Isn’t it too early for Waymo’s self-driving car technology?
Well, that didn’t take nearly as long as we thought it would, now did it?
Waymo has just announced it will be rolling out yet another fleet of self-driving cars, this time in the Phoenix suburb of Chandler. In itself, that is not big news; there are fleets of autonomous automobiles being tested all through the United States (and, thanks to recent changes by the Wynne Liberals, Ontario).
What makes this latest Google car unique — indeed, extraordinary — is that there’s absolutely no one behind the wheel. Yes, Waymo’s experimental Chryslers are the very first autonomous automobiles to be tested on public streets without having their computers backed up with a human co-pilot.
It’s almost impossible to overstate the significance of this experiment, not only as a cultural phenomenon but also as indication that Waymo — progeny of Alphabet-nee-Google — is seemingly technologically ahead of traditional automakers.
Even more pioneering/courageous/mad-as-a-hatter, Waymo is not restricting access to these new, now truly driverless vehicles to its engineers and employees, but is instead offering up its computerized Pacificas — FCA’s hybrid minivans — as publicfacing taxicabs. Yes, the Chandler experiment would appear to be the thin edge of the long-promised, self-driving taxi that Silicon Valley has been promising for the past few years.
If you’re thinking this sounds awfully early to be putting such trust in completely computerized cars, you’re not alone. Most of the industry has been predicting 2020 was the earliest anyone would dare send a car onto public streets without human backup. Now, three years ahead of time, the good folks of Chandler find themselves as guinea pigs to see if the driverless revolution we’ve all been awaiting is really possible.
Why now, why in Arizona and, perhaps more specifically, why such a small portion — reportedly, the test will be conducted in a tightly constrained, 10-squaremile area — of suburban Phoenix?
Well, as for why Arizona, the weather is perfect, the roads are wide and Chandler, like much of Phoenix, is not plagued with the pesky pedestrians that render computerized cars crazy. Oh, and the local politicians, eager for business, have instituted some of the most lax restrictions for autonomous automobiles in North America.
As to why now, that’s a more difficult question to answer.
According to Timothy B. Lee of ARS Technica, it may be a question of what is the right way to gradually phase in automotive autonomy. Traditional automakers, says Lee, have taken the conservative approach, preferring to roll out many partially self-driving cars — i.e., those with safety drivers behind the wheel — to test their computers’ abilities in many different locales.
Lee says that Google initially adopted the same approach, but quickly realized it was having trouble getting its safety drivers to remain attentive, comprehending that “employees who got to test early prototypes started trusting the technology way too quickly and started looking at their smartphones, putting on makeup and even napping in the driver’s seat.” So Waymo decided that it made more sense — indeed, it would be safer — to test 100 per cent driverless vehicles in a very restricted area and then build out the service from there.
Amir Efrati, writing for The Information, seemingly begs to differ. In his recent article, Waymo’s Foes: Left Turns and the Mean Streets of Phoenix, Efrati points out the many failings Waymo’s self-drivers have had, even in those familiar confines of sunny, spacious Arizona. One of the biggest issues, say Efrati’s sources, is plain old left turns. Yes, just like we humans (the U.S. Department of Transportation lists improper left turns across intersections as the most common cause of collisions), computers seem to have problems crossing divided highways. So much so, says Efrati, that Waymo is considering restricting its autonomous Pacificas from making left turns at intersections without green-arrow traffic signals. He mentions unmapped roads into malls as being another potential hazard and claims that cul-de-sacs also flummox the self-driving bits and bytes.
If Efrati’s assertions hold any water, why would Waymo go ahead with such a risky program? Well, one side of the argument would be the dangers are being exaggerated and the self-driving Pacificas will be as proficient as promised. Also, at least initially, there will be a Waymo employee in the back seat who can, if not take over control of the car, at least shut it down.
A cynic — who, moi? — might counter that Waymo’s leadership has more to do with corporate philosophy/experience, Silicon Valley having a long history of taking flyers with new-found technology and using its customers as guinea pigs. America’s tech giants are so used to foisting “beta” technologies on the consuming public that it’s not really surprising Google tries the same thing with self-driving cars.
Of course, the problem with this concept is that when Apple’s IOS burps, your iPhone just freezes. If, on the other hand, a self-driving car has a brain fart, the consequences could be infinitely more dire.
Indeed, perhaps the biggest reason Google seems so far ahead of traditional car companies in developing autonomous automobiles may have less to do with its technological prowess as the fact it’s never been on the wrong end of an automotive wrongful death suit. Having lived through those debacles — Audi and Toyota’s unintended acceleration, Ford’s exploding tires, etc. — automakers are understandably reluctant to send a completely unmonitored car onto public streets.
As simplistic as it sounds, then, Waymo’s leadership in automotive autonomy may be the result of never having faced the wrath of a self-righteous lawyer armed with the death of an innocent sixyear-old as exhibit No. 1.
Google may indeed be the leader in self-driving technology but methinks that Detroit could teach Silicon Valley a little something about the dangers of the American tort system.