open source technologies for the enterprise helps you get from

where you are to where you want to be. Red Hat the open technology to help you realize your vision. Learn more at red hat dot com slash open tech.