The Autonomous Car

Cars 2.0, a brave new world
Over the last century the highly competitive, multi-trillion dollar global auto industry has been defined by ever-evolving styling, performance, comfort, innovation, safety standards and utility, but is now entering a brave new world of ultra-high-tech. In only the next few years every new car sold in the US will use some combination of radar, lidar, GPS, odometry and computer vision in a rapidly evolving world of autonomous driving, ranging from “0” or none to “5” with fully autonomous self-driving.

As of today some high-end new cars are level 2 with partial automation such as adaptive cruise control, radar auto-braking and lane-centering. In only three years, by 2020, many will be Level 4, which means highly automated; if something goes wrong and a human fails to intervene, the car can, in theory, save itself and its passengers. Level 5 will require no driver supervision, and therefore, in theory, will not need pedals or a steering wheel.

The advantages are many. Five million plus traffic accidents in the US in 2016 killed 40,200, with death, disability, medical care and property loss approaching $300b! Lost productivity because of congestion added another $100b, totaling about 2.6% of US GDP. Fully autonomous cars would, in theory, eliminate 90% of all auto accidents, save 35,000 plus lives, substantially reduce congestion and more. Digital chauffeurs would communicate and cooperate among themselves allowing higher speeds, closer following distances, effortless merging without hesitation, and far fewer delays from stop-and-go traffic. Stress-free former-drivers could answer e-mails, return calls, or simply relax in their mobile offices. Insurance costs would plummet and car-sharing services would reduce the number of cars. Give people a button that says “Home”, and they will push it.

The argument for Improved productivity
Some advocate that because of limited population growth and consumption, the U.S. has, in many ways, topped out economically in a race-to-the-bottom global economy. They opine that Americans like to work and in an autonomous world roughly 100 million Americans would convert commuting time into productive time. The average commute time in America is 50 minutes, both ways, giving a 10% plus boost in productivity in an 8–10 hour day. At the least former-drivers could hit the “work” button, put on their favorite tune, recline their seat and enjoy the drive, arriving refreshed and ready to work.

America’s urban landscape is built around people driving to and from work, and consumption, limited in large part by distance, congestion and parking. In a perfect world of autonomous cars you will be dropped off at your desired location while your car finds a space and parks itself, or drives itself home for your wife to use, or take the kids to school, or goes into ridesharing mode to help pay for itself. Parking spaces are now about 31% of urban land and could largely be turned into living spaces, significant because urban populations are expected to grow by 66%, or 2.5 billion people globally, by 2050. Some studies believe that thanks to ride-sharing, car ownership will be cut in half in just the next decade.

A generational shift
Obtaining a driver’s license was once a ritual of American youth, marked by going to the DMV on one’s 16th birthday! Many millennials do not have a driver’s license as iPhones and social media have given them access to their world that would once have been associated with cars. Whereas dodgy steering, stick shifts, and drum brakes once required all of one’s attention, today would-be drivers have better things to do. As per the NHTSA, eating, grooming, talking on a cellphone or texting cause 2.6 million accidents a year. Most major cities are gentrifying with old warehouses made into lofts, or, in my area, entire blocks of older one-story warehouses becoming three Story condos with two car garages, one parking slot for “his” car and another slot for “her” car, with no space for anything as expensive, frivolous or time intensive as a fifty-year-old collector car. Many millennials do not have a driver’s license; don’t like to drive; certainly don’t know how to drive a stick; and those few who do drive, drive new cars leased at low rates with minimal down payments, and have zero space, and zero desire, for a collector car.

To steer or not to steer, Google Vs GM
Having gone through many Darwinian shakeouts, the auto industry is one of the most mature, yet evolutionary, markets on earth. The prevailing philosophy amongst the major manufacturers has evolved from “Driver in charge” in 2010, to “Driver mostly in charge” today, to “Car mostly in charge” in 2020, and will finally evolve into “Car in charge” in 2025.

Google wants to skip the half-measures and go for a moonshot: totally autonomous cars that, regulators willing, won’t have a steering wheel or gas pedal. Thanks to a $670b market cap and a 20% plus profit margin, Google has the luxury of taking a clean-sheet approach. In Google’s world, you won’t just quit driving cars, you’ll also quit owning them. Forget about investing in an expensive and depreciating asset that sits idle 95% of every day. Fleets of autonomous vehicles will pick you up when summoned via smartphone, or smartwatch, or whatever—drop you off, and move on to the next fare. Needless to say, Google’s model is an existential threat to the auto manufacturer’s business model, from production to marketing and distribution to sales! But what happens when there’s half as many cars and the majority of Americans need a ride to work at 9:00 and come home at 5:00? What price a ride?

Yes, John Connor, it’s Skynet
Global auto research and development is about $100b annually. Thanks to Moore’s law, artificial intelligence, deep learning technology, and the ever-falling cost of computer and sensory technology, an autonomous world would have a central traffic authority, in constant cloud-contact with every car on the road, making them crash-proof. Speeds would rise, congestion would decrease, the cost of auto-ownership would drop. What could go wrong?

As per Brian Krzanich, the CEO of Intel, data is the new oil, “If you have rich data, your car will be able to deal with complex route situations, If not, the car will stop.” The more personalized your car, or the more amenities it offers, the more personal information it must incorporate into its services. The question becomes, how much data is for operational purpose and how much is commercial?

Imagine going to your car in the morning, which says good morning and asks if you want to stop for coffee on the way to work, although your car already knows your morning routine. On the way to work your car recommends several restaurants for lunch and asks about reservations for Wednesday’s date night, as your car knows your schedule and patterns. You agree to one of your favorite restaurants for lunch and for dinner, although you don’t remember having discussed this with your car’s operating system. The question becomes, is your car recommending the sponsored one, just like Google ubiquitous ads? Regardless, your car already knows your preferences because it searches for and analyzes keywords in your emails, messages and in-car conversations. Big brother is watching.

Your car as a Supercomputer
Since 2015 all cars have had “black boxes” which track and record basic information such as your speed; gas pedal movement; if the brakes were pressed; seat belt usage; the time it took for airbags to deploy and much more. In a world of autonomous cars multiple cameras, sensors and radars will create about 4,000 Gigabytes (GB) per hour of driving. Counting viewing videos, conversations and any other internet use, the average American produces about 650 MB daily, while each autonomous car in ride share service will generate as much data as about 3,000 people, a goldmine for companies, a bottomless well of data that nobody controls. In 2014 the major automakers voluntarily adopted “Practical Principles on Fair Information” which includes commitments on transparency, consumer choice, minimization of data collection and biometric data. The problem with these voluntary principles is just that: they are voluntary. It’s the HAL factor: “I’m sorry Dave, I can’t do that”!

Death to the Luddites
The coastal elitist high-tech crowd assume that as Silicon Valley goes, so goes the US. Autonomous cars are inevitable and spell the end to human-driven cars, but only if everyone is using a driverless car. As autonomous cars gain market should they be deferential to their inferior human counterparts or should human operated cars be the ones to give way? Will human-driven cars be considered a public menace and so be restricted by higher insurance and registration costs and restrictions? What’s in play is the emergence of a caste-based traffic system, one of high-tech-haves and have-nots, of flawless autonomous cars versus mere humans, many of whom will not give up easily, clinging to their steering wheels. There’s no lack of techie-columnists who opine that twenty-five years from now driving one’s own car will be anachronistic, reserved for a few fly-over state holdouts!

3,000,000 job losses?
While statistics vary, Uber employs about 150,000 active drivers in the US while Lyft has about 75,000 active drivers. Another 175,000 or so are taxi drivers. There are about 150,000 city bus drivers, 500,000 school bus and company bus drivers, 750,000 light truck (UPS, FedEx) drivers and over 1,500,000 tractor-trailer drivers. In addition, 950,000 are employed in auto manufacturing, 2,000,000 in the parts network and 1,300,000 in the dealer network, not to mention the massive secondary or used car market. If you believe that autonomous cars will rule the road and that ride-sharing will cut new car sales in half, then many of these people will be both unemployed and unemployable, unable to make the leap to coder or programmer and unable to survive on a $15 wage flipping burgers. The potential for economic chaos is massive.

Volvo’s 50,000-person survey
Volvo’s business plan is to be a market leader in both electric and autonomous technology and so sent out a survey to 50,000 Volvo owners, admittedly a tree-hugging latte-sipping crowd. 90% of New Yorkers and 86% of Californians feel that autonomous cars could make life easier. Conversely only 60% of Texans believe that autonomous vehicles make driving safer, 9% below the national average of 69%. More telling, 72% believe that old-fashioned driving is a luxury that needs to be preserved, 55% of respondents say they want a steering wheel in a fully autonomous car and 69% feel that making decisions about their journey and which route they take is important. Interestingly, 72% agree that a car company, not a tech company, will make autonomous driving a reality.

An admittedly more middle-American survey by the University of Michigan’s advanced transportation technology group found that 62% of Americans believe that self-driving vehicles are not likely to result in an improvement in productivity: “Fully 23% indicated they would not ride in such vehicles, and 36% would be so apprehensive that they would only watch the road”.

The market realities
The average age of vehicles in the U.S. has climbed to a record 11.5 years, and of the 260m cars on US roads, about 120m are 6 to 14 years old. None have any autonomous features. Autonomous technology is expensive and so reserved for higher-end new cars and well-off early-adopters for the next 5-10 years. The cold reality is that well over 50% of Americans do not have enough savings to pay for a $500 emergency and would need to run up their credit card or borrow from a friend or family to pay for a $500 car repair, let alone buy the latest-greatest new autonomous car. Car market realties dictate that the autonomous car world is further down the road than many predict.

What could possibly go wrong?
In only the last month Equifax, the SEC, Deloitte and more admitted they had been hacked. The mere possibility of a major hack or cyber-terrorism against a fleet of fast moving, closely spaced autonomous cars has horrendous implications. Additionally both the government and automotive industry will have back doors that allow “authorities” to know and, if deemed necessary, control your every movement. Consider the thought of cyber-abduction with you locked in your car at the mercy of your robo-car and its creator or a group of terrorists. How about distracted robo-chauffeurs busy downloading updates or running a system scan, or, as AI becomes prevalent, busy messaging other robo-chauffeurs. Since each manufacturer is evolving their own software and systems, envision a hack that tells every Mercedes to make a hard turn left now? Totally pilotless commercial aircraft are technically possible today, but how would you feel about getting on a pilotless 777 for a flight to Paris?

It’s only weeks since the US suffered through Harvey, Irma and Maria. Most Autonomous cars will be EVs, ZEVs or Hybrids in which the floor is composed of high-powered liquid batteries. Now imagine that car in three feet of flood water? And how do you charge or drive your ZEV autonomous car when the electricity is off or the cell towers and internet are down for a week or more?

The ethics of Autonomous cars
What of the many ethical dilemmas along the way? When accidents do happen, does your autonomous car protect the “driver” at all costs, or alternatively should it do the least amount of damage possible to other people or vehicles that may be in its way—and perhaps sacrificing the “driver”? What programmer decides if your car is Nietzschean and drives right over everything—why not? Or perhaps an Albert Camus option, which would pause in the middle of the highway—and then suddenly shoot off because the existential leap must be made? Perhaps an Ayn Rand free-market program that would scan the bank account of each pedestrian, calculating their net worth, swerving to miss the makers, and mowing down a taker or two—who needs ’em? Or perhaps a Woody Allen program which would start apologizing as it presses on the gas and continues all the way home?

The argument for Collectors cars
Driving is one of the more intensely social interactions in our lives in that it requires us to trust and predict the behavior of dozens of other people as we cruise down the freeway at 70 mph plus. Experienced drivers are constantly aware of unstated nuance around them at all times (understanding that the driver in the car over there is likely to want to change lanes well before (or without) turn signals). Perhaps they are lost or out-of-towners? And yes, we all do it. Giving total control of my life, and the life of everyone around me, to a computer in the clouds, terrifies me.

Cars have always been about status as well as mobility; many of those who can afford them will want to own a trophy car. I opine the market will segment into three camps: automated vehicles that simply fulfil transportation needs, the latest-greatest luxury autonomous cars as Veblen goods, accessible only to the super-rich, and a third but sizable die-hard cadre of collectors and enthusiasts who own and drive collector cars for their visual beauty, their mechanical intricacies and their marque history. An EV or ZEV has never won the Mille Miglia or raced down the Mulsanne Straight at 200 mph and can never match the rumble of a 550 with a Tubi exhaust at idle. Even better, vintage cars can never be hacked; never need to be charged; and are not at the mercy of electrical and internet outages, ransomware, or cyber-abduction. They are, and will be, valued by those who believe that old-fashioned driving is both a pleasure and a luxury that needs to be preserved.