Disruptive innovation. Disruptive. Innovation. You know, I think that might be my favorite two-word phrase of all time. Better than “first car” or “first kiss,” better than “Buick Regal,” “I do” or even “Your place.” There’s just something about those words that sings true to everything right (or at least constructively anarchist) in my being.

Disruptive innovation.

Punk rock.

An innovation is said to be “disruptive” when it forces a rapid enough change to an entire market that those involved will either be forced to adopt an entirely new business model, or shut doors forever. It creates entirely new markets and value systems by designing for a new set of consumers, and eventually drops prices for existing consumers. Some examples might include the Model T for automobiles, Wikipedia for traditional encyclopedias, LEDs for light bulbs, computer printing to movable type, and of course (here’s looking at you, kid), the internet. Disruptive innovation is like the asteroid that wiped out the dinosaurs; it’s London burning and plague rats rolled into one – particularly if you happen to get caught on the wrong side of it.

Here in this century – nevermind electric cars or clean energy – self-driving cars will almost certainly prove to be the most disruptive innovation since the steam engine. Or at least the Model T. Not just to the automobile industry, either; to every industry with which automobiles are even tangentially connected, including mining, manufacturing and of course, energy. That might sound like hyperbole, but you can bet that the guys with the big wallets take this looming disruption very seriously indeed.

You have to wonder if the dinosaurs did the same – just before it was too late to matter.

In this two-part article, we’re first going to look at the surprisingly long timeline of the self-driving car, from the first radio control systems proposed all the way back at the 1939 World’s Fair, to the ultimate evolution of its visionary concept. We’re going to go way past the “Level 4” self-driving systems most people imagine when they hear the words “autonomous car,” and crank this disruptively technological dial all the way up to 11. You might be surprised at how close we already are to unbelievable things, how much closer we’re getting to the almost unimaginable, and how much we already owe to those who imagined those things first.

The second half of this article will be something like the second half of the “Future of Electric Cars," published last month. In the second part, we’re going to go into the social (and more importantly) economic aspects of the self-driving systems we’ll have soon. There’s a reason manufacturers and oil barons are terrified of these things; you’ll find out exactly why they fear this particular asteroid so much in the second half.

Then again, it’s not as though they didn’t have plenty of warning. Unlike the dinosaurs, auto manufacturers have seen this fireball coming for almost a century. And that’s where we start – with the first glimpse of one disruptive innovation, and a techno-punk vision that will shape our future.

Continue reading for the full story.

History and Design – How to Build Half a Road

1939, World’s Fair opening in Flushing, New York: Times were tough in 1939 for a nation fresh off the Great Depression and facing the prospect of yet another world war. Most people had given up on the idea of personal car ownership, and saw little reason to own one in a nation where not even half the roads were paved. With everything going on in the world at the time, people were ready for a bit of optimism; and they definitely got it with General Motors’ World’s Fair exhibition.

“Futurama” wasn’t just the namesake of a show about a purple-haired cyclops, it was a vision of the world 20 years hence. GM’s display was something like a massive train set built into a scene from The Jetsons – which Futurama directly inspired. Appreciate that irony for a moment.

...a scene from The Jetsons – which Futurama directly inspired. Appreciate that irony for a moment.

However, instead of steam trains, General Motors’ display featured lines evenly spaced of cars moving at incredible speed, almost bumper-to-bumper, on a massive superhighway system inspired by Germany’s Autobahn. What made the Futurama scene so amazing was the fact that it seemed plausible. The whole thing was designed by Normal Bel Geddes, who’d done a huge amount of research on solving traffic problems – decades before any such problem had ever manifested itself.

One of the features GM listed was a method for keeping those huge lines of high-speed, closely spaced automobiles moving without crashes or delay. The cars, GM said, would carry radio speed controls to maintain a perfectly even three to five feet of separation between them, even at triple-digit velocities. With such a system in place, GM promised that large highways could easily hold all the traffic any city could generate. Only problem was, those highways didn’t exist.

The Futurama exhibit hit American culture like a bombshell, and for quite a while afterward everything was “-arama.” People began demanding the high-speed roadways that GM promised would usher in a new future for transportation, built around cars that essentially drove themselves. It all seemed fantastically plausible, and America wanted those Futurama highways.

1945: During the War, radio control got a test run – one that went tragically wrong. Witnessing the success of Kamikaze attacks in the Pacific, with droves of obsolete bombers at their disposal and fighter-filled skies over Germany, Allied Air Command embarked on a top-secret experiment. The idea was to equip an old bomber with radio controls an tons of explosives. The flying bomb would take off with a skeleton crew, who would bail out afterward and turn control of the aircraft over to a radio controller in a chase plane.

Planes of the era already used rudimentary autopilots, which themselves probably represented the very first step toward self-driving vehicles. Using radio controls to direct the autopilot seemed like a no-brainer, and very similar to what GM had proposed in Futurama a few years before. Unfortunately, something went wrong: The flying bomb detonated prematurely, before the flight crew could bail out.

Among those killed was one Joseph Kennedy, eldest brother of future President John F. Kennedy. The details of the experiment, and the cause of Kennedy’s death, was kept under wraps until after the war. When that information was finally released, it put a somber new perspective on the current state of technology, and the reality of trusting one’s own life and family to these new electronic and radio controls. While it isn’t and never was fair to blame those controls for the bomber explosion, this incident nonetheless put a bit of a damper on the idea of self-driving cars.


The public still wanted their Futurama roads and high-speed travel. With or without Geddes’ carefully choreographed lines of self-driving cars.

As America entered the Cold War, congress approved a massive new roadbuilding project. As is the way with politics, proponents sold it on two emotional fronts to different audiences. Those cowering in fear of attack from the Red Rooskies signed off on America’s Interstate system as a national defense measure for evacuation in the event of nuclear war. That’s still the story you’ll read in the history books. But on a much deeper level, America’s optimism and hope for the future was itself seen as a weapon against Communism. Or at least, a defining difference between America and the USSR. It was ultimately that emotional appeal that sold the Interstate project to the American people, and it was Futurama that planted the seeds of emotion. Or, seen from another direction, America wanted its futuristic superhighways, and Eisenhower pushed the military defense angle to fund them.

America wanted its futuristic superhighways, and Eisenhower pushed the military defense angle to fund them.

It’s entirely fair to say that without Futurama, there’d almost certainly be no Interstate system today. There probably wouldn’t even be any four-lane divided highways. Those roads were part and parcel to the Geddes design.

America ultimately got its futuristic superhighway system, and all seemed well for a while. At least, it did until the public began to realize what happens when you only use half of a good idea. After several decades of gridlock, congestion and death on the highway, somebody smart suddenly remembered that the Futurama highways themselves were only half of the total system Geddes came up with. They were designed from their inception to work with self-driving cars, and only self-driving cars. Frankly, considering that, it’s amazing freeways aren’t a lot more murderous than they are.

All of which begs the question: What would this transportation system’s designer say if he knew that, more than 80 years after he came up with it, we’d still only built half of the thing?

Present Day – And Into the Future

Self-Driving Cars, Timeline Exterior
- image 633048

To be fair, we have come a long way toward the self-driving car. It started with two of the very first drivers’ aids, both of which (ironically) came from aircraft in WWII: Anti-lock brakes and cruise control. In fact, it was ABS that directly evolved into the electronic traction/stability control systems that directly preceded today’s functional auto-driving systems.

Ironically, it’s Mercedes – the first to develop ABS into traction control – that still leads the way on commercially available self-systems. The biggest breakthrough came back in 1999, when Mercedes first introduced radar-based adaptive cruise control to the market. Next, came optical “lane departure” warning systems. Now, Mercedes has a new system that combines optical and radar-based navigation; in typically understated German fashion, it’s known only as “steering assist.” What Mercedes should have called the system (combined with adaptive cruise control) is “The car freaking drives ITSELF!

Sort of.

And this is where we get into a discussion on the different levels of self-driving vehicles. Technically, while the Mercedes system is fully capable of controlling the car at speed, it’s only semi-autonomous, somewhere between Levels 2 and 3 on the self-driving car scale.

Level 0: The driver completely controls the vehicle at all times.
Level 1: Individual vehicle controls are automated, such as electronic stability control or automatic braking.
Level 2: At least two controls can be automated in unison, such as adaptive cruise control in combination with lane keeping.
Level 3: The driver can fully cede control of all safety-critical functions in certain conditions. The car senses when conditions require the driver to retake control and provides a "sufficiently comfortable transition time" for the driver to do so.
Level 4: The vehicle performs all safety-critical functions for the entire trip, with the driver not expected to control the vehicle at any time. As this vehicle would control all functions from start to stop, including all parking functions, it could include unoccupied cars.

Self-Driving Cars, Timeline Exterior
- image 633046

Of course, as we all know by now, Google has come a long way toward full-on Level 4 autonomous vehicles. Its self-driving cars have been tooling around California and Nevada for a while now, relying on a combination of GPS, LIDAR-based navigation and optical camera systems to get around. For a while, it seemed Google was the only player in the fully autonomous game. But Google’s fully realistic threat of perfecting the driverless technology by 2020 has forced almost every major manufacturer to begrudgingly get serious about their own.

GM, Volvo, Mercedes-Benz, Audi, Nissan, BMW, Renault and Tesla are all slated to have fully autonomous vehicles for sale by Google’s projected 2020 release date. Say what you will about the free market – but once disruptive innovation becomes a serious threat to business, things tend to move along pretty quickly. And you can bet, GM and Mercedes aren’t about to lose market share to Google, or pay the licensing fees for its surely patented software.

So, there it is, right? After almost a century of being constantly “just around the corner,” driverless cars will one way or the other become a commercial reality by 2020. It’s going to happen. You will be able to buy a 2025 Cadillac that drives itself to the golf course, with or without you. In fact, the IHS predicts that by 2025, there will be about a quarter-million fully autonomous vehicles on the road. By 2035, they expect more than 54 million, and by 2050 it’s expected that every car in the world will have some autonomous function. So, end of story, and pop the champagne. Right?

You will be able to buy a 2025 Cadillac that drives itself to the golf course, with or without you.

Not exactly.

As of right now, four states have legalized autonomous cars, and so has the U.K. and France – in certain areas. And that’s exactly the problem, because most other areas contain one glaring monkey wrench waiting to fall into the works: People.

There are certain to be a lot of iterative steps along the way, as technologies improve and people begin trusting robots with their lives. I know I’m not entirely prepared to as of right now. Though I’d have to imagine a blind, deaf, emotionless pile of silicon would almost have to be a better driver than my last girlfriend; someone for whom a diesel Ford Expedition should almost certainly be declared a weapon of mass destruction. So, perhaps it’s all a matter of perspective.

Here’s my personal perspective on the future of autonomous cars, and a rough timeline of what we can expect as this disruptive innovation takes us from today to the highways of Futurama.

2025 to 2035: Several manufacturers, notably Tesla, General Motors and Mercedes, release fully autonomous vehicles capable of operating without a driver. Very likely, fully auto mode will be prohibited in most populated areas, and all but the far right lane of interstates in areas where they are allowed to roam without a driver.

These early vehicles will network with GPS systems, and will not permit owners to program in a destination where autonomous vehicles aren’t allowed. Say, for instance, you want to program your car to pick up the kids from soccer practice, but if the soccer field is in a no-auto zone; the car’s computer will simply tell you “no,” and you’ll have to drive it there yourself.

Likely, there will initially be a three-tier system: “Driverless Zones” where the car can go on free roam by itself, “Driver Assist Zones” where drivers can engage the autonomous function to take over the wheel, and “Manual Control Only Zones.” You’ll likely see Driverless Zones in rural and outlying areas, or on seldom-used roads. Most places in and around cities will be Driver Assist; pedestrian-heavy areas and school zones will probably be Manual Control Only.

As of right now, for liability reasons, drivers still have to maintain some degree of control of the cars they’re driving. Or riding in. Full-auto mode will be reserved for passenger-less vehicles in certain areas. If there are occupants in the vehicle, the “driver” will still need to manually tell the car when to change lanes, make turns or merge into traffic. If this were a symphony, a driver wouldn’t be so much “playing an instrument” as “conducting an orchestra.” You might tell the car to change lanes by using the turn signal, tell it to pass the cars ahead by pressing a button that says “pass,” or just verifying turns suggested by the nav system.

If this were a symphony, a driver wouldn’t be so much “playing an instrument” as “conducting an orchestra.”

This isn’t for the most part technically required; at this point, the car could drive itself on full auto with or without someone telling it what to do. But, if it hits someone, they’re going to want to know who to sue. And computers don’t pay insurance claims. By keeping a driver in some measure of control, it puts the responsibility for what the car does on the driver.

2035 to 2040: By this time, the government will have mandated a common communications protocol, in much the same way that they did with OBD-II computer programming language back in 1996. Call it “Car to Car Communications, Series 2” protocol. All cars built after a certain model year will be required to carry CCC-II, and manufacturers will offer software updates for cars built before CCC-II.

With CCC-II in place, cars can start talking to each other on a regular basis. Not all cars will use the protocol, but enough will that they can start using “swarm” programming to work together. Swarm programming is exactly what it sounds like: A type of “artificial intelligence” based on the swarming behaviors of insects, a “hive mind” that allows individual units to self-organize, work together to accomplish tasks, and stay out of each others’ way. We see the same programming used today for experimental drones and robots; even now it’s very nearly good enough to control a dozen or more full-sized automobiles. By 2035, swarming will be down to a science.

What’s really interesting about this technology is that, for once, it’s actually far beyond anything Geddes could have hoped for in terms of making his future highway system work. As he originally envisioned it, cars would simply form “virtual trains,” simple “follow-the-leader” groups called “platoons.” Forming vehicles into simple trains allows them to run almost bumper-to-bumper at freeway speeds, which saves fuel and actually enhances total group safety.

Self-Driving Cars, Timeline Exterior
- image 633052

Remember back in the caveman days of 2013? Way back then, the SARTRE (Safe Road Trains for the Environment) Project in Sweden used a human driver in a tractor trailer to control three fully robotic cars following behind. The “road train” ran almost bumper-to-bumper at freeway speeds in Europe. Using a very simple “follow-the-leader” strategy, the Sartre Project showed that simple platooning is a reality even with today’s technology. There’s no real reason why Sartre couldn’t have lined up a hundred cars behind the human-controlled truck. In that alone, we’ve already gotten pretty close to Geddes’ vision.

Realistically, Mercedes itself is only a CCC-II type protocol away from making that a reality right now. Today. But, by 2030 or 2035, autonomous cars will be doing a lot more than just playing follow-the-leader.

By 2030 or 2035, autonomous cars will be doing a lot more than just playing follow-the-leader.

True, that will be the majority of what they do; platooning as imagined by Geddes and demonstrated by Sartre really is the best way of moving a group of vehicles down the road. But the swarming programs we have today – nevermind those we’ll have 15 or 20 years from now – will allow huge numbers of cars to form super-platoons, and react both independently and as a group.

Say, for instance, a tree falls down and blocks the road ahead. The individual cars under the falling tree would use their laser sensors to see it coming. They’d communicate that to the cars behind them. Those who couldn’t stop in time, but could make it through, would be ordered to speed up, and the swarm ahead of them would speed up to close distance with each other and make some room.

But what of the cars behind? Do they all just slam on their brakes? No – that’s what a platoon would do, and it would result in a massive pile-up. Just as the swarm ahead of the tree is ordered to speed up and close distance, the swarm behind would slow down. But it doesn’t stop there, because the cars behind need more room between each other to stop safely.

This swarm would react by directing every other car to simultaneously and preemptively swerve into the breakdown lane. That then opens up another one or two car lengths for each one to slow down. Then, each individual car, tracking the position of every other car, would provide just enough brake force to avoid hitting the vehicle ahead of it. Eventually, all traffic would come to a stop, as cars a half-mile back from the fallen tree began gradually decelerating in anticipation of the stopped traffic ahead.

Would some unlucky soul hit the tree? Quite possibly. But this system would prevent a deadly pile-up, even if that tree dropped into the middle of a busy interstate. The fact that every car a mile back became instantly aware of the falling tree even as it fell gives them all the time in the world to slow down before crashing.

You could extend this same scenario to a manually driven car suddenly swerving in and out of traffic, or a tractor-trailer crashing through a guardrail, or a government surveillance drone crashing in the middle of I-90. Because, obviously – you know they’ll always be one around somewhere. That’s a different article, though.

Point is, by 2040, these super-platoons – or more accurately, super-swarms – will have become among the safest things on the road.

2050 to 2060: Up till this point, manual control was thought to be the safe, fallback method of control. Now, with all the road data in and swarm systems proving their safety and reliability, manual control will only be permitted outside of major population centers.

It’s kind of a weird inversion of how the “zoning” system started out; but eventually, swarms will come to regard those weird, erratic things known as “humans” as illogical wild cards intent on messing up the swarm’s day. In major population centers, cars controlled by humans will come to be seen as bulls in the china shop, or boulders in the river. Assuming human control isn’t outright outlawed in major population centers and on busy roads, human-controlled cars will probably end up squeezed into the far left or right lane.

In major population centers, cars controlled by humans will come to be seen as bulls in the china shop.

Of course, you’ll still be “allowed” to drive your own car where you’re not in danger of disrupting the swarm; but the onboard GPS system will automatically put the car into auto mode when you enter certain areas.

Think of that scene in the movie Independence Day, where Will Smith pilots the crashed alien fighter up to the mothership. He’s got full control of it all the way up until the ship’s airlock opens and he goes inside. At that point, the mothership’s computers take over control of the ship, and guide him to the dock on autopilot. That’s how entering most major cities will be; though it’s probably safe to assume cities will retain a network of "Level Zero" roads for use by old fogies driving antiques from 2015. Maybe around retirement villages – also known as “Soylent Green Factories.”

Wow…that got dystopian quickly.


For the most part, by about 2060, manual control of cars anywhere near civilization will come to be seen kind of the way texting and driving is seen today: dangerous, stupid and sociopathic. Especially considering the fact that by 2060, we’ll have truly entered Futurama, and cars will regularly run bumper-to-bumper at triple-digit velocities. Every highway in the civilized world will turn into one, long NASCAR track. Which is awesome, and very plausible considering that running bumper-to-bumper means you’re constantly drafting the car ahead of you. You’re always in a tunnel of near aero-vacuum; so it wouldn’t cost much more in terms of energy to run 200 mph in a tight swarm than it would 55 mph on your own.

Every highway in the civilized world will turn into one, long NASCAR track.

Might we see 200 mph super-swarms? Could be, on long stretches in the middle of nowhere. But who decides how fast you go?

Most likely, it will be something like the Independence Day scenario above. The super-swarm will likely communicate with a central traffic control computer, the “city mothership,” if you will. It’s actually pretty unlikely we’ll even see set “speed limits” by 2060; in most places, the speed limit will be whatever central control says is ideal for the flow of traffic. Your car will probably also select from routes suggested by the central control board, which themselves will be optimized for minimum travel time and maximum energy efficiency.

And now, here we are – finally — living the future.


I’d imagine that, if you’re reading this on Topspeed, something probably rankles pretty foul about this scenario. I know. It is a little Orwellian. You’re waiting to hear the word “CentCom” somewhere. I bellyfeel your badfeel to groupthink, and truethink your fun is un-thoughtcrime.

Something about this is a little Orwellian.

But, rest assured, it’s extremely unlikely that even the most Big Brother of control systems will cover the entire country. The most we’re ever likely to see is mandatory auto-drive on interstates and in large population centers. Odds are good you’ll still be able to bash your pickup truck around the trails and over the backroads if you like, and helm the controls yourself most other places. Especially in America, with its vast tracts of land between major population centers. I won’t say it would be impossible to radio-network every road in the country; it’s not impossible to do a lot of things, including take a shower with your cat. It’s just pointless, and really,really stupid.

But, don’t take that to mean that self-driving cars won’t affect every aspect of your life, no matter where you live. They will, and probably a lot sooner than you think.

This article was all about the innovation, but as you know, “innovation” is only half of my new favorite phrase. In the next article, we’re going to get to the fun part – disruption. Disruption of the entire social order; not just the social order of our nation, and not just the basis of our economy, but that of the entire world. We truly are looking down the barrel of an epic shift in everything we ever knew. Without being melodramatic, the next 15 years will probably end up being not just a turning point in the history of the automobile, but a turning point in human history.

All thanks to a few self-driving cars, and the power of disruptive innovation.

It’ll be anarchy.

See you at Part II.

What do you think?
Car Finder: