Autonomous vehicles: They’re not there yet

Autonomous vehicles are here, and they’re causing some problems.

Reports over the past year show that driverless cars like the General Motors Cruise and Alphabet’s Waymo have been occasionally getting glitchy in cities like San Francisco and Phoenix. In one instance, a Cruise vehicle nearly collided with a San Francisco light-rail train. In another, five Waymo cars tried to pull over simultaneously when dense fog rolled into the city.

Andrew Hawkins, transportation editor for The Verge, says it’s a confusing moment for driverless cars. Most of the time, they work remarkably well, until suddenly, they don’t.

Marketplace’s Meghan McCarty Carino spoke to Hawkins about the state of autonomous vehicles and an industry beset by technological and financial problems.

The following is an edited transcript of their conversation.

Andrew Hawkins: Over the last year or so, we’ve seen mass layoffs hit a lot of the companies that are working on autonomous vehicle projects. In addition to that, there have been a lot of negative headlines about the rollout of autonomous vehicles. In cities like San Francisco, companies like Cruise and Waymo are ramping up their robotaxi services and are being met with a lot of outright hostility from the people who live in those cities. That’s mainly because the vehicles are not acting as perfectly as they were originally portrayed.

These vehicles are blocking traffic, they’re rear-ending city buses, they’re getting confused, and it just seems like we’re in this strange liminal period right now. Autonomous vehicles are coming out and they’re real, despite some predictions that they would not actually make their way onto our public roads. Now, they are actually on public roads, they’re making money for some of these companies, but I think we’re in this strange in-between period. We’re experiencing the very well-predicted tensions between these robot-driven cars and our very human-centered street and road infrastructure in the United States.

Meghan McCarty Carino: Are there specific technical issues that have proven difficult to overcome?

Hawkins: The issue that we’re facing is what’s referred to as the long tail of nines. It’s the idea that you can get a vehicle that is 99.9% as good as a human driver, but you never actually get to 100%. And that’s because of something that’s referred to as edge cases.

Edge cases are these things that are unpredictable events that occur out on the road. It’s almost impossible to ever get to the point at which the vehicle is at the same level, or even better than, a human driver. And that’s what we’re really going for here — a technological system that is not just as good as human driving but is actually an order of magnitude better than human driving in order to make the case for itself. It has to make the case that this is something that we should use to replace the current system that we have today.

McCarty Carino: A lot of the problems that have arisen have seemed to be low-stakes issues and near-misses that didn’t result in injuries or death. What do these kinds of problems tell us about where this technology is and what the remaining challenges are?   

Hawkins: They tell us a couple of things. But I think most importantly, it tells us that as human beings, we have an extremely high tolerance for human-caused chaos, crashes and deaths that occur on the road and an extremely low tolerance for any sort of problems caused by robots. That’s because we’ve trained ourselves over generations to assume a certain amount of risk when we get into our vehicles in the morning and drive to work or when we step out onto the street and interact with cars. It’s very possible that we might crash our car or that a car might run into us and kill us, and that’s just a risk that we’ve accepted as human beings in order to have these vehicles that can take us where we want. But when those cars are driven by robots, all of a sudden, that tolerance and that risk acceptance evaporates almost immediately. And if it’s a robot that’s responsible for the mistake or the crash or the traffic jam, that’s when the fingers start being pointed and the blame starts flying because we’re very used to the former system, but we really have no experience with the system that’s about to come.

McCarty Carino: When it comes to the market potential for this technology, one report from Intel in 2017 predicted a $7 trillion economy annually for autonomous vehicles by 2050. Does that seem realistic?  

Hawkins: They were just laughably off-base. That’s not to say that this won’t eventually start to bring in revenue, but it’s certainly not at the moment. Some companies are able to charge passengers for their rides, but there’s some regulatory hurdles still.

There’s a question of whether or not the industry will actually begin to see a return on their investment because this technology has cost hundreds of billions of dollars. I think that that’s why we’re seeing a lot of contraction right now: startups that are shutting down, layoffs, consolidation, pauses in commercial deployment. This is an industry that is very much a product of a zero interest-rate, pre-pandemic environment. That’s not the environment anymore.

McCarty Carino: Is there anything that’s going well for autonomous vehicles?

Hawkins: The companies that are on the road and are doing the testing and have been at the forefront, they are continuing to expand. They’ve had to lay off employees and they’ve maybe had to go a little bit slower than usual, but it does seem that very slowly, this technology is spreading out over the country. It stands to reason that it will eventually start to touch more people’s lives and people will start to see how it actually can integrate into their lives and whether or not this is a product and a commercial service that they would want to use. At a certain point, the autonomous vehicle companies are not going to be competing against each other anymore, they’ll be competing with human-driven transportation systems and services like Uber and Lyft.

I think it’s going to be a question of how quickly they can get these cars out on the road and what markets they are going to be focused on. Will this technology become anything more than just a fad or a tourist attraction for people? Like, will people go to San Francisco, ride the autonomous taxi, take some selfies and then go back to their human-driven lives? I don’t think that that’s what any of these companies actually want, but I think that there is a piece missing right now to ensure that they’re going to be able to get over that hump and into a bigger market with more possibilities for them in the future.

Related links: More insight from Meghan McCarty Carino

If you’re hoping to get your own selfie in the backseat of an autonomous vehicle, San Francisco is a good place to start, as Hawkins said, but Phoenix is also home to a growing AV population.

Waymo is starting to roll out driverless cars in Los Angeles. Eager Angelenos can get on the waitlist to be among the first passengers. (Hey, maybe I should do that!)

Meanwhile, Cruise is expanding further into Texas, with cars already on the road in Austin and soon in Houston.