Self-driving cars will change everything
I’m utterly fascinated with self-driving cars. They’re in the headlines every day as Tesla, Uber and Google spar over technical advances. It’s not only the tech that fascinates me though. I’m enamored with how self-driving cars are going to change the world we live in.
I’m assuming you remember life before the iPhone and the shift it’s brought with it in less than a decade. I argue that vehicles are the only other consumer device on par with smartphones both in market penetration and impact on day-to-day life. There are more passenger vehicles on the road in the US than there are smartphone users (260M vs 220M), and that number doesn’t include commercial uses. The arrival of autonomous vehicles will fundamentally redefine our relationship not only with our cars, but also our cities, and each other. And it will cause shifts in quality of life in both directions as it arrives.
The state of the art today is Tesla’s Autopilot. Yes, Google is driving around Mountain View, and Uber’s first self-driving cars hit the road in Pittsburgh any day now, but Autopilot has logged 130M real-world miles. (In fairness, the efforts shouldn’t be directly compared as they’re currently working on different implementation levels of autonomous technology: Tesla’s Autopilot is a level 2 implementation, Google is shooting to jump immediately to level 4.)
Though certainly not perfected, Tesla will continue to improve Autopilot, both through better hardware on the road and through collecting exponentially more data when the Model 3 launches. Musk predicts fully autonomous cars will hit the road by 2018. He’s known to be aggressive on timelines but not by much. It’s reasonable to say that some cars on the road will be driving themselves by 2020. Depending on how quickly the roll out goes, children born in 2004 may not ever need a drivers license. A child born today certainly will not.
This opens the door to some very interesting questions that our society will face in the next 10–20 years. For example, what happens to car ownership? If the cars drive themselves, that means Uber doesn’t need drivers to bring their own cars. Cars can be on the road nearly 24/7, dispatched by software, and run at a fraction of the cost because there’s no human driver.
The benefit of owning a car is having instant access at any time. With autonomous vehicles, surge pricing and delays caused by capacity issues become a thing of the past. This gives end users effectively the same level of access as ownership would provide, without requiring the capital commitment. Why would anyone need to own a car in a world of instant and cheap transportation?
On one hand, it’s easy to assume that incumbent car manufacturers will continue to rely on a direct sales of their cars, but GM’s recent acquisition of Cruise should be looked through the lens of their $500M investment in Lyft as well. Mercedes, BMW, and Audi all currently operate car sharing services so have some proficiency with operating in a service model. And Ford’s gone so far as to say their initial autonomous vehicles will be prohibitively expensive for consumers and will only make economic sense for a service providers. Surprisingly, it’s Tesla that’s most publicly asserted their intent to sell autonomous vehicles to consumers. Part Two of Tesla’s Master Plan includes car sharing as an assumed use case, though expect that to work best with a Tesla-branded service.
Children born in 2004 may not ever need a drivers license. A child born today certainly will not.
Of course there will be push back on giving up the idea of ownership completely. Car ownership is a deep-seated part of the American Dream and a lot of individual personality is tied to one’s choice of car. That makes it a hard consumer behavior to break. But we’re already seeing car ownership in decline today as the general population moves to cities with better transportation options. Autonomous vehicles will accelerate that trend.
There’s also an increasingly strong argument to take human drivers off the road. 38,300 people were killed on U.S. roads in 2015, 1.25M globally. 95% of those were determined to be the fault of a human operator. It’s the top cause of death among people aged 15 to 29 years.
It’s easy to see the argument for human drivers being restricted from driving on certain autonomous-only roads or dedicated lanes early in the adoption cycle. From there it will be a slippery slope of restricting human drivers on all freeways (which will then run at higher density and speed), before being relegated to closed-track days altogether.
With the carrot of lower cost and the stick of restricted road access, it’s hard to imagine a world where car ownership will be anything but a fraction of the level it is today.
In fact, a recent study predicted that if we gave up traditional car ownership for a Lyft Line/UberPool-style car sharing model, we could take 90% of the cars off the road without any reduction in transit time, while also reclaiming 20% more space inside cities currently occupied by parking and roads.
This means autonomous vehicles will create opportunities for urban planners to redesign the basic building blocks of cities over the next century. When all vehicles are self-driving, cities can reduce the footprint of roads and parking, and optimize infrastructure, particularly traffic signalling to improve traffic flow. Similarly, smart city tech companies will need to step up to enable cities to gather data and provide automation-friendly data to individual cars. The no-stop self-driving intersection is only a simulation today but it’s a possibility in the future with the right urban planning. Even our homes will change in a world where we no longer need parking for our personal cars. Imagine when having a garage is as outdated as having a carriage house is today.
Still, it’s very hard to predict how our cities themselves will react to easy access to transportation. Does that enable more density in city cores or a resurgence in bedroom communities as the pain of longer distance commuting is eased? Either way, there’s an opportunity for those in real estate to speculate on the changes automation will bring.
We could take 90% of the cars off the road without any reduction in transit time, while also reclaiming 20% more space inside cities currently occupied by parking and roads.
Of all the changes that autonomous vehicles bring to our cities though, perhaps the most fascinating is the impacts on mass transit. In a world where large shared self-driving fleets reduce traffic levels and drop the cost of getting from A to B, what does mass transit look like?
In the last decade, there’s been a revitalization of public investment in rail-based mass transit in the US. But if general availability of autonomous vehicles is five years away, it’s unclear whether decade-long rail infrastructure investments actually make sense.
For example, today’s choice of trains over expanded bus service is usually one of throughput: the amount of people that can be moved between two points in a fixed period of time. Within cities today, trains on dedicated right-of-way can’t be beat. But buses on dedicated right-of-way (called Bus Rapid Transit) have many of the benefits that rail and signalling provide trains today, and there’s no reason the dedicated busways couldn’t be underground either.
When autonomous is added to the mix, buses should be able to run at higher speed in closer proximity with larger vehicles to boot, making them very comparable to trains while still maintaining the flexibility to run in mixed traffic if needed. At the very least, this is a compelling reason for current rail plans to support mixed-mode right-of-way that allows both rail and bus at the same platform, as Seattle’s downtown stations do today.
Autonomous buses may not need to run fixed routes either. A fleet of smaller buses could operate in a Pool/Line-style mode and provide more convenience on lower traffic routes. The main blocker here isn’t so much the buses but adoption of smartphones by the general population to signal for a ride. Regardless, those smaller buses start to look a lot like minivans running in an Uber Pool. In fact, Lyft has already done a partnership with the MTC to support carpooling.
If general availability of autonomous vehicles is five years away, it’s unclear whether decade-long rail infrastructure investments actually make sense.
This begs the question of whether mass transit infrastructure should be owned by the public at all. Taxed-subsidized transportation options are almost certainly still required for public good. It’s unclear whether that’s still best provided by government-owned fleets or whether the complexity of operating an autonomous fleet and dispatch service is better provided by partnering with the private sector.
It’s worth pointing out as well that there’s a dystopian monopoly scenario here. It’s easy to assume that the cost savings get passed onto the consumer and healthy competition between options remains. On the other hand, one can imagine a world where one car sharing service wins and creates a monopoly on transportation. Though it’s an unlikely outcome, it’s a reason for local governments to invest in public options and ensure that multiple players have a chance to compete.
Still, it’s hard to understate this: all of humanity will benefit from the arrival of autonomous vehicles. All forms of transport, whether that’s driving, biking, or walking, become dramatically safer without human drivers at the wheel. And because they’ll be less likely to be involved in crashes, cars can be built lighter and more efficient, requiring less energy to operate. Combine that with a drop in the number of vehicles on the road and we’ll be creating a meaningful dent in humanity’s energy consumption while improving life expectancy at the same time.
Autonomous vehicles are also going to be particularly impactful for those with limited mobility today. Elderly and disabled will have personalized transportation options, giving them access to life-changing freedom and potentially life-saving as well. In addition, the sensing technology could scale down apply to scooters and other personal mobility devices to help owners simply navigate their homes. Similarly, it will make safe and trustworthy mobility available to parents of young children enabling similar freedom to both.
All forms of transport, whether that’s driving, biking, or walking, become dramatically safer without human drivers at the wheel.
Not everyone will benefit equally and it’s similarly hard to understate this: Any person who makes their living as a driver or supporting drivers will be impacted by the arrival of vehicles that no longer need human drivers to operate.
In the coming decades, 2.5M truck drivers, the most common job title in the US, will disappear. Human taxi drivers will disappear as well, though they may take some bitter solace in seeing all Lyft/Uber drivers going as well. Bus drivers and all mass transit drivers will follow. Delivery and garbage will remain in vehicles for a longer period of time, though primarily as passengers until we figure out how to interface with buildings automatically.
But the impact is not limited to drivers. Anyone who makes their business serving drivers (think roadside diners and motels) will be affected by automated trucks cruising by in the night. The car rental industry will disappear except potentially to provide specialized equipment like UHauls through Uber-like dispatch. And of course, the entire auto insurance industry collapses when individuals stop owning cars and vehicles are involved with a fraction of the accidents.
Any person who makes their living as a driver or supporting drivers will be impacted by the arrival of vehicles that no longer need human drivers to operate.
There will be a lot of responsibility placed on the shoulders of government to step in with social programs supporting those affected by autonomous vehicles. Basic Income is the silver bullet du jour but the transition will require a portfolio of income and tax relief as well as retraining. It will be rough but it can’t be avoided.
At the same time, it shouldn’t surprise that a change as drastic as this will create opportunities for many new types of businesses as well as modernized versions of existing ones.
- The base building blocks of autonomous vehicles are the obvious short-term opportunities, though incumbents are quickly building or acquiring teams here. There’s sensor technology: Ford and Baidu recently invested in Velodyne, which makes one of the most popular LiDAR sensors (a great example of a decade long overnight success, Velodyne’s founders build their first LiDAR to run the initial DARPA Grand Challenge in 2005). AI/image recognition is a big gap for most automation manufactures. And Google, Apple, Tesla, and Uber are all heavily investing with their own custom 3D mapping platforms while Zenrin has launched to develop 3D maps of the country to support self-driving in the country in time for the 2020 Olympics.
- Companies building dispatch software on top of autonomous infrastructure, ie the Lyft/Uber equivalents for other industries. You can see that in trucking now with Convoy, or Shyp with scheduled pick up. There’s absolutely room for whitelabel Lyft/Uber dispatch system as well for local governments that desire locally-owned and operated systems, or bespoke services with specialized or premium vehicles. Companies like TransLoc, already building smarter routing for public transit are well positioned for to ride this wave.
- Robotics companies working on vehicle-to-building interaction. Getting objects other than people into and out of vehicles will be complex and keep a lot of humans on the road as passengers for quite a while. The obvious place this will happen first is automatic refueling or charging. More complex problems will be delivering food or packages, collecting garbage, moving, etc. The solution will likely involve some compromise on both sides of the interaction as automated delivery to a BufferBox-style unit is a lot more feasible than to individual mailboxes.
- There will still be opportunities for small-scale vehicle builders in the world of autonomous vehicles. GM and Ford currently offer drivetrains as a platform for building vehicles with specialized purposes. Autonomous tech will need to apply to motorhomes, limos, moving trucks, and other specialized vehicles too. It’s an open question whether the incumbents continue to offer that platform or whether Tesla or some other upstart can provide it.
- There will be a lot of work by policy makers to support self-driving vehicles, whether that’s federal legislators or municipalities looking at urban planning and infrastructure. There’s absolutely the opportunity for those working in government today to be thoughtful about how to position their cities, states, and countries as autonomous vehicle friendly and help their constituents benefit from the potential quality-of life improvements in the coming decades. This applies equally to efforts to support those whose ability to make a living will be hurt by the move towards automation to help make that transition a little less painful.
That’s a lot of opportunity and certainly not an exhaustive list. Many of these topics deserve deeper exploration on their own and many of the questions will become clear in the months, years, and decades ahead as companies place their bets and communities react. The technology is already moving incredibly fast (just keeping up with day-to-day autonomous news triggered several rewrites and additions to this post).
Still, I hope you’ve got a deeper sense of the scale autonomous vehicles’ potential impact. Our relationships with cars will change as we abandon ownership, our relationships with our cities will change as they no longer have to accommodate idle vehicles, and our relationships with each other will change as we take advantage of new levels of freedom and the opportunities they unlock. In my mind, autonomous vehicles represent one of the most interesting and impactful new platforms, both for business and society, in our lifetime. I couldn’t be more excited to be along for the ride!
Thanks to Boris Jabes and Josh Jenkins for the editorial assist.