Autonomous Cars 101, with Brad Templeton

Software Architect, Internet Entrepreneur, and Futurist
Self-driving cars are about more than taking your hands off the wheel. Brad Templeton, who consulted Google on its autonomous cars project, says a car that drives itself is a car that can deliver itself to you. Such an on-demand service will fundamentally change how we relate to our vehicles, says Templeton. Currently, we're given to buying cars for their least practical qualities. City-dwellers buy SUVs for a couple skiing trips each year, or a weekend carpenter will buy a gas-guzzling pickup and only use the truck bed occasionally. In an on-demand world, we'll have greater fuel efficiency and, as a result, a much better world.
  • Transcript


Brad Templeton: The first thing to understand before looking at cars that drive themselves is just how much of our lives cars have become, especially in the United States but really all around the world.  There are 33,000 Americans killed every year in car accidents.  More killed in car accidents in the United States than in its entire history of war going back to the Revolutionary War.  We have 1.2 million people killedaround the world.  It's one of the world's major diseases - if it were a disease - in terms of killing people.  And we also give 25 percent of all of our energy to personal transportation and 25 percent of our greenhouse gases are going to the car.  Now this is not true in Manhattan, but in Los Angeles it's estimated that more then half of the land in the city belongs to the cars, its garages, its the driveways, its roads, its parking lots, all these things land that belongs to the car.  We have given up so much of ourselves and we depend on the car for so much it's shaped our cities, it's shaped our lives.  The fact that now cars are going to be a computer thing.  That the computer is going to be the most important part of the car, the thing that drives the car will be more important than the engine.  And this will be on the Moore's Law curve we've talked about getting better, not the engine getting better every year but the important part that drives it getting better every year the way computer technology and network technology do.  We're going to rewrite really important elements of our society when we make transportation one of these computerized technologies. 

Most people thought that cars that drive themselves were something from the science fiction.  I still routinely run into people who say "This is not in my lifetime, 20 or 30 years away."  But if this is clearly wrong and it is becoming more and more clearly long as time goes on.  The most famous project and the most advanced project has come not from a car company but from Google.  And I worked actually on Google's team for a while in building that vehicle.  So they have both logged about 700,000 miles driving on ordinary city streets with their vehicles.  And they've now just released in May of 2014 they have released a new vehicle that's they're building from scratch, designing from scratch which has no steering wheel in it, no pedals, you just get in and you have given it a destination probably on your phone and away it goes and takes you there.  And that's the real game changer when it comes out. 

The car companies also are all working on efforts.  Every major car company has some sort of effort and Nissan and Mercedes and a Volvo have all announced they'll be selling cars in about 2020 that look a little bit more like traditional cars.  But what Google's car does it's a vehicle that can run without a steering wheel and thus it can run unmanned and that's where it gets really interesting.  Because a vehicle that can run on it's own is a vehicle that can deliver itself to you.  It's a vehicle that can refuel or recharge itself without you having to worry about it.  And it's a vehicle that can store or what we used to call park itself, although it may mostly function as a taxi.  And as a taxi it wouldn't even park it all it would just go and pick up the next person it has to pick up.

So today when people buy a car they go into the car dealership and they look for a multipurpose car.  They ask: "What car do I need for my life?"  Because they think "Well I ski twice a year so I need an SUV, even though I live in the city and I need an SUV to get me up to the mountains every so often." Or the number one selling car in America, the Ford F150 pickup truck, which is not what most of those people need.  But we all go and buy at least a five-passenger sedan, a multipurpose car for all of our needs.  When he think of cars as fungible objects almost that can be delivered on demand because you pick up a cell phone and you ask your car to come to you, well suddenly you don't have to get a multipurpose car, you can get a car for that trip, the car that you need today; the car that you need right now.  That is a very different car from the cars we've made for the last century these big fat boxy things.  This is a car most of the time for one person because almost all your trips are alone.  Sorry, unless you're a very friendly person, they're alone and they're just short trips across town.  But we buy a car that's able to take five people hundreds of miles and that's very wasteful.  It's very energy inefficient and it takes a lot of space on the road too.

So if most trips come in little small electric vehicles in particular that are good enough for one person or two people face-to-face and they have a desk in them and a screen so that you can work or watch videos and do things and takes no time, it's a very different vision of transportation and it's a much lower energy vision of transportation, particularly as you enable electric vehicles.  Now, today people don't buy electric vehicles that much, they buy some; there's certainly some press being made about it, but it's really a small portion of the car market and they are worried that the car will run out of range or it will take hours to recharge it when they're done.  They don't want the hassle of plugging it in.  But a robot doesn't care about any of those hassles, and robot doesn't care about anything.  And because of that you can get a vehicle that is electric because you don't care about the range.  You don't care about the power train at all.  It's very different from the way cars are sold today.  So we have lots of small efficient vehicles.  These vehicles are so efficient that they don't just beat out the cars we're riding in now, they beat out the trains and the buses, even in Manhattan, even in Japan.  That's how efficient small lightweight electric vehicles can be at carrying people in terms of energy used to send a person a mile.

So, this would mean the United States would no longer have to import oil from overseas.  And you've probably seen there's a pesky habit that the U.S. has of going to war over the oil imports from overseas.  It would be a very nice habit to be broken of.  It also means reducing greenhouse gas emissions by 200 million tons.  It means getting rid of all the parking lots we have in our cities.  Because if we don't have to store vehicles in the same way, suddenly there's a land of bonanza in so many towns, maybe not the dense downtown part but all of the surrounding territory where so much of that land is parking lots and or cars.  Suddenly we're going to turn some parking lots into parkland, that's going to be pretty nice.  We also have tricks to deal with congestion and that's pretty big.  If you invest in real estate you probably have heard a mantra that says the value of real estate is based on three important factors right, it's an old joke, location, location and location.  We're going to change what location means because transportation is really what location is about.  It's having a short trip either on foot or on wheels or in transit or other things to the things in your life that you need to get to - shops, people, work.

The whole nature of retailing is going to change because we're also going to build delivery robots that never carry a person and they're small and they're very, very cheap.  And they can get you anything in 30 minutes, not just a pizza.  Now we do have a society that's decided the pizza is the most important thing in the world; you have to be able to get it in 30 minutes no matter when, no matter where you are, but what if you can get anything cheaply in 30 minutes?  What does that mean for shopping, for retailing, for how we use goods, how we own goods?  It's going to change a lot more of our lives then people think to have cars that are smart in this way.

A lot of people want to know when this is going to come out.  And I make jokes about naming a specific day and time.  So we've seen forecasts from many different companies.  Sergey Brin first said Google he thought would have cars in people's hands and 2017.  Widespread deployment not then but it's happening in the later part of this decade and real deployment in the early part of the 2020s.  The car companies have said 2020 is sort of their target year.  Most of them have said that.  Although Tesla, because Elon Musk, the CEO of Tesla, has the largest testicles of anyone in Silicon Valley… he said he thinks he's going to have a car in 2016, we'll see if he delivers on that.  Volvo though has also said they'll have cars running around their headquarters town in the hands of civilians in 2017.  So there's a lot of people willing to make predictions for really quick delivery.  Now, will they spread like wildfire like smart phones or will they spread like electric cars really, really slowly?  Well, I'm betting on more like smart phones, but that's not a bet that anyone can promise the answer to. 

We're not talking about vehicles that will be perfect, it would be hubris to suggest they could be perfect.  There will be accidents, there will be injuries and there could unfortunately even be fatalities.  And we do have a strong fear of being hurt by machines, and good reason.  It's something to be afraid of.  But the social question is if you don't like being killed by robots you'd rather be killed by drunks, because that's what's happening today.  Forty percent of the fatalities on the roads here have drinking involved in them.  And robots, for better or worse, very rarely drink.  They'll make mistakes but for different reasons.  And so as a society it's an interesting question, something you're very scared of but which is actually safer on the whole.  Nobody is going to release this technology until they've got to the evidence to themselves, to their lawyers, to the public that it's doing a better job than people are at driving.

So you've got something that is better than people but it's more frightening and you could indeed see societies jumping back from that.  However, whichever country decides to jump back, another country will jump in.  And there are many countries which can change their laws and their liabilities and their rules very quickly in a way that the United States maybe can't.  And so there will be global competition to see who rules the new car industry.  And I really mean the car industry.  There is a lesson that we teach over in Silicon Valley about what's called disruptive technology and when we go through disruptive change in a technology.  It's very uncommon for the big players from the earlier generation to make it through to the new world.  How many of you are using a Kodak camera, for example?  Not too many.  So I fear for the car companies because they will die, some of them, as we move to this new world.  And the places where cars are made may shift.  Detroit has a big challenge in front of it and it's going to face against wherever the new industry comes from.  But Japan, Korea, Germany, they also need to be afraid; they're the centers of car manufacturing today.  But politicians actually are aware of this and they're interested in seeing ways to make sure that their economy becomes dominant in this world.  And that's why they've already passed laws in several states and now in other places around the world to enable this technology rather than slow it down.

Directed / Produced by Jonathan Fowler and Dillon Fitton