My article last week—Apple Car or not, driverless cars are an ethical dead end—looked at driverless cars and the associated ethical issues they raise, and it elicited the sort of response I’d anticipated from tech true believers. I especially got a kick out of the “ignorant troglodyte” epithet lobbed my way. And “stuck in 1995?” Try 1965, which was around when my automotive enthusiasm really began to gel. No apologies for that. Jean-Louis Gassée in a blog I reference further below, quotes his late countryman, the French literary theorist, philosopher, linguist, critic, and semiotician Roland Barthes, once observing about the contemporaneous Citroën DS 19 (the car driven by the Patrick Jane character played by Simon Baker in The Mentalist is a quite similar later DS 21): “I think that cars today are almost the exact equivalent of the great Gothic cathedrals; I mean the supreme creation of an era, conceived with passion by unknown artists, and consumed in image if not in usage by a whole population which appropriates them as a purely magical object.”
I like computers, too, as tools, and occasionally as objects d’ art, but the hum of a power supply and images on a screen will never match the orchestral musicality of a V8 engine exhaling through a set of dual tailpipes, the tactile pleasure of executing crisp gearchanges with a good manual transmission, or the overall satisfaction of driving quickly (and safely) through a challenging stretch of twisty road.
I was disappointed by the dearth of reasoned argument regarding points I raised in the article. There was plenty of ad hominem critique of my alleged ignorance and obsolescent thinking, but nothing on the issues cited themselves.
And even if we set the associated ethical complexities aside, there are still many practical hurdles such as how driverless cars would be able to cope with snow and ice covered roads (which is what we’ve had unrelentingly in most of the eastern U.S. and Canada for about three weeks now), remain ignored and unanswered. Ditto for dirt roads with no markings and minimal signage. I would be interested in hearing how the techies propose to address those issues. It’s one thing to test driverless car technology in sunny California, quite another in Maine or Massachusetts in February. Just one of many examples why driverless cars won’t be going live for consumers anytime in the near future.
Happily, I’ve subsequently run across a few really good reasoned arguments supporting the views I expressed in the article. Former Apple exec. Jean-Louis Gassée in his MondayNote blog this week first throws cold water on the Apple iCar rumors, reasoning sensibly that profit margins in the carbiz are way too thin to be of interest to Cupertino. He points out that Ford Motor Company made $835M net income last quarter, which translates to less than 4% of their $34B in sales. Meanwhile, Apple turned a record-breaking $18B profit in the quarter. Mr Gassée observes: “There isn’t an inkling of an explanation for why and how a superior product designed and built by Apple would bring superior returns,” concluding that “The fantastic Apple Car is a fantasy.”
Former GM chief Dan Akerson, who knows a bit about the car business, agrees, commenting to Bloomberg News that Apple would find itself in a highly competitive, highly regulated market with low margins and high costs.
As for autonomous cars, Mr. Gasseé acknowledges that they’re good PR, but says that while to some they may seem like an inevitability, it’s quite possible they may never actually happen, citing some of the same obstacle issues I did, noting that, “Actual, in-the-wild autonomous driving is fraught with countless intractable exceptions. What happens in heavy rain or snow, or when the software behind the camera has trouble recognizing objects that are blown onto the road? What happens when your car approaches a a last minute detour around new construction site?”
These are apparently real world problems robocar fanbois don’t want to address. Writing in Slate Lee Gomes notes the “… surprisingly long list of the things the [Google autonomous] car can’t do, like avoid potholes or operate in heavy rain or snow. Yet a consensus has emerged among many technologists, policymakers, and journalists that Google has essentially solved—or is on the verge of solving—all of the major issues involved with robotic driving.” That pointing out such uncomfortable truths is flame bait, indicates to me that a sort of “The Emperor’s New Clothes” dynamic appears to be setting in.
Gomes, who has written about the Google robocar for the prestigious MIT Technology Review, suggests that it “… may turn put to be the automotive equivalent of the Apple Newton,” a “timid, skittish robot car whose inferior level of intelligence becomes a daily annoyance.” He observes that in order “… to be able to handle the everyday stresses and strains of the real driving world, the Google car will require a computer with a level of intelligence that machines won’t have for many years, if ever.”
For example, Gomes points out that the Google driverless car can’t travel a single inch without extensive map data to be loaded into its memory, and that for practical general purpose self-driving cars to ever become a reality, some four million miles of public roads in the U.S. alone would have to be extensively and accurately mapped—a tremendous undertaking in itself. And even at that, there would be the countless driveways, off-road trails, and everywhere else you’d ever want to take the car that would still be left to do, to say nothing of automotive thoroughfares throughout the rest of the world. So far, only a spit in that massive ocean has been achieved with the mapping of few thousand miles of major roads, mostly the highway environment surrounding Google’s Mountain View, California, headquarters.
Who knows? Maybe someday real world show-stoppers such as dealing with changing weather conditions and comprehensive mapping will be overcome, but Gomes notes that even Google, when he was researching his MIT Technology Review piece, conceded to him that the sort of mapping necessary for the Google car’s consumer deployment “are an order of magnitude more complicated” than anything they’ve done to date, and that “the process it currently uses to make the maps are too inefficient to work in the country as a whole,” let alone the rest of the world.
Thus, anyone who thinks these problems will be solved inside of a decade or two at the earliest is woefully, almost touchingly over-optimistic. Which makes old school automobile enthusiasts like me optimistic that robotic vehicles will remain analogous to, say, colonization of outer space for the foreseeable future, and that cars requiring human control will be around for a long time yet—probably long enough to see me out.