What I, along with many others, thought would be the massive technical improvement of the 2020s seems to have failed. Four years ago, there were so many articles about that effort that I could have dedicated this entire blog to it, and in fact, for a few months, I did just about that. It seemed worth it, as it promised titanic changes, equivalent to cars becoming widespread a century before, many unexpected. As the Ford assembly line was to transform romance, autonomous vehicles, just to name one, might have slashed smoking, as vehicles filling their tanks without occupants would have made human trips to gas stations, during which many cigarettes are bought, unnecessary.
Here is my latest file of what has been written. Given that I did not save pieces only on future possibilities, whether judicious or hyped by companies providing them, I had to go back nine months to have enough for a post.
The oldest is Aaron Mak’s “The Most Disturbing Part of the Latest Tesla Crash,” from April 19th in Slate. This story was about an incident not involving a fully autonomous car, but an electric vehicle with an “autopilot” feature, abused by the occupants as neither was in the driver’s seat. That was nothing against current or future autonomous vehicles, but was ripe to be misperceived as such.
Next is “Silicon Valley is resetting expectations for self-driving cars and settling in for years of more work” (The New York Times, May 25th). The unbilled author started with “after years of hype, bullions of dollars of investments and promises that people would be commuting to work in self-driving cars by now, the pursuit of autonomous cars is undergoing a reset,” and seemed to blame excessive expectations, yet in 2017 such vehicles were already quite adept and drew intense and extensive efforts zeroing in on remaining problems. Per this piece, Uber and Lyft, expected to be early and large customers, have sold and “offloaded” their driverless divisions, and, in the year before the article date, “three prominent self-driving start-ups have sold themselves to companies with much bigger budgets.”
Less optimistic is Ashley Nunes’s May 31st Business Insider “The dream of the truly driverless car is officially dead.” In this self-described “opinion column,” Nunes wrote about the lack of truly self-sufficient systems elsewhere, without acknowledging the value of a mere reduction in human involvement. Despite requiring overseers, store self-checkout lanes have hardly petered out, as the number of people per unit is less than one, and the same could work here. The reason for autonomous vehicles not becoming widespread is elsewhere.
Distant emergency command is prominent in “How Germany Hopes to Get the Edge in Driverless Technology” (Jack Ewing, The New York Times, July 14th). A Hamburg “ride-hailing service” used vans which could “steer themselves, but technicians working from a remote control center keep an eye on their progress with the help of video monitors,” allowing these workers to “take control of the vehicle and steer it our of trouble.” That was the old “Level 4 autonomous driving, in which a vehicle can steer and navigate by itself most of the time but may occasionally require human intervention.” Although German law still required at least one human driver or supervisor on board, and such vans were only used to “navigate a predictable course, such as from an airport parking lot to a departure terminal,” such niches may be where autonomous efforts can thrive.
The count of foolish accidents is probably increased by sloppy use of promotional language, as shown on August 30th in the New York Times, in Cade Metz’s “Tesla Sells ‘Full Self-Driving,’ but What Is It Really?.” That is the name of a “package… which can be purchased as an extra on Tesla cars,” and is no such thing, rather “a collection of services that add to Tesla’s Autopilot,” also an overstated title. Indeed, that is the subject of a customer lawsuit mentioned here, alleging misrepresentation.
Beyond “Alphabet’s Waymo launches autonomous taxi service test in San Francisco,” by Brock Dumas in Fox Business on August 24th (didn’t that happen in 2018?), we got a post-mortem of Tesla’s bad things mentioned here in “Inside Tesla as Elon Musk Pushed an Unflinching Vision for Self-Driving Cars” (Cade Metz and Neal E. Boudette, The New York Times, December 6th). Musk, Tesla’s owner, understandably touted his technology, but seemed to have conflated what the hardware could do with progress in software and testing.
What can we conclude about the true state of driverless cars? Metz, two articles above, quoted a Consumer Reports’ Auto Test Center director as saying “These systems are good at dealing with the boring, monotonous stuff. But when things get interesting, I prefer to drive.” That is our overall situation. I had thought that with test-ranges and scrutinizing problem situations, the technology would be developed, and it still could be. I refuse to believe it cannot be done.
The real problems are societal. Can we accept occasional, even fatal, mishaps in the service of removing America’s 27,000 annual driver-error-caused deaths? Can companies find enough people willing to go years without product for a potentially gigantic long-term gain? Can we regain the national will to embark on long, expensive, and uncertain efforts? Until then, all we will have here, outside a few well-protected enclaves, is driving assistance technology. That, though valuable, is a far cry from the benefits true vehicle autonomy would get us. Yet again, the choice is ours.