Of the futuristic technology we see in movies and television, few have the appeal — and the sense of being tantalizingly close — as self-driving, or autonomous, vehicles. Today, we have cars and trucks with some autonomous or driver-assist features, but they aren’t quite the promise of the driverless cars we see in science fiction.
The conversations about this mostly revolve around passenger cars navigating in cities, which still depend on a person who can take the wheel in case the computers crash and the technology fails. But perhaps we are a bit closer to that futuristic scenario when it comes to self-driving semitrucks.
Christopher Mims, a tech reporter at The Wall Street Journal, recently wrote about the specialized technology behind autonomous big rigs. Marketplace’s Kimberly Adams spoke with Mims about it. The following is an edited transcript of their conversation.
Christopher Mims: You need a lot of different kinds of sensors to perceive the world around the truck — lidar, which is lasers; you can have radar, which is just like this stuff at the airport. And then, all that gets fed into a system that is a hybrid between [artificial intelligence] and so-called rules-based systems. So that system may have learned like a traditional AI, but it may also have a bunch of modules in it that engineers just kind of built. And all of that, of course, gets informed by the many miles that these trucks have already driven on America’s roads.
Kimberly Adams: Yeah, you highlight that in your piece, that in some ways it’s easier to do self-driving technology for big rigs than individual passenger cars. Why is that?
Mims: Big rigs spend most of their time on highways, and highways are simpler. They’re well marked. There’s no cross traffic. They’re not zero pedestrians on highways, but there are many fewer. That said, even though weird things might happen less often on the highway, it’s no less critical that these systems successfully deal with them.
Adams: How does this technology compare to tech that car companies are using to try to get us to full self-driving car mode?
Mims: So the main way that self-driving big rigs differ from self-driving cars is that engineers have said, “Hey, self-driving big rigs can still be useful, even if the self-driving tech really only works under very limited circumstances.” So on these highways, on these routes, at these times of day, in good weather — that would be not so useful for a robotaxi to limit its operation that way. But for a big rig, I mean, those are miles and miles and miles that humans are driving now. If you can swap that human out with a robot, there’s all kinds of economic reasons you might want to do that.
Adams: Can you talk about some of the economic rationale, then, for these trucking jobs to become automated in this way?
Mims: Part of the rationale is just eliminating labor costs. Another part is that a robot truck can be on the road nearly 24 hours a day. And so companies get a lot more use out of those trucks, and they can start competing with routes that in the past you might have had to fly on an airplane. But now, you can put those goods on a truck and will get there about the same amount of time.
Adams: Which companies are investing and leading in this effort to create these autonomous big rigs, and what have they said about the timelines?
Mims: So there are two companies that recently went public — and, frankly, are under a lot of financial pressure from investors to deliver — who have promised the earliest arrival date for completely driverless trucks on America’s highways. And those are TuSimple, and then there’s Aurora. The third company that’s taking a little more of a cautious approach is Waymo. I mean, they’re the ones that are already running autonomous taxis in three different locations in America. And they seem to have realized as well that they can probably start making a profit on autonomous trucks, even before they make a profit on robotaxis.
Adams: As somebody who follows this pretty closely, what do you think of these goals of some of these companies to have this stuff ready to go by 2023? Think it’s going to happen?
Mims: Well, no company has ever met its stated deadline for when it was going to roll out commercial autonomous technology. So I’m always really skeptical. That said, I do think that given the demonstrations that we’ve seen already, we are close to robot truck drivers that are safer than humans, in the circumstances in which they’re driving. So will we see it by the end of 2023? Maybe not. But could we see it soon afterward? Absolutely. I think that it is coming soon. And perhaps more importantly, once it’s here, it could grow at a rate that’s faster than people expect.
Related links: More insight from Kimberly Adams
Mims’ piece for The Wall Street Journal has photos of some of these autonomous trucks, which basically look like regular big-rigs, just a bit sleeker and with some extra gadgets onboard.
Mims also pointed us to a recent documentary by The New York Times looking specifically at Tesla’s effort to reach a self-driving future and the cost of the speed of those efforts.
The National Highway Traffic Safety Administration looked at almost 400 crashes over 10 months involving “advanced driver assistance technologies.” In those crashes, six people died and five were seriously injured.
And the publication Innovation and Tech Today poses an interesting question as we move toward a more autonomous future: How is car insurance going to work?
The future of this podcast starts with you.
Every day, the “Marketplace Tech” team demystifies the digital economy with stories that explore more than just Big Tech. We’re committed to covering topics that matter to you and the world around us, diving deep into how technology intersects with climate change, inequity, and disinformation.
As part of a nonprofit newsroom, we’re counting on listeners like you to keep this public service paywall-free and available to all.