Shared publicly  - 
Hi all,

Here's Part 3 of my driverless car series at Forbes on why I think #driverless cars will arrive sooner than most think. I address the hurdles and offer three scenarios which might jumpstart the adoption process: Google Fiber Redux, The China Card, and a Big Venture Play. Please take a read, and I look forward to your comments. "Google's Trillion-Dollar Driverless Car -- Part 3: Sooner Than You Think"
Jos Echelpoels's profile photoChunka Mui's profile photoDavid Vorriccelli's profile photoharish upadhyay's profile photo
About trusting the computer and liability: It's amazing how much trust drivers put into the computers that control traffic lights. They gladly trusts that the light is showing red to the other guy when they see the green light.

Anyone know who is liable if the traffic controller suddenly shows green on all sides and two cars collide in the middle? 
Absolutely love this series and I sure hope it's widely read. Great analysis +Chunka Mui

Few comments:
The trust thing - The advent of driver-less cars is similar to the advent of the elevator. People always had to climb the stairs and then someone figured out an elevator. Of course there were tons of people that would never get into one of those boxes that magically took us up. Today, driving is just this thing that we have to do because we want to get to Place B. Why the hell do we have to do it ourselves??

The "love their car" factor - People will still own cars, but they'll be recreational. You'll have a sports car to enjoy on the weekends. But to get to work and other necessities, you'll get picked up.

Google self insuring is an unbelievably great idea. It ends that issue completely. They just say done - we'll stand behind it. As for GEICO and those guys, I don't see them being proponents. This quite honestly is their future demise. The insurance industry as we know it is no longer.

The New Co business model is a harder stretch, but we're already seeing the pieces coming together. The car sharing guys, uber taxi are all laying the groundwork for what we'll have.

Imagine this. You're getting ready to leave work. You let the app know the time. Your smartphone (if we'll even call it that) knows where your heading (either home or your next appt). A car in the area is summoned. You'll have a choice. Pay for a larger car, midsize or compact. Share a ride with others heading in the same direction? All of these choices will just have a different per mile charge. Your account is deducted for your usage just like your EZPASS is today.

Why would we choose to drive??? 
I really like Part 3.   It's a great series.   It has inspired several great discussions both online and off.  
Loving the analysis so far +Chunka Mui.  With 1.3 million deaths per year worldwide, automobiles kill people nearly as quickly as the concentration camps at the height of the holocaust.

The public outcry out to be commensurate with that fact.
Hey Chunka, Been a fan of your trillion-dollar driverless car concept.  I wanted to suggest a good contender for the vehicle of the future.
This electric motorcycle is balanced with gyroscopes. It literally cannot fall over. If you combine it with a network of power stations similar to what Tesla is implementing in the USA, you could charge these vehicles easily(really they would charge themselves). You could also change the seat layout of these cars so they face each other since the driver would be unnecessary. 

Obviously the technology is there and the price will plummet with economies of scale. The question is, will this technology be provided by one company to many others, or will one new company literally wipe the car industry clean? 1) Google licences it's LIDAR/navigation system and gives the APIs and tech to other car companies to implement. Over time, those companies fight to be the main driverless car provider. I can't see how having lots of choice is going to compel many entries to survive in the market. It's basically going to become a monopoly. 2) Google buys a car concept/engineering plant/everything it needs - and makes its own fresh monopoly out of the driverless cars. 

It will be interesting how the government works with driverless cars, I almost think it could be nationalized (conservatives will hate that) but it will put safety first, and the gov't needs to be in line with such a large and important transportation network. 
+Travis Statham I'm not sure you're familiar with monopolies and how they're formed.  Neither of the scenarios you suggest with Google would result in a monopoly, because it is simple for other corporations to create competing standards.

The only scenario which would result in a monopoly is the one you suggest:  Nationalizing the entire industry ^^
I don't think Google is interested in making cars. I believe it's just the software that they want perfected. I think multiple parties will be making them, too. Similar to android, but this time they'll be getting some cash for the licensing. And finally I see lots of the car companies racing to take the lead on it even if it means they might sell less cars later. They know it's coming. No way they are going to be left out. 

The government is not going to nationalize car making, but will certainly participate in the regulations and impacts on the highway system. 
Right, I meant that if the gov't were to say that Google's licensed tech would be the standard  for all driverless cars, then the competing car companies would be forced to use that. I think driverless cars will be so inherently simple that that it would drive out competition because people won't care as much.  Or, instead of having people buy cars, the new model will be a company buying massive amounts of cars that they can rent out to people. When you purchase such large amounts of relatively similar cars, the industry will collapse and focus on the several car makers that make the best versions. 

Google may be interested in making cars - they make their own hardware for their software these days. They certainly have the capital. I think they could easily buy another car company and use it's tech to manufacture driverless cars. The confusing nature of the entire new industry would be perfectly suited to letting one company do all the work. 
+Travis Statham This is the only thing you said that made any sort of sense in the context of reality:

" Or, instead of having people buy cars, the new model will be a company buying massive amounts of cars that they can rent out to people."
Your articles give me hope that driverless cars will be a viable choice for consumers.  Google giving driverless cars to employees should help out a lot to get the word out.  Also, working on lobbing state governments to allow the cars as well.
As far as Google building cars, I think they would want to get all the data produced by a driverless car.   What Google would go through to make sure it got all that data would probably dictate it's involvement in building cars.  
I think we can do better with timing than the vague 'in our lifetime' or 'sooner than we think'. Urban uptake is going to be the most rapid- taxi or +Zipcar  or +car2go customers have no sunk costs to prevent them from immediately switching to a competitive autonomous service, I assume that will happen within five or ten years. Car owners driving in cities are going to switch increasingly when they would have bought a new car anyway- the decade for average car turnover applies there- some of those people may opt for the service model (because it will be superior to current taxi and rental services that are not good enough for them to get rid of their car now) and maybe the wealthier ones would want to own. Maybe half of all urban cars could be autonomous in 15 years, and 90% or more another decade after that- so 25 years plus or minus five years? It ought to look like an s-curve.

+Bob Dunmyer Don't forget that prices would vary with congestion as well.  Someone might choose to share a ride during rush hour because it's less expensive (because there would be fewer service cars to go around and longer average trip times making them less efficient, the service providers ought to charge accordingly), but then for the same money they could choose the luxury of being a single occupant at another time.  I also suspect that planning a trip further in advance would cost less than one made with no notice, so making the same commute every day with the same service provider would be less than wanting a car in five minutes.  The bigger the provider (or the more cooperation between providers) and the more advanced notice with the scheduling the better the service provider could do traffic control, routing cars down different streets, synchronizing them with each other and traffic lights and draw bridges and so on.
+Lucas Walter 5 to 10 years?  Perhaps sooner than you think—if one or more of the scenarios that I discussed in the article play out.  I think that this is one of those technologies that is now more contingent on breakout events than the passage of time.
+Chunka Mui The software seems to be pretty much done. The testing and refining needs to be done to prove to the public that it works. Maybe a couple of years. Meanwhile you could do better prototypes and try to force some action into governments. I think 5 years is the max amount of time until we have a somewhat compulsory driverless transit system. 
+Chunka Mui  The lynchpin I see is the +Velodyne Lidar on the top of the current Google car and the Darpa Urban Challenge winners- it's very expensive and doesn't integrate well into the body of the car.  It's hard to imagine them getting the cost down by an order of magnitude in much under five years, and that is assuming decent economies of scale.  The kind of much less expensive lidar and radar and stereo vision systems that exist now and are sneaking into the front and back bumper and windshield areas of luxury cars and as high end options could be the path forward but they don't provide nearly the same quality and volume of data about the position of neighboring cars and obstacles.  Being located closer to the ground doesn't help with seeing further out and being able to react to events further away from the vehicle.  It is the difference between the Google car and the European models- the lower end sensors provide lots of clues suitable for lane assistance and automated braking and cruise control and parking but they don't provide the same level of confidence for completely autonomous driving.  

The ideal I think would be to field a much smaller device that could retain the fully 3D high speed scanning laser approach, and have multiple lower profile packages on the roof and front and back bumper areas, and combine video and radar also- though the numerous sensors means the cost really has to come down.

I'd love to be wrong about the timing and pricing of the critical sensors, even if I'm right if it's economical to field a taxi service in a car that costs $100K within one or two years of now then Google or someone should definitely do it (and engineer and deploy it in Seattle, I'd be first in line as a customer and prospective employee...).  
+Matthew J Price Well, if you need a platform to bring all parties together to collaborate, that's where the single standard will come from. 
+Jay Payne In my opinion, Google will only go into making cars themselves if they believe that cars need to be "re-invented".  They can't be interested in the small margins that car manufacturers today work with.  Buying a plant or even a car company would not be the problem, it is what they would do with it.
+Jos Echelpoels  Yes, you're right.  Why would Google want the bad margins of the car business?  Car makers have been complaining for years that they do all the hard work and everyone else makes the money.  But if car makers can insert themselves in the downstream services, like they did with auto financing, then things get interesting.
And just imagine how smooth merging is going to be. Exiting and entering highways. As one car is driving its route, the other ones nearby will know the moves it needs to make. It'll make room knowing the other car needs to get on or off.

Lighter weight vehicles due to less accidents means less fuel use. Closer driving as all cars communicate means less fuel. They might not be going 70 during rush hour, but when you remove all the stop and go, the commutes will be tons faster.
There's also a chance that driverless technology will start with trucking companies, speeding up the deployment in the real world. They have a lot of incentive to decrease cost, increase delivery speed, and increase safety...
I really like the trucking idea, and completely  forgot about that potential. Supply chain management will realllllly like it, and will provide a lot of corporate push for such cars. 
+Dobromir Montauk Actually, I believe that public trafic (bus, taxi) will be the first ones to adopt once the security aspect is agreed to be a benefit rather than a risk.  E.g. In Belgium we had last year a school bus that had an accident in Switzerland and more than 20 children and teachers died.  That kind of accident may well have been caused by a human factor.
Hahaha... "The China Card". It's nice to think that China might elect to use high efficiency vehicles to safe their money, but (as a person living here) I can tell you that system efficiencies are typically seen as job destroyers. Friedman's "Why build it with shovels, when you could build it with spoons?" still rings true.

Looking forward to the discussion of Trucks :)  
+David Vorriccelli  Perhaps you're right, though the opportunity to dominate in one of the PRC's designated strategic industries might outweigh that consideration.
+Chunka Mui You're right in so ways with regard to driverless cars, but I disagree that strategic industries would be sufficient cause for China to invest in a job-destroyer like driverless vehicles. There are a few reasons that I seriously doubt the China option:
1. Building infrastructure is a big deal in China. The vested interests are not going to give up on road to nowhere  projects. There is less than zero demand for efficiency here, because big infrastructure & roadway projects keep the party's constituents well fed and permanently dependent on the government's good will. Remember, even though the party is capable of making top-down decisions that apply across the entire marketplace fairly easily, those decisions are made by old-guard, change resistant individuals with financial interests 'more of the same' (similar to your argument about "car guys" in Detroit). If you didn't hear it even mentioned it in the recent 5 year plan, I doubt that you'll see it as a meaningful part of the next one. Further, driverless cars will be too expensive for the middle class and undesirable as an avenue to facilitate the movement away from an export economy, so the party is even less likely to encourage it. 
2. As yet, and for the duration of the mass-migration from the countryside to the cities, money is spent in ways that others can see it.  When the middle class spends significant money here, the perfect phrase is "conspicuous consumption". After houses, the next product purchased is a car. Owning a car here is still about 'freedom' in the sense that American 16 year olds would think of it. Once you have a car you have a place that is yours and it may be the only place where you can have a moment of privacy from the over-bearing in-laws who live with you. It's not uncommon to see a person resting peacefully in their Audi, right outside their apartment. Driverless vehicles will not be 'fleet' vehicles here. 
3. China hates Google. That can not be understated. Facebook, Twitter & Google are all blocked here for reasons related more to allowing for the unchallenged development of sino-equivalents like RenRen, QQ, Baidu. You're finally seeing products like WeChat released in the west, so it wouldn't be absolutely impossible to see a Chinese driverless car in China, but it's certainly not going to be a Google product. 

I stress that I fully support the development and implementation of these technologies, but I really think that China will resist them until after they are relatively mainstream in "the West".  (edit: looking forward to Part 6!! )
Add a comment...