Thank you all for agreeing to participate in this forum.
Let's jump right in, and then I will recede into the background. The conversation should be open ended, and the questions a springboard for the discussion.
At this year's CES, driverless car technology was not just one of the hottest trends, it was THE hot trend. So let me kick off this conversation by asking the following: what did we learn about the future of driverless cars at CES 2016, and what in your opinions is still unknown?
There was quite a bit of autonomous driving tech and a number of such vehicles shown off at CES 2016, but there's still not much on the market for ordinary car buyers.
The most full-featured and ready-to-go bit of autonomous car tech on display was in the newest iteration of the Mercedes-Benz E-Class. It's far from affordable for most car buyers, but it showed that self-driving is ready right now in recognisable cars. The entry-level price is about US$50k, but that likely will not include the self-driving tech when it launches soon.
Very importantly, the new Merc has been shown running in heavy rain (by Robert Scoble on an open road; see this video on Facebook).
As recently as late 2014, autonomous cars were seen losing control or getting lost in the rain, so that kind of major barrier has already come down and is ready for the market.
If you want all that in a more affordable package, we'll have to wait for what Ford teased at CES, where the automaker said it'll have fully autonomous cars on the road later this year in testing.
Along with Tesla's recent software updates for limited self-driving and automated parking, it seems the near future of self-driving cars lies in very familiar vehicles - not in something that looks like Google's little koala car.
There's a lot still left unknown - like how much extra such tech will cost. And if this is a pricey optional extra, will people even opt to buy it? Although gas prices are now so low, electric or hybrid cars offer some incentives to buyers - save money! - but autonomous tech doesn't have such practical appeals. What do people really get out of it for all that expense?
At Mobileye we are already working with 13 car manufacturers for their autonomous driving programs, including GM, VW, Tesla, Volvo and others. And it is clear from the industry that the primary reason that they are accelerating their autonomous programs is to address the need for a safer driving environment.
Safety is the main reason that this will become the norm, not just a fancy feature for premium cars. Vehicles with multiple sensors that have awareness of their environment better than human drivers, and without being distracted, will undoubtedly lead to a significant reduction in car accidents.
And to Steven's point, using Mobileye's solution of surround-view cameras, combined with secondary sensors such as radar or Lidar, the total cost to the car companies is only around $1,000 total - so this is going to be very affordable. Affordability is important, because it can then be more easily regulated to become a standard across all new vehicles.
As Yonah just said, affordability is key to adoption of self-driving in cars.
In my view, autonomous driving is a non-essential feature right now in a car (although it has safety benefits), so it has to be priced low. Ideally, the cost should be sunk into the price tag of the car.
Yes, for this tech to really take off, car-makers will need to make it standard. And standard on even the cheapest model. Until then, it's a toy.
But even then, we're looking at a long time for autonomous driving tech to become commonplace in cars. Remember that the average age of a car in the US is now 11.5 years old. That's a new record. Consumers are still hurting from the economic depression that started in 2008. So that will slow down mass uptake further (same for electric cars) unless the economy improves and people get back into freer spending.
The one big question that is the elephant in the room is around regulation. Certain states like California have taken a crack at it, but this is something that will take considerable time to reach consensus from state level policy makers. Ideally, there would be a national standard set in place within the next 12 months.
In countries like India (where I'm based), I see the government at least 2-3 years away from even raising questions around autonomous driving. In India, we're grappling with totally different social questions since autonomous driving would put millions of Indians out of work. Certainly something that will take some time for policy makers to get comfortable with in India.
We've touched on the issue of safety and how driverless cars will have an overwhelming net positive benefit to society by drastically reducing automobile accidents, but are there any unintended consequences? Are we sure that people want to give up their autonomy on the roads?
Governments need to do a lot to get autonomous cars rolling. First they need to encourage innovation in their countries so that their own automakers don't fall behind. And then they must do something more radical like creating a framework where insurance premiums will drop drastically for people who use self-driving tech most of the time they're on the road. That's also an incentive to consumers - something that's sorely lacking right now.
Greg touched on driverless cars taking away jobs, but I think we're a long time away from that. At least a decade. For now, a human must be in charge of self-driving cars. And those in charge of other types of vehicles, like motorbike taxis in Indonesia or "auto" (motorised rickshaws) in India, are in a safer position because their rides are much less likely to become automated in the decades to come.
It will likely be more developed countries where the tech is being pioneered where the law will change first. That's where we'll see the next stage of automated driving (e.g.: no human needed to be ready to take over; perhaps the steering inputs and pedals will be removed) happen a lot earlier than in most countries.
I think the next decade will see a slow and rather messy transition between human drivers and robocars. In this time, human drivers will not be outlawed in any country because mass adoption is too far away even in the US or western Europe.
How do Yonah and Greg see it panning out from now to 2026?
I certainly agree with Yonah's assessment here. This will no doubt be a gradual process. My primary earlier point was that when it comes to technology that has a potential to create major long term structural changes in terms of employment, India tends to be very slow to adopt. This is evidenced in the deployment of robots across industry. In this field, we actually have way lower penetration than exists in the US or Europe. I'm hoping that autonomous cars represent a slightly different reality because of the huge quantifiable benefits (reduced accidents, lower societal costs, etc.), but this is far from guaranteed.
Even the luxury cars in India don't have the basic automatic parking functionality yet so we still have a ways to go.
There will be a long period where humans and machines will share the road, but it should not be "messy" at all. In fact, there will be far less mess than we have today, due to improved safety on the roads.
Many new vehicles already come with accident prevention systems, and now that auto-emergency-braking is being regulated globally by star-safety ratings, this level of protection will soon be a standard feature for almost all new vehicles. So even cars being driven 100% by humans will include machine-based safety - a must, since 93% of accidents occur because of human error (distractions, mostly). And over time, drivers will appreciate the benefits of having their vehicles manage themselves, so that they can manage other things.
As for modes of transportation in general, it is highly unlikely that the basic concept of a small vehicle will change. The dialogue seems to be focused less on personal drones and more on the fact that there will always be cars, just far less quantities than today. Autonomous driving lends itself to an atmosphere of "shared mobility." Consumers today only use their cars about 4% of the time, making it one of the least efficient consumer-owned products in terms of its total capabilities. It will be smarter, less expensive and safer to share vehicles on an as-needed, pay-per-use basis, like Uber without the driver.
Both the auto industry and technology leaders understand that this is going to happen during the next decade or so, and they are gearing up. So if you do own a vehicle in 15 years, it will likely be used by all members of the family throughout the day, perhaps taking one member to work, then returning to take others to school, or elsewhere. More likely, neighborhoods will have enough vehicles ready such that one will be able to pick you up on a moment's notice and bring you to your destination, then get put back into service for someone else. There have been excellent studies that quantify the benefits to all interested parties in such a scenario - individuals, cities and towns, the environment, available real estate (imagine how little need there will be for parking), etc.
It may sound extraordinary today, but think about the advance in so many tech-driven items in the past 2 decades. The same is coming for road mobility, with the ultimate reward being the vast reduction down from over a million fatalities and close to 50 million injuries every year. This is why we at Mobileye and others in our industry are so passionate about the greater impact of the success of our mission statement - to end car accidents, period.
From a time perspective, this transition to autonomous vehicles will certainly take well over a decade in India. One fascinating element that people are already talking about in India is the ability to target the passenger with targeted ads as they travel past certain places/regions. This hyper-local, uber personalized advertising is no doubt the future and autonomous vehicles will make it much easier to execute on this vision.
Although the evidence suggests we're a long way off from autonomous cars that won't need a human back-up, once we are at that stage it'll unlock very exciting possibilities that blend public and private transport and which change the nature of vehicle ownership.
Subscription-based car services will be a real possibility once we have fully autonomous cars/pods. Imagine being able to choose, to pick an example from the air, a "BMW Basic" subscription for US$3,000 per year that will grant you access to BMW's entry-level self-driving pod models. You can pay more for higher tiers to get fancier or faster BMWs. So the notion of ownership will change, although it may never die out.
That's the next paradigm shift on the horizon.
I don't think we'll see any other massive changes in transport until after that car revolution. Things like Hyperloop are decades away.
Once Hyperloop is ready, we'll probably see it first in Asia. The US doesn't even have bullet trains (the average speed between Boston and Washington, D.C is a comical 109 kph), which Japan put into service in the 1960s. So I think the car-loving US will be very happy with the driverless car revolution and won't bother implementing the next public transport evolution.
The irony is in India, we may end up seeing the Hyperloop before fully autonomous cars! Will certainly be fun to see how it all unfolds over next few years.......