Shared publicly  - 
The Unintended Effects of Driverless Cars

Google has been working on driverless cars for a few years now. The obvious selling point is that the cars will be much safer without a human behind the wheel.

Currently, a car spends 96% of its time idle. Compare that with planes which spend almost their entire lifetime in operation/airborne. Idle planes aren't making money, and they need to recoup their hefty $120M price tag. There is an unforgiving economic incentive to make sure it is always in use.

The proliferation of driverless cars will have a similar effect. Cars will spend less time idle: why would a household buy 2 (or even 3) cars, when they only need 1? Ride to work, then send the car home to your spouse. Need to go grocery shopping, but your kid also needs a ride to a soccer game? No problem, a driverless car can handle that.

What will begin as households cutting back to a single car, will expand. Why would a family need an entire car to themselves? That's crazy! It may start as extended family in the same area sharing cars, then neighbors sharing cars, and then entire apartment/condo complexes in cities offering driverless cars bundled into their HOA/rent.[2]

The operating percent of a car will go from 4% to that 96%. But back to my leading statement: there are unintended consequences. Parked cars will be a relic from the past. What happens to car insurance prices if a driver is no longer part of the equation? And if cars are receiving 20 times more actual use, that would imply that there would be 20 times less cars sold.[1] This is the kind of disruptive change that can reshape the automotive industry. The recent GM/Chrysler bailout may have been for naught.[3]

[1] Of course, this isn't exactly the case, as the cars would need to be replaced more often due to nonstop usage, but the point stands.
[2] Hell, I'd share a car with my condo complex. I currently don't own a car, I walk or take taxis basically everywhere.
[3] Of course, car companies realize this. And I can guarantee you, they will lobby against driverless cars.
Harald Sandø's profile photoScott Junner's profile photoJohn Wilkinson's profile photovaleriya smelova's profile photo
Wow. These are some possibilities that certainly never occurred to me when thinking about Minority Report-style cars.

Ain't the future awesome?
It sounds like you are redefining public transportation as much as anything else.
Yeah, because cars aren't a personal item? why don't we just make it public transportation? oh wait, they already have that. It's called a bus, train, subway, taxi. Sorry I like my gas car, because I can modify it and make it go fast.
Dude, this is an excellent deduction. Nice new vein to think about. For some reason, it's actually got me to thinking of mini-mega trains, where individual units break off and join up for massively travelled routes, and are able to deliver passengers to and from unit to unit.
I don't know that the cars would need to be replaced more often. How many cars are replaced cause they actually stop working after keeping up required maintenance? I'd guess not many. Many are probably from accidents or wanting something newer. So your thinking is probably pretty close actually.
+Austere Grim No one is touching your car. Why are you so upset that other people may want to cut down to a single car per family and save money?
+Bryan Paul in the short term yes, our generation will probably feel this way - but people who grow up with self driving cars probably wont care at all.
Taking it one step farther, it makes electric cars that much more affordable and manageable. They can wait at a charging park until called into service. Cars could be "pooled" so that one is almost always at the ready ... just use your smartphone to schedule your ride. Cool beans. Also might note that an electric vehicle would need far less maintenance/repair.
+Bryan Paul I feel like with driverless cars they become less status and more appliance. Most people don't buy a toaster to show their status, they buy it to make toast. As people share cars, like they share toasters, I believe this will happen with cars as well.

Sure, there will still be people that want to arrive in limos or whatever, but who notices the model of limo? It's just that it's a limo.
Interesting idea. Though, this requires that cars can drive completely driverless, that is, with no one in them. There is a difference between an "autopiloted" car and one that is completely autonomous. That's a whole new level of reliability that is needed. And that over and above the reliability that is needed for an autopilot car. Apparently Google with their cars have gotten to an impressive level. But I imagine we are still a long way off from anything close to this for many reasons.

This isn't to say I don't like the idea or thinking. It's great. But there are many, many barriers even beyond the technical.
Great idea but probably won't be put into effect until long after I'm dead, reminds me of iRobot and imagine a bug or fault in that system :-\
the comparison to planes is incorrect, btw... personally owned planes do spend a good deal sitting on a tarmac. Public transportation planes, are in use, but so taxis would fit equally into this equation.
+Reece King Remember the Star Trek communicator? May not be as long as you think.
Making us more efficient just means we will save money and make more wealth. It's the future.
Ignoring the fact the driverless cars would be more expensive, making up for a good amount of the loss from selling fewer. And as you said, it would be used/replaced/maintained more. A lotta jobs...
I made a public bet with a friend that i believe that by 2031 everyone will use robocars as their primary means of transportation instead of buses or personal driven cars.
Why have jobs that are merely inefficient? Do we still mourn the loss of horse poop scooping jobs due to the proliferation of the car?
+Ken Pietro The jobs thing is a very unstable point on which politicians bet, primarily because the days of large industry, the only things that require masses of bodies, are ending as things become more and more automated. The more automation, the fewer jobs OR the more innovation required to be viable.

This guy has done a lot of the thinking behind this stuff , and is in fact a consultant for Google with their robocar. Really really fascinating stuff. The only barrier practical really seems to be legal ones, such as liability, etc.

I'll just say this: The more money I save from not having to own a car at all, or even maintain a drivers' license/insurance, can be used in other ways. Jobs will be destroyed and created, much like every other technological innovation has done the last 150 years.
Some interesting ideas here. Makes me think that the future of Zip Car has some definite advantages.
it is called car sharing and besides the cars do not move themselves, it is already real over here
you can get a car for as low as 29cents an hour (most do cost a lot more though) and in some carsharing groups you are allowd to drop the car wherever you want

just pic your smartphone and it tells you were the nearest free car is and off you go
+Steffen Arntz The annoying part of car sharing, and the reason I don't use Zip Car, is that I gotta walk like 5 blocks to the parking lot just to get to the car.
I think the main problem is with people (like me) that really like to drive and the different kind of car that we like.
My family have a normal family car but I have a sporty car and I really dislike when I have to drive my parents car, it's slow and it's scary on the curves the way I drive.

I think I'll like driverless car for the home/work commute, I can really start thinking of work related stuff or simply wake up later and fully wake up in the 20 minutes to work.

I drive at least 4 time a month from Milan to Genova (400km trip) and i really enjoy it full of curves on mountain passes, I can't think of that with a driverless car... and probably working during the trip.

Another big Problem how much will it be the benefits from the Driverless cars if there are still human drivers in the road?
If these new cars can account for my wife changing her mind every 20 minutes, I'll buy into it... Otherwise, we'd be very unpopular.
+Koushik Dutta but they potentially could be in a great position if they adopted driverless cars
+Koushik Dutta hmm that is no problem here, carsharing companies have reseved spaces in parking facilities they keep stocked up
and another feature is, parking is free :) you won't have to bother for a ticket or anything
clearly this concept only works in highly populated areas, but it works already, there is no place where you would need to go further than ~200m to get to a sharable car :)
+Austere Grim You're right. Taxi is actually quite close to the "shared car" idea, the only problem is that you have to pay the driver. (and even if you have to pay the driver, it's sometimes cheaper than paying for the parking in the city) Now imagine that you can drop the price down to the price of fuel + car wear + some small margin for the taxi company. You may suddenly realize, that owning a car is a luxury you don't really need. And you don't have to share the car with your neighbor, that's more expensive, more complicated and less reliable than using the car as a service from some big "taxi" company. (With their volumes, they can buy cheaper cars, cheaper fuel and can have cheaper repairs than any private owner)
Wow. I never really took the idea that far, and it's definitely a shakeup but I think I'd rather live in that world than this.
Simon B
+Koushik Dutta - thanks for that. I hadn't really thought it through. But it makes sense. Point 3 makes the most sense too, sadly.
If your car sharing concept relies on the car's ability to drive itself (without a passenger), then gas prices might throw a wrench in the economic feasibility, since any gas spent driving the car without a passenger is wasted compared to having multiple cars. Would only work with non-oil-consuming cars (including fuels that are produced with electricity generated from oil).
The potential is amazing. I would love a driverless car. Being in the UK on some of the most congested roads with the angriest people driving them, driving is stressful. It only takes slight incompetence of drivers to back up a motorway, brakelights multiply like the butterfly effect, and jams are constant. Taking humans out of the equation is a great prospect. Fewer driver errors. Less speeding and slow driving, constant speeds would fix traffic jams. Driving stress, gone. Rush hour crush, gone. And that's just taking the drivers away. This makes the whole proposition even more appealing.
Driverless cars will still be a status symbol, because that's our culture. With the example of becoming a appliance, even with appliance, people buy different toasters, from different companies with different features.

The people that wouldn't care about the make/model of their driverless car are the ones that will pay down on a community version like some people do with a zipcar.

There are tons of carpooling groups out there but how popular are they really.

Even with a driverless car that won't stop people from buying more than one. I can say that I need my car and if my kid were playing high school football and has practices, I still want access to my car.

What this will do is create new markets but the car industry won't suffer at all. If my car was driverless, I would take it on more trips more often, because I could enjoy the trip for the cost of 3 tanks of gas $180 and take 5-6 people instead of paying $300/ticket on flying. I would say that the airline industry would suffer more from this innovation then the car industry.

But that's just another viewpoint, not that anyone is wrong but I can see I vast array of situations occurring from this innovation. I also agree with +Jonathan Berry this all depends on what type of type of driverless car we are talking about. Right now I saw reports that they are working on it to only be driverless on certain stretches of highway and after that stretch, the driver would assume control again.
Driverless cars is the worst idea ever! I LOVE driving. I love burning gasoline! I fill up 93 octane every other day and drive hard. All that would go out the window and the joy/experience of driving a car will be lost. I hate electric/hybrid cars and enjoy taunting them on the road. :)
+Brandon Hilgeman 93 octane, that takes me back.... Minimum you get in the UK now is 95, but more commonly 97 or 99 octane if you pay a litter more.
You just described taking the bus. The only differences are the bus has a driver still and you have to pay a fare.
+Chris Harrison You're fortunate. Here its, 87, 89, 92 or 93... There were some pumps here in town that used to have 105 but it was $5/gal.. 87 here in South Carolina is $2.95/ gal, but I always opt for the 93 at around $3.30/ gal...
This isn't a viable strategy for breeders. I could not afford to send the minivan all over Hell's green acres (commute o work and back twice a day would cost about $500/mn) and comunal service vehicles, public or private, will not meet demands for holiday travel. I still need 2 cars and will eventually need 3.
While I think it's an interesting perspective and I think in the long run you'll be right (by the long run, I mean another generation away), I have to wonder, "So what?" Yes it will be destructive to the industry. Cars were also destructive to the horse and buggy industry. Computers were devastating to a huge number of industries. It may be a problem in that car companies will make it more difficult for driverless cars to be allowed on roads through lobbying, but I don't think it's a societal problem. Either the car companies - scratch that, the American car industry; Japanese and German automakers generally already make far more than just consumer cars - will adapt, or they'll die. It won't be an immediate shift anyway, so they'll have a few years to try and figure out how to change. So my question is: so what?
+Miroslav Prašil I know what you're saying; but, for example, those who live in the boonies, where does this shared vehicle help when I drive 20-40 minutes away nearly every day. Yeah it can be computer controlled to pick up the next passenger, but then we're back at it's not "my" car, or a few owned car. (It's a poor concept to share a personal, 1-5 man, vehicle. Economically it doesn't work.) Where yeah if it became more public transport, okay, great, gov't sponsored, like a bus. You pay for a ticket, etc. Now what about vandalism?? I don't live in the greatest of neighborhoods, and these types of areas and people who don't care about this vehicle are just going to abuse the privilege and destroy the vehicle... In a perfect world, where no one is "bad", and we all worked every day, and have 2.4 children. Live in a nice 2 story house on a cul-de-sac, in the suburbs. Yup this would be great. (But this isn't the world we live in, and people don't magically become naturally "good" in 30 years.)
It just take human behavior, without supervision (like a taxi), into the equation... Sure, a perfect world where everyone does the same thing, it's fine. You could easily throw numbers into a spreadsheet, or a model with randomness, and see that for a small 1-5 man vehicle, the cost grows exponentially. Due to damage, vandalism, wear, upkeep... where public transportation like a bus, is economically more advanced...
+Brandon Hilgeman that's cheap fuel. Petrol now starts at £1.29 a litre, and if a gallon is about 4 and a half litres, well you can scale for that and the currency difference of £1=$1.60 ish... Yea we have good fuel but we pay for it.
This is an interesting idea, but I think it fits better for urban environments versus rural or even some suburban areas. People are too widespread in those areas to really see the advantage of sharing cars, but the driverless vehicle would definitely be a godsend for people like me who have long commutes - especially through a somewhat dangerous mountain pass.

But with the rural/suburban areas, the cost of driverless cars would be prohibitive for their benefit, and for those people they'd probably still driver their own cars. The benefit of intelligent driving systems then is diminished since there is still a human on the road that can't communicate to other cars. I could see a semi-automatic driverless system like in iRobot, but because of the unpredictability of humans on the road, the advantages of driverless systems would not much in safety as it would be in convenience.
+Sean Payne I think it was BMW (On an episode of top gear) that had showed a concept car, that you could record a track (road) and it will drive it as fast as possible, beating even their best driver. This took into consideration that the computer knew the weight of the vehicle the turning ability, scanned the road and knew the conditions of the road, understood turning capabilities, and handling. The computer did it all on the fly, with GPS coords, and knowing the car. (knowing the physical location, weight, the direction, pitch roll, speed, angles, air pressure for drag, torque break point for the wheels; basically modeling the conditions of the car before it does a maneuver.)
It literally put the pedal to the floor, and raced the car for the driver. It's not impossible for car manufactures to do this in every car, and autonomous vehicles would drive 120+mph and be safer than anyone ever could be in a car. Along with stopping capabilities in cars now that will stop the car safely if it sees a wall or obstacle in the way of the path, long before a driver could react.
granted the computer/technology to do this probably adds $5k-10k.
As far as i remember Nicola Tesla scientist was working on a project of making electricity wirelessly,consider charging your car as you would use 3g network.I myself belive it is posible and the future really have some nice surprizes.
Drunk driving will also become obsolete, so the weekends will become extra drunken. Big Alcohol is the clear winner here.
Yes one of the related unintended consequences would also be lesser emissions and traffic congestions because of reduced number of cars on the road... Isnt that something we would love ?
+Steve Kondik The thing is, this'll have both the support of Big Alcohol AND Mothers Against Drunk Driving.

+Ashwin Guru Techincally 20 cars being used 1% of the time produce the same emissions as one car being used 20% of the time. What it would do is spread the traveling times out so rush hours wouldn't be so congested leading to less wasted fuel.
This is a good example why the actual economy system will doom us all. It rewards inefficiency and lead us to endless consumism.
I imagine a "driverless" car with sensors would not take a left in front of me like many "driven by human" cars do now. OK as a motorcyclist I say go for it, the fewer idiots behind the wheel of cars the safer I am.
Cars will go the way of home cooked meals...slowly life engages us in such a way that we compromise. Another status symbol will rise from its ashes...An icon of western wealth and waste. But, what about farm equipment? Maybe we'll all end up like those fatsos on wall-e. Or more so.
Works only if time is not a shared commodity, which it is. Car time for you is car time taken away from other people. Say you need to transit 30 mins east of town to work, while your spouse needs to go 30 mins west of town for work. If spouse goes first, it'll be one hour until you get to reach your work place. Of course we can always reach our destination early, but then, there're always circumstances where that cannot happen.

Cool concept though.
Haha, so much from our driving time being a sanctum from the profiteering world. Now we can (and will) be expected to work during our commute
It's interesting to consider this from the perspective of energy consumption. It doesn't seem like a particularly good use of energy for a car to drive my wife to work, then come all the way back home so that it can drive me a few miles to my appointment. Of course, this can be mitigated in the shared use scenario - but that's more of a replacement to taxis than anything. Still, one could imagine some economies of scale that make taxis more cost effective when you remove the driver, better optimize the distribution, and assume a more consistent reliance based on predictable use cases (e.g., commuting patterns).
+Jeremy Caney I was thinking the exact same thing its rather an oversight to forget the energy use of the empty cars in the family use situation.
Also, right now car companies have an incentive to build cars that last 5-10 years. Not that they can't be maintained longer, but it's not cost effective to produce a car that will last 50 years, because few people are willing to pay that much extra for the level of engineering and quality control and componentry required to make that happen. And, in fact, it's arguable that the market has been evolving quickly enough that there's a good case for that - in nearly everything but style and horsepower, we're significantly ahead of where cars were at in the 1960s. That said, if cars were being used this effectively, if one car could service an entire family or if a handful of cars could service an entire apartment building or community, that changes the economics a bit. Things like the standard "100,000 mile / 5 year" warranty become obsolete because the usage pattern changes. I'm sure the car companies would be happy to sell you a $100,000 vehicle that has a 500,000 mile warranty on it. And if they're not, another company will. That's the beauty of a free market.
riding in car without the need to drive it, lots of people sharing the car? That's called a bus, it already exists :)
+Cory McConnaughy: I partially agree in the sense that this would require better distribution of commuter hours, but that's been an incentive for decades in response to traffic (from the user perspective) and infrastructure (from the state's perspective) and yet this is an institutional shift that's been difficult to fully accomodate. In fact, many companies that have experimented with it come full circle, due to the difficulty in coordinating efforts when people aren't universally available. That said, I do think there are scenarios where this makes perfect sense and could reduce the number of cars on the road - albeit, at the cost of energy utilization, as mentioned above.
Thats quite a vision.

I would love to see it executed...this calls for a huge shift in the mindset of not only the consumer but also the manufacturer.
I don't think it works like that. You can't assume that people can wait on the car to be available in order to travel. Rush hour doesn't happen simply because we all think it's a grand idea to go sit in traffic twice a day, every day. Multi-car families frequently need to be in two places at the same time: while one parent is sitting in traffic coming home from work, the other is taking the children to after-school activities.

You won't get anything like 96% utilization ever (even planes are usually only active 10-12 hours day, save for the longest intercontinental flights). The problem with mass transit of any sort isn't the cost of operating the vehicle, but actually getting utilization to make the cost worthwhile. Peak periods and routes are brought down by long periods of hugely inefficient use and inefficient but necessary routes. This includes airlines too, to a large degree. I don't think even the notional speed and fuel economy gain by having the cars go faster (since you remove the "stupid human" element) and not having to pay a driver really changes the economics all that much. Having the car off when not in use probably improves the picture, but maybe not enough to matter.

It probably works well as a replacement for driving to the subway/train/bus station and taking that to work that's common around a lot of major cities. It works a little better than buses in suburban sprawl, but not as well as one would imagine. Cars still have to be distributed to ensure reasonable response, which means you need a lot of them handy. As population density drops, the numbers just get less and less attractive. You also just can't park them anywhere, unless they can be readily moved on a whim (I may need my driveway later). And remember, all those cars that are out here in the country aren't cars in the city where they can be used by urban folk who need to get around.

In short, I think you're trying to optimize the wrong problem.
+Adam Skutt What makes you think you can't order a car in the future, like you order a pizza now? It shows up when you want it.
I hope Google can make this happen soon! That would be a very welcoming change to the car system that we have now. My car sit Idle probably 99% of the time because I only drive 2 miles to work and my girlfriend lives on the way!
Interesting idea. Maybe i can buy a car and get back the cost by giving it out as driverless taxi.
+Troy Mcilvena Not really. My 5-seater compact car gets 75% of the miles/gallon of a 2-seater Smart Car in the USA. A few these days get equitable MPG.

You have to go to something like a motorcycle or diesel to win on MPG, both of which have a whole host of other problems.
Thanks for always doing the right thing google - all services you provide only seem to help mankind. If only all companies could thrive on good principles like google does.

I am shocked that there are so many naysayers commenting on this post. I guess there are a lot of people that are scared of change and lack vision.

Technology like this holds the answer to many of the questions the country is dealing with:
1. Reducing our need for foreign oil
2. Reducing CO2 emission (even if VMT were the same (although I would venture it would be reduced) everyone wouldn't need one or two 6000 lb SUVs that are really one needed for 1% of their trips), so the fleet would have a much higher mpg on average
3. Getting the elderly off the road and providing better travel options to everyone - might even make public transit meaningless
4. Saving Americans thousands of dollars a year locked up in the capital to purchase and maintain a vehicle (albeit that all the potential issues that Koushik raised about effects to the economy are valid - savings to you means an impact to the economy - however, I would venture that the economy would shift and survive).
5. Decreasing accident rates and saving lives

Just as a few benefits off the top of my head, I am sure there are many more.
Good idea, till the car gains self awareness, views you as a threat, then tries to kill you!
Or, you know, you could take public transit!
Here's an apologetically long and snarky alternative perspective...

I think it's an interesting thought experiment but Lucky for the car makers, you won't be sitting in a driverless car anytime soon.

I think this is a conversation being had at minimum 20 years before its time. For god sakes we're still using combustion engines in 99.9% of all vehicles. This reminds me of when I was in 5th grade and they were hyping up Moller Flying Cars( ... Michael Jackson even pre-ordered one back in the early 90s.

The exponential growth of a processor is not the exponential growth of technology as a whole. See Peter Thiel...

Technology requires commercial viability and adoption to be successful. Anyone remember Iridium? (

Let's assume that tomorrow the technology was available and licensed on all interstate highways...

Let's first look at consumer demand...
-Why do you think that there are classic car collectors paying millions for cars?
-Why do men spend thousands tricking out their car to make it louder and faster?
-Why is NASCAR and Auto Racing as a whole one of the world's most popular sports?

What an amazing future... We roll out of bed a robot dresses us, we are placed into our automatic car, driven to our office where we sit in a desk and our thoughts are wirelessly transmitted to the computer which dictates our very important ideas and we are fed an optimal amount of food daily at 12pm in the form of nano supplements.

Can't wait for this blessed future...

Google, can you work on the Star Trek Transporter instead?
won't work for multiple uses, or how is it different than taking the taxi? i think most people choose to drive themselves because they don't have to wait for the car or specify a schedule for the car.

car companies against driverless cars? don't kid yourself google. Toyota is years ahead of you in development. they've had one for you to try out in their Odaiba(Tokyo) showroom for a few years now. it's not in effect because most consumers aren't for the idea, even if they say they are, how much do the general public really trust a computer when there are other drivers on the road? and the infrastructure costs for integration on pre-existing roads?

i think it's fair to say, as a start this will work better first as a separate lane(or replacing carpool lane) on the highway for popping an autonav switch. baby steps or it won't happen.
Think of all the extra gas. I drive to work and back 2 trips. My wife drives to work and back 2 trips. My car now drives me goes and gets my wife drives her to work then gets me brings me home then goes and gets her and brings her home.... So instead of 4 trips it is now 7.

I can see something like this for public transportation, Taxis and what not that can manage where they are picking up and dropping off but for the average person this seems a bit much.
The comparison between personal vehicles and commercial airplanes is an invalid one in my opinion. How much time does a personally owned Cessna spend idle? I would imagine it's much closer to the 96% figure than a commercial jumbo liner. I would also imagine that a company-owned (not privately owned by the driver) taxi cab spends close to 90% of it's time in utilization also.
Of course then you think of the gas that will be used. We will have to transition to hybrids to say the least. In Tokyo, 90% of the air quality issues come from taxi emissions. The cars are on while waiting for customers. Cars are not built for running like airplanes. Maybe future cars will cost 200k, and can be used for 20 years? I'm sure that the cost structure will just shift, and will end up costing more somehow...
Here's a related questions: What will cities look like with driverless cars being shared by groups of people? Parking needs will diminish significantly, allowing cities to become more dense, yet because the cost of driving will be reduced (by splitting the ownership costs of cars, by increasing the rate of shared commuting, and by allowing commuters to spend their transit time being productive), demand for urban land for residential areas may decrease overall. So my best guess is that in many cities commercial areas will become more dense, but residential areas could enter a decline. Any other ideas?
GM wasn't bailed out - they filed bankruptcy. I know, I owned a few shares of stock that instantly became worthless. However, I'm hoping for driverless cars because I'm 63 now and I don't want to become a danger to other drivers when I get older. But I do still want my independence. If we get driverless cars, you will all be much safer. ; ) Also, there is a very successful car sharing business in Chicago now. I think it is the wave of the future, especially in urban areas.
Yeah, +1 to the argument that this sounds pretty much like the brilliant idea that started what is now called public transportation. So a better argument would be that it would eliminate the jobs of bus and taxi drivers while possibly making this mode of transportation safer. But I don't want my car smelling like the STM buses....
Why are dirverless cars a horrible idea just because you like driving? I think it is worth noting that no one is implying that driverless cars means you will never be allowed to drive a normal car again, or for fun. I don't see why this cannot be a great general transport idea and not a replacement for driving a car yourself. You can still ride a horse if you want to right? (Well maybe not to work in a city and back, but you can still do it for recreation.) Point is if driving is something you truly love, you will do it either way.
Good - get the 'appliance drivers' off the streets! Then, when I drive my car I can enjoy it even more.
I'm a huge proponent of AI driven vehicles. You are very correct to point out the insurance implications. This stands as the most significant barrier to the adoption of driverless vehicles in my understanding.

Of course the concept of a 'car' would be shifted significantly by this technology. A vehicle would be a mobile power-source and wireless transponder. The driving performance of the vehicle becomes much less of a feature of the vehicle's design.
+Chris Harrison +Brandon Hilgeman I believe octane rating vary between countries. Here in Australia, same as the UK, the Research octane is shown, ie 91, 95, 98. In the US, the Anti-Knock measure is shown, which is typically around 4-5 points lower. I think we all use the same fuel.
+Brandon Hilgeman +Chris Harrison +Jordan Hotmann You have to keep in mind the conversion between RON (used in Europe), MON, and R+M/2 (used in the US). The actual antiknock properties of "regular", "premium" and "super premium" in the US and Europe are roughly the same. And the lousy octane rating of high-altitude gas only becomes a problem on forced-induction cars.
I think where this will end up is like a low-infrastructure-cost version of the various Personal Rapid Transit proposals that have been dreamed up over the years.

I don't think privately-owned or human-driven cars will go away. I think driving a car will just become a hobby for rich enthusiasts, much like flying a plane is today. And some enthusiasts will want to own their cars, driverless or not, much like some people own their own planes.
I imagine maintenance is going to be expensive on these things if it all works like that. I would imagine there are going to be more points of failure than on cars currently on the road. That is going to make maintenance and repair more specialized and costly. Cost of ownership will be higher than with a conventional car, I would wager. That would almost require the proposed group/community ownership.

The car company resistance is probably a minor obstacle. With forward thinking leadership in car companies (a big "if") they could transition from almost purely production to production/service. Factory authorized and trained service centers could cater to all the maintenance needs. This might almost be required with the added complexity of the systems over what we already have. It would be even better if the car would just drive itself in for regular maintenance in the middle of the night and/or other regular idle time.
I think if the kind of usage you suggest becomes reality then the cars would be engineered to sustain the increase of wear-n-tear. Necessity, demand and purpose drive innovation, pun intended. :)
But that will put an end to drinking and driving! Why would anyone want to do that?
I would imagine that in a future where this kind of complex vehicle autonomy is possible. Would there really be a reason for me to go anywhere?
I don't think driver-less cars are coming anytime soon. A car that can drive itself with a person in the driver seat doing something else, yes but an empty car no. The reason for this is while a driver-less car can handle 99.99% of the situations it encounters that .01% it will fail hard on. Specifically I don't think a driverless car could follow directions given by a police officer, nor are we anywhere close to being able to do so. These situations are rare per car, but happens everyday and with a large number of driver-less cars this could balloon out of control and cause congestion everywhere.
I don't know about you, but everyone in my family goes different places at about the same time. There's still a time gap between getting the person there and the car going back to pick someone else up. Also, this would pretty much end spontaneous usage if you needed to call in your car from someone else. I'm not doubting that driver-less cars will happen, but one car per family might be tough as people's schedules get busier.
everyone needs their car in the morning and after work. it sits idle during the night and day... drivless cars would not help this..
boo Jay
This would make a pretty incredible taxi service, that's for sure.

Who wants to share their Porsche with me?
The creation of car not driven by humans is a great idea. It may collapse one industry but it will bring in another.... Jobs may get lost due to the creation of a fully automated vehicle. There are also plenty of ways to prevent from lack of production and that is if people were to share cars like that then it wouldnt be 1 persons car but like a cell provider youd pay monthly to use the transportation device. It would also prevent from human error. Like taking away jobs it would create them... There will always be downsides with the upsides of innovation... There would be a lot less deaths as well, what most people don't know is America is actually in a negative population boom. Less kids are being born to a set of parents of this generation. Many imply that it would be a bad thing but there is already too much human fault on the road that it would not be smart to find ways of fully automated non-human driven vehicles. It should be implemented ASAP... AI> Human Emotion + 3000+ lb deadly weapon. And everyone is already looking at screens, cell phones, TVs, PDAs, tablets, laptops, even reading kindles while driving, why not let people get this and their work done while riding in a fully automated vehicle to work, school, home, etc. this would give so much more free time. No one would ever be late, timing would be accurate, no one would have the excuse of traffic, bc there would be an estimated time of arrival with a +- of like 5 minutes. And there are plenty of other ways than just motors with AI brains, such as guided roads with electro-magnets<- also in testing.... This idea has been worked on for quite sometime now, and I think people need to expect and hope for the future more than finding its flaws. Its takes some hard falls to get something right. even guided rail systems all throughout a city and not just in certain perameters. The major major downside is if the product failed it might cause a few problems. but that is why now-a-days there are emergency backup of backups.... like multiple backups backing up another backup :P. fun phrase... anyway I wish they would continue with the push toward fully-automated vehicles, it would solve quite a few problems we face today!
+Koushik Dutta Hey, I care what model airplane I fly in. Of course I am a little biased.... Cough.... Boeing.... Cough. ;-) If it ain't Boeing I ain't going, lol .
Quite frankly, this would never work in countries like South Africa where the cars would be hijacked more often than not.
Lojack really does fix problems such as that. I don't see thieves making it very far when the gps embedded in the roof of the car activates.
Good points, but I think you are forgetting about people's laziness and also the feeling of... freedom. The comfort of having a car, which is available at all times and is not limited in any way is way too precious to be replaced. Again, your points are very valid, but I see them more as an addition to current transportation system rather than replacement...
The big benefit for the short to medium term for driverless car research is in better driver information/safety/enhancement systems. True driverless cars -- where you wouldn't need to be paying attention to the driving process, invoke so many complicated policy issues (not just liability and insurance) but also nontrivial human factors issues. For example, as we've seen in airliner disasters, autopilots tend to create situations where pilots become inattentive (or inadequately trained) to deal with extraordinary situations that occur outside the programming of the "normal law" mode of those autopilots (or when those systems fail). Even after all these years with autopilots and highly skilled human pilots that operate them, these issues have actually become worse with newer generations of aircraft with more advanced systems, not better. So it's going to be a long road. No pun intended.
The fault with this is that we are depending on people getting in the car and knowing where they want to go. I doubt that people will so easily give up or share their primary means for leaving their home.
wrt 1. It's an interesting subject.

Let's assume for the sake of an argument (the exact distance is not really important), that a car breaks down after 100,000 miles and requires a major repair that could cost higher more than the value of the car. Hence -> scrap.

Average commute distance in the US is about 30 miles a day, so it will take about 9 years for a car to die and be replaced (it can change hands multiple times before that, and this number somehow should influence the rate of new cars being build, let's take the best scenario: people buy 1 car in 9 years).

If we assume that this is 4% of car usage per day, and increase it to 90% then we will reach 100,000 in about 3 and 1/2 months.

Let's assume a car has a single driver, so 4% to 90% bump is a daily usage of ~23 drivers. In 9 years (best case) they would buy 23 cars. With the new "burn" rate, this looks more like they need ~30 cars to travel the same distance in 9 years, more than 20% increase in car sales and manufacturing.
The future sucks. Where's my jet pack? 
Driverless cars also means that alcoholism would rise as drunk-driving would not longer exist. Everyone will simply get trashed every night and get escorted back home by their driverless car.

The population will also increase as people are no longer (or very rarely) killed in automobile accidents.
stop dreaming about driverless cars okk US people try driving in INdia with goats and cows on the main road and with children playing cricket . Do you even think a driver less car can see and understand when a road crossing guy just signals with his waving hand that hey you can go ahead ?? can a driverless car distinguish wherther the guy crossing infront is old or young to be able to cross swiftly or run ??? all crap imagining people stop dreaming ...
If car sharing on this scale is really going to work, then we've got to get past the idea that everyone must arrive at work at the same time, and leave at the same time too. Staggered business hours are a must.
+Ashwin Guru There wouldn't be any fewer cars ON the road - that number relates to the number of journeys people need to make, which isn't changing. There would be fewer cars OFF the road. So emissions and traffic stay the same.
This would destroy second-hand car market, but will not affect the primary market that much, because cars are driven up to the end of their lifetime at the moment by 2nd or 3rd owners. The cars will just get faster to the end of their lifetime, which leads to a higher frequenzy of replacement investments
Willing to be most people who support driverless cars, or think it'll be the standard whether they like it or not are from urban areas. When you have the easy option to hail a cab or jump on a train, you never really develop a connection with having your own car. Im a single guy with THREE vehicles of my own. A truck, a 4 door 'winter car' and a 2 door. The truck is utilitarian, the cars are my children, and nobody is going to change that.
Unfortunately, Car driving seems to be part of each persons need for expression and free will. We are nomadic often in nature, and cherish our ability to" journey" as part of our life path. It is "unthinkable" for many to have their "wings clipped" so they cannot journey as they wish. If they choose a partner or companion for a jaunt it would not be because they have been told to. Individual transport is the expression of our free will and for many sactisanct. I am certainly for better and more efficient fuel and resource useage, but car manufacturers have long had the technology to make vehicles that are more efficient and much less expensive. It is not in their or associate industries to develop these technologies until they have sqeezed every last drop of money out of their outdated investment.
Ultimately our own nature plays right into their hands (pockets!!) As soon as the petrol runs out, there will be a miraculous discovery of new technology to transport us about... and for those that can pay for it.. it will be be individual and freeing.. the cycle of technological birth and redundancy all over again!! Isn't their some joy in being able to drive? there is always a choice as to how we get about. Isn't There ? :-)
Wow. How did you even think of this? I wish i had this skill because i can never project unintended consequences about anything. This is brilliant.
Combine Artificial intelligence with Driverless cars and in the future we might see cars that have free will and own themselves.
Or manufactors will increase driverless cars (and insurance) 30x ...
I would argue that these are the intended effects of driverless cars. Driverless cars are just as much about saving lives, as "fixing the design of transportation systems in the city". I expect both goals will have tremendous positive economics impact. Inria - Cybercars and the city of the future
'Sounds like a lot of passenger-less cars as well. I don't think there's room on the roads.
Hey you know what. I was just thinking. With driverless cars, if the concept becomes reliable and workable, you could even cut out the postal and delivery system for a fair number of transactions. Imagine having to go to the post office to send a small parcel to one of your friends. You could just stick it in your driverless car, input the destination address and let the other person at the receiving end attend to their package and send the car back - (although there would be security issues with that, but its all do-able)
+Jeremy Bell Not quite so. If you think about a trip where a parent is taking a child to the cinema and then driving home alone. The gas on the return journey would be used whether or not there was a driver in the car.
car companies dont even like energy efficient or well built cars. Why would they want a car that can drive itself and doesn't get into accidents? However to fully make a driver-less system practical a network should be established that also links them allowing greater mitigation of an accident by allowing them to share data and space on the road. Companies that make cars dont have any issue making disposable vehicals, terminating hundreds of thousands of employes nation wide. Companies have no issue sending just about all manufacturing to other countries to be made by slave labor or robots. The USA government has no problem letting all the manufacturing in the USA be sent to other countries.

It's funny though, with respect to the developers of the automated cars.
-Small personal flying vehicles are nothing new. But receive every little attention and very little development, thus keeping costs high. But why? Because the FAA refuses to allow personal vehicles in the air citing concerns of tracking and safety. Of course if they were not using 30+ year old radar and information systems, this wouldn't be an issue. In the day of GPS and Cell Phone based internet services a vehicles personal automated flight and navigation system could pull data from the same sources a commercial airlines pilot can, and in a few seconds or less make course corrections. Air craft automated flight and navigation is nothing new. This used to take massive computers. But most personal PCs have more computing power now. In-fact I would wager your smart phone now has more power than a commercial airliners nav system. Yet the FAA still says the air is unsafe. -Safer than on the ground with drunks, and ego maniacs on the road. -what it comes down to is fear.
-FYI, if you were to buy a flying car. It is very unlikely you would be allowed to actual fly it. The systems would be entirely automated. Just as your regular car soon will be.

Eventually cars will drive themselves. Therefore the cost of using a taxi will drop to approximately the same price per mile as having your own car. As you're not using your car most of the day it will be inefficient to own a car.

Once everyone uses public transport the system will become much more efficient, when you book your journey the system will find other people taking a similar journey, and bundle them together, 4 people will share the taxi, bring costs down even more, if 10 people are taking the trip a minibus will be used, 50, a coach or bus. If longer journeys are required, or more people are travelling the smaller vehicles will travel to the train station, or airport. Journeys can be timed to the minute when there's predictable traffic flows, and if you're rich you can travel alone and pay for the pollution you're creating to be removed.
+Koushik Dutta I'm sure that's how it would work, but the car has to get to me sufficiently quickly, or I won't want to use it. The nearest town of any actual sort is almost 15 minutes from my home, about half of my commute (and partially the wrong direction). The nearest truck stop is 7 minutes (the wrong direction). Outside of rush hour, any daily trip I make is unplanned. The change you're asking me (collectively, us) to make in planning our needs to go out is quite considerable. On weekends, it's usually no big deal, but on weekdays it would be a big deal because what little time I have at home is already quite precious. Moreover, in the winter, you can't leave the cars on the roads because the plows have to have access. Driveways pose a similar problem.

I really think you should take a look a population density map of the whole USA and then figure out where the cars would have to end up (or write a simulation to do it) and how many you'd have to have. I'm quite confident the numbers get really ugly once you get out of the cities really fast. The only public transit that comes around here is a school bus, and they don't even offer door-to-door service (not that I'd want it turning down my street anyway). Plus, school buses take everyone to the same destination and have static routes that can be optimized well in advance. Your solution has neither luxury.

+Troy Mcilvena You're correct that decreasing weight can improve fuel economy, but the weight decrease from 4/5-seats to 2-seats isn't enough to make a practical difference, especially in the USA due to various safety requirements. Seats just don't weight that much, and the added space doesn't increase drag enough to really matter. The margin just isn't wide enough to be compelling: even if I only use my back-seats a dozen or two dozen times a year, the car rental cost alone means the 4/5-seater is the better buy.
Yes, you are 100% right, car companies will certainly lobby against Google's driverless cars.. but you forgot to mention the insurance companies- they will also lobby against driverless cars. These corporate bastards are the reason why America always lags behind Japan.
+Raymond Lulling For all the outsourcing of manufacturing, the USA is still the 3rd largest manufacturer of cars and other automobiles in the world. You're greatly exaggerating the situation.

Also, your understanding of how commercial aircrafts communicate with the ground and what's required to keep everyone safely airborne is very poor. GPS and cell phone connections won't create a personal aircraft revolution and how can they? Neither is enough.
+John Young Hybrids aren't a exactly a slam dunk either. In a city, it makes good sense. For shorter commutes, it makes OK sense or at least isn't any worse. For extended commutes (largely highway) it frequently makes no sense, as long as prices stay as they are. Even if prices fall considerably, it's tough to come out ahead over the life of a vehicle. Especially if you compare to them diesel. Though as it stands today, diesel is also hard to make win in the USA. Both my wife and I have 30+ minute commutes one-way, so I've done the math on all of this.
a car is trialled a thousand times going through all the pressure test... checking all the test...
Excellent post +Koushik Dutta , thanks for sharing such a great vision. Another angle to explore is what role is Google targeting in this picture. They probably see the potential you describe.
Okay, I think it's time to claim the domain - #cod for "car on demand" - and start writing some html5 code and some apps ...
I just hope they keep Verizon's hands off it, so we dont end up paying more for a auto car when we need to use it, than we would driving it ourselves..
Been saying this for years.. only take it one step further... a driverless system could transform mass public transportation... all you would need is the fleet and a public system of ordering transportation to pick you up wherever you are and drop you wherever you wan to go. The IP world + Driverless transportation = revolution of mass transit.

and to +Disha Patel, Why? cause 99% of all accidents involve drivers. Computers can scan, analyse and react to 360 degrees of input while the car is in motion... a driver can see in one direction at a time.
need to hurry up and change to a Resource based Economy (RBE), as opposed to our current economic model where profit incentives kill sustainable innovations such as this. Jacque Fresco has referred to the car situation in his information at
This is an interesting topic with many societal implications already mentioned. I wonder if anyone has created a mind map of this topic?
I'd still like my own car, thank you. I mean, who's going to clean that thing when 500 people are using it daily? A car is an extension of a home to some folks. Might be tough to make them turn that into public transit
Adam, Diesel isn't a slam dunk becuase the auto manufactures wont promote it. Good technologies are often left to defend themselves while lesser efficient technologies are hyped every time they up the ability by a notch. Unlike gasoline diesel is a renewable resource and although shortages may be possible, it will never run out.

Adam, As for comercial air craft. I understand far better than you will give me credit for. I studied flight in elementary, I studied recreational air craft in high-school, learned about the FAA standing on flying cars, and have followed the technology and policy ever since. In-fact several of my customers, are retired commercial pilots and a couple are private.

Hybreds... Adam is right. Their claim to fame is stop and go intercity driving. -Highway is no significant pay off, unless you run out of gas, then you have a short emergency run in limp mode. Ultimate highway fuels are solar hybred, Diesel, and Hydrogen. Sadly only one is even slightly common.
Wow! You've invented taxicabs.

If you lived in New York, you'd know that large buildings have been running car services since cars existed. A big chunk of ZipCar's business model is built around this.

Welcome to 1920.
I'd like to see a real study on this. I'm sitting in an office complex with probably 4 or 5 thousand workers that all pretty much show up an the same time and leave at the same time. Across the area there are probably a few 100 thousand on the same schedule. All of those cars have no need to go anywhere.

There are so many areas in the auto industry that could improve their environmental impact for so much less than making cars driverless.

Think about your desktop computer and how much of it allows replacing parts from many other computers. Think about your car, there is very little of it that you could replace with anything from any other model car.
Seats, no, brakes, no, heater/AC, no, gauges, no, mirrors, no, lights, no. Engine and nearly anything on it, no.

How incredibly inefficient is that? I'd like to see someone design an open hardware truly user serviceable car. Leave a place to plug in the robot driver. Being able to keep a car on the road for 20+ years would save way more resources than a driverless car ecosystem that will never happen will.
There are already efforts to enable individual mobility for people without a car (like ). Now imagine a fleet of small, autonomous cars roaming in your city transporting people on their own. No more searches for a parking lot, far less costs for transportation and so on....
It is your third point "[3] Of course, car companies realize this. And I can guarantee you, they will lobby against driverless cars." that bothers me +Koushik Dutta because I suspect that it is probably true. Surely any who prohibit or delay the introduction of driverless cars will have the blood hundreds of thousands of needless deaths on their hands? For this reason if no other these cars should be introduced and those who would stop them should be prevented at all costs from doing so.
Don't get me wrong but you should just cut our balls and private us from one of the last true freedom that an average person can have.
A driveless car is just one more walled garden. Nowadays cars are boring because car companies made them boring so people who can't be bothered to learn how to drive and behave in a public road can still drive a car.

Basically car companies said: "You are too lazy, have this boring boat with 4 wheels."

Google now says: "You are to lazy, stupid and dangerous, have this car but don't touch anything."
Never gonna happen. Would cost the auto makers entirely to much. They've got the money to lobby, and they would also have to be the ones to develop the cars.... So yeah not gonna happen.
+Raymond Lulling No, it's not the car manufacturer's unwillingness to promote diesel that makes it unpopular in the USA. It hasn't managed to escape it's stigma from the past (and potentially never will) and so many buyers simply aren't interested. Diesel is quite popular and does well among those who support it, but has never been able to break out of that niche even after all the problems of the past have been fixed.

The problems with your claims about aircraft is physics simply says, "No". Unassisted GPS isn't sufficient for plane navigation, especially for take-offs and landing (but terrain is a problem too). Building a comm network to pass the GPS information around isn't trivial but also necessary. Cell networks cannot be used reliably because of co-channel interference problems at altitude and a lack of ability to ensure adequate service. Your suggested solutions simply won't work.
Super Cool! And cars can tail gate (just like train wagons) to reduce traffic jams and way more efficient to use the length of the existing roads. So much benefits ...
That will be the coolest this ever. It will save the world!!!!Implementatiooooooooonnnnnnnnnn
It doesn't have to be a major auto company. They've outsourced all their parts, and now small startups can compete a lot easier. Heck, Google could do it.

First families scale down to one car. Then they start renting it out when they're not using it. Then somebody makes a web app to help them do that. Google could do that too...every car it sells could be hooked into the system. I wonder if driverless cars can be considered taxis under the law...
It could even help with climate change. One problem with electric cars: you have to plan ahead or risk delaying a couple hours while you recharge. A fleet of driverless cars could keep itself charged, and if you're on a long trip and run out of juice, just have another car pick you up. The Pony Express lives again.
I think the real paradigm shift will be realized when we stop thinking of "driverless cars" and when we think of "driverless tractors". There is no reason why the "car" part of an automobile can't be personalized, and separated from the drivetrain. I think GM had embarked upon a project to do this at one point. Historically, having a carriage company custom build your "car" was the norm, a BodyWorks. Back to the point, if the "driverless tractor" looked more like a Roomba, and engaged the "carriage" at a common interconnect…well, we'd all be the "trailer" for a fleet of small-scale robot truck drivers. Once that happens, the robots could swarm and hook your "trailer" to a "train" and duck out to handle another more local task, communicating ahead to have a tractor at the destination ready to peel you out of the "train" and handle the last mile.
As for the personalization part, the consumer would purchase their "body" or rent/lease one. Of course, some consumers might simply choose to outright buy a full car or truck. Whatever. But there could be clear incentives to join the robot collective, such as tractor-sharing to reduce costs, tractor-service (maybe Walmart provides better upkeep/maintenance, more available tractors, less wait time), and more flexibility in body changing (consumers could rent a tricked-out "digital officespace" coupe for the work week and a SUV-looking jalopy on the weekends--whereby that same SUV may be used by another single mom of 3 during the week (tho I hope she takes time to clean out the fries stuck down between the seats)).
Sometimes history has an odd way of repeating itself: we've gone from carriages to train cars, to trucks, to cars, and I think we're now set to go backwards. There just will be robots instead of horses and rails.
people in the same family buy cars not only to drive but also so they can say that they possess it or to show off .......whatever
BTW how can you stop driverless cars from crashing into each other????????
People don't want to wait, nor sit in someone else's mess. It's going to be a long time before vehicles are shared. Do you want to share a car with a family that has 4 kids in car seats who spill crap everywhere? While the usage of a vehicle could go up, it doesn't mean it will. Planes carry lots of people, cares carry few. They also travel great distances vs. small. There's lots of problems with this argument. The automotive companies will build these cars and people will buy them if there's proven value.
I agree on some points of this, but I think that people that can afford their own cars, will get their own. I share a car now and it sucks, you are stranded, and have to adjust your times for everyone, its just terrible. Also, I don't know if technology is advanced enough yet to handle driving on its own, let alone driving unattended. What about hazardous road conditions like potholes, jay-walking pedestrians, flashing red lights, other drivers making mistakes etc? God forbid something like a flat tire would happen or the car starts leaking fluid when no one is in there. If something goes wrong, no one is there to fix it.
It will free capital for more productive use! #win.
Since you are using the planes-are-expensive analogy, why not apply this tech first to public transportation here in the US? When I go to other countries (e.g. HK, Taiwan), you can go everywhere easily because the trains and buses comes every 10-15 minutes. Buses (and trains) are expensive, just like planes. So, what if you can lower the costs of the buses (minibus) and run them in a loop all the time? Think what you can do if you have automated "Google" minibuses, with "Google wifi" of course, that runs every 10 minutes during business hours. Wouldn't that take the burden off having to have a car for every person going to work because they can have more leeway over current transportation infrastructure? Wouldn't this also decrease costs of running the buses in the first place, and lower fares at the same time, which would increase ridership?
I fully expect my grandkids to be absolutely shocked that cars were driven by people. "That would be so dangerous!"
Something else to consider, especially if cars go communal, is the effect on the parking industry. Not to mention all the land that could be reclaimed cause the need for large plots won't be necessary!
I thought they already had a driverless car. Its called a train.
Techcrunch thinks you work for google now. congrats on the new job. :P
Watch what happens after the first pedestrian gets struck by an autonomous car. I'm impressed by Thrun's work, but am not so sure they've thought through all the non-technical aspects of this technology.
Interesting idea, the best part of it all, nothing will go to waste.
with the price of gas, no I'm not going to send a car back 40 miles to run someone to the store. They can wait till I get home or buy their own car
Its part and parcel of any technology advancement. Auto industry has already gone through this phase when automation came in..
I think that driverless cars could revolutionize taxi and car rental services the way Netflix and Redbox revolutionized video rental. Cheap video rental wasn't possible with VHS, but DVDs enabled them.

Similarly, driverless cars could enable cheap on-demand car rental. I call this hybrid service taxicar. The cheapest taxicar would be a small single person electric with a short range. This isn't possible with current taxis, because of the driver.

For most trips and commutes, you would use a single person taxicar. Even if you want to travel further than its battery range, you would simply hop from one taxicar to another as necessary for the desired range. You don't have to wait for the taxicar to recharge.

There would also be demand for two person and four person taxicars, as well as longer ranged taxicars. These would be good for commuters who can carpool, but it's infinitely more flexible than carpooling because everyone still has the option to use single person taxicars whenever desired. The point is--car size and range would match requirements, on demand, rather than everyone being forced to buy cars which match their peak needs.

None of this is to suggest that automakers would be in trouble. Cars don't really last that long, so these driverless cars would still need to be purchased and replaced at roughly the same rate as before. The miles per person could even increase by some degree as there's some overhead in zero passenger cars traveling to/from the customers.
yep.... as time marches on so does technology, and these days particularly computing. Driverless cars will happen and probably faster than anyone expects. Imagine driverless taxis you could summon using Siri and Google Maps! No driver = less cost, = lower fares = more usage = few privately owned cars. This could change cities forever.
I would amend the concept as follows. Only cars with passengers would need to travel at high speed. Empty cars could be relocated to where they are needed using slow speed conveyor belts.
All interesting points, and I for one look forward to welcoming our future automated automotive overlords.

I gave up driving when I moved to London 5 years ago, and don't miss it one bit.
Of course, using public transport in London, one of the first things that springs to mind is that if everyone's using driverless cars, surely we're going to see an increase in people throwing themselves in front of them? Suicide by train is quite popular over here unfortunately, and I'm guessing we'll see similar behaviour on our roads if all cars are automated. I wonder if Google's engineers have accounted for this inevitability?

But otherwise, it can't come soon enough for me. Being able to roll out of bed, into a robot car and continue sleeping without fear of being robbed while being taken to work sounds like bliss.

The other thing that this gets me thinking about is driverless trains. Surely this is an easier technological goal than cars? If so, why hasn't it been done? There must be plenty of studies that have looked into doing this and decided that retaining the human factor is the safest option. Why?
I like the way your mind works! But I'm sure, as mony is still the only currency that makes this world do anything, the auto industry will stop this perfect idea in it's tracks. Because they would sell much less cars and - come hell or high water - thats what they want to keep doing: making profit.
In my home town of Helsinki, traffic isn't really a problem. There are a few places where, for 25 minutes in the mornings and 30 minutes in the afternoons, traffic slows to an average of 15km/h rather than 30. However, downtown parking IS a problem. I've long conjectured, that if only 10% less people wanted or needed to park in downtown, there would be no parking problem. Where you now have to drive around a couple of blocks to find a space, there would be 5 free spaces in every block. That's more than enough. This is the nature of bottlenecking issues - you don't have any, until you do, and then it's bad immediately.

Most downtown occupants don't own more than one car, and many don't own any. But if car sharing was practical, because cars would be easily available, many, many more people would turn to car sharing, I'm sure. In-city car ownership would turn to a relic, and businesses like Zipcar, or the local City Car Club would boom.

And interestingly, this would only need a limited form of driverless cars. Regulation could still require that a car with a passenger needs to have a human driver - as long as cars could drive themselves (empty) to the next nearest user, parking would be solved immediately. In many municipalities, I'm sure these driverless, passengerless cars could even be limited to only certain routes, further limiting the possibilities that a human driver would need to find themselves in a traffic situation with a robot. All of this would likely make it easier for regulation and communities adjust to the idea that a car can move autonomously, and for the tech to further mature.
Thanks for the post! Oh - it will become a wonderful world. I'm really confident the software in the car will steer it as good (or as bad) as the current average human driver (Patching and updating them is by far harder, haha). And, hey, no need for the cars to honk the horn! Did you ever hear the TCP/IP protocol honk during handshake? What a silent world. And when it's time for winter tires I send the car alone to the mechanic during my working hours. (Jokes aside, I REALLY like this).
+Isaac Kuo A one person car is a bad idea unless it's a motorcycle. Adding space for one or three other people doesn't eat up very much in terms of fuel or space efficiency (even in an electric). Since your idea is only workable in the city, I'm not sure what advantage it has over a bus on busy routes. The bus makes up for its poor fuel economy (a few mpg) by the fact it can carry 60-90 people. On sparse or less active routes, I get it, but I'm not sure electric buys anyone anything, at least in the current state of the technology. You'd need a lot of places to park the cars so they can charge and to cover the downtime of your fleet. Hybrids are honestly a better option if you were going to build a solution like this.

Cars also last a long time these days. Increased emissions and safety standards have forced cars to generally be pretty reliable across the board; it's hardly unreasonable to expect most cars to last well over 100k miles. Plenty of new cars don't even need any standard maintenance besides tires, oil, wipers, and the occasional bath before 100k miles. At 15k/yr, that means /7 years/ before you have to sink any serious money of any sort into maintenance (unless your tires are 180/wheel ;_;). This is a considerable part of the reason that the used car market has done so well.
+Mat Linnett Several train and subway support an Automatic Train Operation (ATO) where the train is controlled remotely. I believe the trend is slowly heading in that direction. Regardless, most (all?) ATO lines still keep operators on board because their salary is very small potatoes in the grand scheme of operating a railway line, especially when compared with the cost of an accident because the ATO equipment failed and there was no one on board to operate the train.

Again, trying to take the driver out of the equation is simply optimizing the wrong problem. The inefficiency in car transportation comes from supporting all the concurrent routes (unique origins and destinations that must be serviced simultaneously) efficiently. It's nice to believe you can just change times around freely or that you can merge routes together heavily, but you really can't. You probably can do better than today, but you won't get 96% utilization or even 50% utilization. Everything else is small potatoes: removing the driver does little (he/she was probably going there anyway or there are too many passengers to matter), changing the size of the vehicle is a small gain, even being able to shut the vehicle off when not in use isn't all that big.
Too many comments to see if the other obvious benefit is mentioned: capturing billions of hours of advertisement-viewing time!
I didn't read through all of the comments, but isn't there a key piece here that is being ignored? Traffic congestion would be WORSE and we would need to build MORE travel lanes. Why? Because right now when I drive to work it is a single round trip (home->work->home). Now image the trip (at a minimum) needed for a driverless car: previous destination -> home -> work -> new destination -> work -> home -> next destination. That doesn't even include any miles traveled in the "new destination" trips in the middle of the day. If people are still riding alone in a driverless car, then the amount of miles driven per passenger carried will go through the roof. Rush hour all day long, and just imagine the total gridlock in L.A. if the current traffic was replaced with the scenario I listed above. The only real solution to reduce traffic congestion is mass transit.

And a somewhat more practical question: Would the cars be cleaned after each trip? Imagine a car arrives and the previous user had spilled something all over the seats. Wouldn't it suck to miss your kids soccer game because you had to wait for another car? At least on a bus or a train I have the option to look for a different seat or stand.
The traffic won't be a problem thank to computer moderation, but it's terrible for car makers and taxis!
+Anthony Prattico The practical assumption is the routes are divergent enough to not create continuous rush hour. I see no real reason to believe that's not the case. Depending on the scenarios, rush hour might end up extended, however.
You are merely scratching the surface of what could happen. Schemes like city cars would become much more popular, especially as you wouldn't have all the cars ending up in city centres as they could drive back to the outskirts. Car ownership would drop massively but some would still want to buy cars to drive for enjoyment.
Every taxi driver could be out of a job. Same goes for lorry drivers, bus drivers and driving instructors. Would you even need buses? You could just pay for a journey in a driverless car. In my own little utopia this could be similar to the bike schemes operated in Paris and London. The car itself would totally transform. The current interior layout could be totally different. The family run about could have seats facing each other rather than pointing in the same direction. The effect on the emergency services could be positive. Roads filled with automated drivers would be able to move over to allow the services to pass through at high speed without danger to other road users. With no limit on the age of drivers young people could be freed from the limits of where their parents will take them. Drink driving would become a thing of the past. Everyone could have a lie in as they eat their breakfast, do their hair put their makeup on as they "drive" into work. Interesting stuff!
Wish I could read all the comments but too many. Some of them are really interesting. As +Adam Skutt pointed out, there are a couple of things to figure out to make this happen. First, this car needs to be folded or do whatever to make it smaller so that it will take up less space on the road and consume less energy. Secondly, it seriously needs to be energy-efficient otherwise it doesn't pay off the cost of another car. Thirdly, it needs to travel on a different route like underground or something like that when there's no human passengers so that it doesn't add to the traffic congestion. And I'm sure there are many more things to solve before this becomes our reality.

At any rate, as +Chris Grey said, it will be interesting to see how this thing will change our lives in the future.
The insurance industry will find a way to keep charging is a San Francisco peer to peer car sharing service that is starting to take off. It is in an interim step towards the main idea in the post. Shared cars, very close to where you live, bookable via your phone. Very convenient.
I like the idea of GetAround but it would work better if it wasn't "my car" other people we're getting around in. Who is going to get a car rental company to step in and make this happen? They have bulk insurance, fleet pricing, maintenance, etc. Condo Car Share seems like the best perk ever! Online scheduling, small Fleet Nissan Leafs (insurance, etc), maintenance - mostly cleaning out irresponsible user's junk, replacing removable seat covers, washing windows, checking tires...add to this a back-up plan using Groundlink, Uber or the like. Come on you mathematicians - there's got to be a user : cars : urban density : travel requirements golden ratio.
We need to think a bit more saner / wholistic and far-out , Guys , why does one need to "GO" anywhere .. very primitive in my opinion. Need to be like the really-wise-eastern-ancients .. Let things come to You .. Use Tech to do stuff from "home" -- immersive tele-presence etc No point "going" somewhere for economic reasons. .. Other than for educational/recreational purposes ..see no real reason for "people" to move from their base - not in the least for "jobs" etc. - most jobs today require f2f interactions (checks and balances , modern management theory etc. ) because of all the inherent ambiguities associated with human weaknesses and frailties ( greed, deceit , dishonesty , egos etc ) - All this Tech has its place - we need to focus on the basic human stuff first .. New solutions will automagically emerge .
Is it too early to start a Hangout to think up the malevolent things that can be done when transportation goes AI? Every digital technology that is developed with the expectation that people can be trusted and that they will act with common, public-spirited good sense has become fertile soil for exploitation and crime too. The internet is flooded with traffic designed to con people into something. The security companies can't protect our PCs and the US and Israel have already committed acts of cyberwarfare.

So for all the taxi-drivers, truckers, bus drivers, UPS guys, pizza delivery boys, and auto-workers who are going to have to find a new livelihood I think it's only fair to offer a forum where they can get the jump on The Dark Side of AI, starting with driverless cars. What will it take to hack driverless cars and trucks and turn them to other purposes and exploit their users? Time to start thinking innovatively about coming opportunities in the underground economy! If hacking and cracking are defeating security professionals now, just think what AI in the hands of ne're-do-wells will offer?
As long as there is an override or interactive mode, I'm OK. If the cloud system goes down, or the car loses it's connection, having it just stop could prove troublesome.
It will be all roses until someone figures out how to use an unpiloted car to kill someone intentionally and then blames it on "malfunction".
+Koushik Dutta Your creative thinking is wonderful. Have you considered the effects of driverless cars and car sharing on other forms of transportation, or other transportation needs? For example, would some people choose a driverless car trip over a short-haul airplane flight? How would driverless trucks for shipping change things? Finally, what about pilot-less planes or captain-less boats?
Basically correct but you've overlooked the availability factor. Most people need their cars at specific times - such as when it's time to go to work in the morning. That means my neighbor and I both need my car at the same time. We don't carpool now because we work in completely different directions. Therefore, driverless or not, we both need cars.

Not everyone lives in cities, you know.
96% of it's time idle? I would like to see some numbers to back that up. I live in the most congested metro in the US (DC), and I have been using my Mercede's trip computer to monitor average speed over the last 15000 miles. I'm averaging about 37MPH, which would indicate that I'm usually well above 5MPH.
+Evan Pols I think he mean "not in use" by idle. If you commute for two hours total and use your car for nothing else, then you use your car for 2/24 hours = 1/12 = 0.083. A person who commutes an hour total therefore is at 4%. Even if we change a day to be something like 12 hours instead of twenty-four, it means a person who commutes an half-hour is therefore at 4% which is somewhere pretty close to the stated average commute time by various places (15 minutes one-way).
Just think what it will mean for all the parking lots out there. Sudden increase in useful real estate. Also, streets will get smaller for more pedestrian/bike use/parks.
I wrote a some more about what I think driverless cars will do to parking problems at my blog at,-and-a-complete-rethink-of-in-city-traffic-management. With regards to the question of number of cars sold vs time-to-reach-100k vs amount of maintenance needed, you need to look no further than taxis. 2-5 person cars will never reach the utilization rates of airplanes, because they don't cover the same distances, and thus will not need 24/7 operations, but they can easily reach the utilization rates of a taxi shared by 2-3 drivers.

I don't have any data to back me up, but I believe the average lifetime of those vehicles (at least before being resold for some other use) would be somewhere around 3-4 years, perhaps a year less in municipalities where the city or its taxi association wants to maintain a high standard. One automatic car will be able to serve 2-4 rush-hour customers, more if it's combined with carpooling. The assumption that 20x less cars would be sold is perhaps a bit optimistic, though I can certainly buy the idea that 20x less would be parked anywhere you now see a lot of parked cars.
+Osma Ahvenlampi If I can't get utilization to 96%, how do I reduce parking by 20x? These magic driverless cars can also convert themselves to pure energy and cease taking up space?
Sounds like it will be the end of taxi drivers. But ultimately, It will probably be one of those things a lot of people will go against for a lot of valid reasons... in vain.. since you can't really stop the future. Society will just take a little time to adapt. We now have robots and machines handling cleaning, checkouts, and an increasing number of things. I guess taxi drivers will just get replaced little by little just as cashiers are now.
+Damien Vaillant Actually those self-checkout registers are being removed in a lot of grocery stores: they don't really save money over cashiers and lots of people dislike them. The notional advantage is checkout is faster with them, but that turns out to be false for a variety of reasons (mostly lack of trust in the customers using them and bad machines).
It wouldn't happen. At least not in near future. Guys like me (car geeks/petrolheads) will not buy these cars. I like to drive a car (and a lot of my friends are the same) and enjoy driving a car.
"Cars wich drives you to any place you ask were invented in early 20th century. They're called taxi" (c) James May from Top Gear.
A car with autopilot is like a non-alcoholic beer.
+Adam Skutt you don't need to reduce parking nor busy routes traffic by more than 20-50% to eliminate the problem. Congestion is a bottleneck problem, and has a very clear inflection point between works fine/doesn't work at all. Besides, during hours of lower utilization, automatic fleets can divert excess capacity to neighboring areas with more room.
+Osma Ahvenlampi Reducing congestion does nothing to change the fact I still have to have places to park cars when they're not in use! Local parking even goes down at a slower rate than the global need, since local peaks can and will persist well above the global average. As an example, consider a sporting event: the area immediately surrounding the stadium doesn't need all those cars, so you're forced the send them on increasingly distant routes to improve utilization. At some point, this breaks down and you cannot get the cars back to the stadium in time to ensure they can return home. As such, a considerable number of cars get stuck at the stadium with nothing to do. Could the amount of parking go down? Yes. Will it go down 20x? No.

You have the same problem with Holiday shopping at the mall: even if the parking is totally empty 300 days of the year, you still need the space for the excess cars the other 65 days. There just aren't enough routes locally to cover all the vehicles parked at the mall, and not enough time to send the cars to service further away routes and expect sufficient coverage for when people leave the mall. Shopping is particularly problematic since the car may have to remain parked anyway. If I go to the store and buy a shopping cart's full of goods, then that car has to remain with me until I return home and put the goods away. Expecting me to take my previous purchases with me is unreasonable for many reasons, likewise so is expecting me to trust random strangers with my goods in the trunk. The car gets stuck regardless until I go home.

So, we're left with my original conclusion: your expectation is that driverless cars can cease taking up space.
I read the beginning of this thread on Friday, and it left me with a nagging concern that I couldn't quite put my finger on until this morning. Basically, it's that pretty much everyone on this thread is a technologist or a technocrat - we're all confortable with Vorsprung durch Technik because we know enough about the underlying tech to figure out what to do if it's not functioning as expected. The several interconnected layers of tech necessary to make first self driving and then network federated cars a reality are expected progress to us.

My concern is that the superset of the population that has no idea how any of this works - they just expect, that like dialtone, it'll be there when they need it. That's partially mitigated by us; when our less tech-savvy friends escalate to us, or when they Google something to get the info needed to address the issue themselves. But as we add layer upon layer of complexity to already complex systems, many people lack the basic deductive or inductive logic to figure out any semi-complex issue.

An extreme example: a friend of mine works in a Lexus dealer's service triage desk - she got a call from a new owner who was panicking because some lights were already regularly blinking on the dash. She'd inadvertently activated the hazard lights, and didn't even know what hazard lights were. She hadn't even left the lot yet.

I think we also take for granted that the next generation will be more comfortable with these layers, and to some degree they are. They are certainly more comfortable using them, but few really understad the back end systems and how they work. Hence more throwaway tech. I feel these appliance cars will fall into that category of "too complicated, just get another module/drivetrain/car" because only a few will understand how to fix them, and those will be some combination of us/mechanic. This is already true to some degree, but I am concerned that we're not concerned about it as a society.

I guess what I'm saying is that we should be aware that there's a potential for class divide here, along the sysadmin vs. user lines, but at a much greater scale. And the less people understand about tech and what we do, the more we seem like curmudgeonly sorcerers, who refuse to fix the problems others want resolved, because they have no idea of the complexity of what they ask.

reminds me of a conversation about complex software problems related to the inability to finish the nextgen FAA systems:

and why we use old tech when the mission is critical:
I'm in favor of this as long as we can get "the right to own and operate a car for recreation" added to the Bill of Rights. Seems like driving for fun may get pegged as dangerous and sometimes fatal; which it is, but so is all the other fun stuff in the world.
Driverless cars can be used right now in controlled sectors like farming, agriculture, parking, and add that to the usage pattern of today's cars and an increasing population - yep we need more utilization and optimized usage of the roadway systems #aiclass
Driverless cars are unavoidable, nobody can stop the future to come.
What happens when cops find drugs in the car? Can you claim it was just left there from the last guy who used your communally-shared driverless car?
Neil S
Kottke pointed me here. As previously mentioned, Brad Templeton's done a lot of thinking through this.

Here's my little presentation / place-to-store-thoughts (it tapers out towards the end) I was plotting to pitch this to City of Vancouver (who are spoiled with Modo, Zipcar and Car2Go - none of which are electric, and none of which come to you).

But I now think the taxi and bus-driver unions will stop it ever happening.
It would be like Knight Rider! > I will be able to call the car from my phone
+Matthew Jacobs nothing ye olde good CCTV cameras can't deal with. I imagine that in that kind of service they would become a must
I expect insurance to really push the conversion. When it costs twenty times as much to insure a manual car as it does for a autocar, any extra expense in paying for the autocars extensive tech will be offset.

Consider what it will do to infrastructure. If the automobile fleet goes 100% driverless, why would it need traffic signals, road markers, speed limit and hazard signs. Highway construction would drop off because the driverless cars would be able to pack in more tightly.

Distracted driving would become a plus! Watch the news during your commute. Watch a movie on a road trip. Of course, your boss would want you to work while you're commuting between customers. Think of how a car would be re-designed if the driver no longer needs a cockpit. Perhaps there would be "long-haul" varieties designed to allow overnight trips. Go to bed in LA, wake up in Denver.

Ooh. And speaking of work... a lot of truck drivers would be out of business, not to mention taxi drivers and couriers (no, the unions won't stop this). Trash pickup, too. Ya don't need so many cops to man those speed traps anymore, and a lot of small towns would need to rethink their tax base. Fewer firefighters and ambulances would be required, as there would be fewer car accidents. Fewer trauma centers. Fewer rehab centers. Life expectancy would see a boost.

How would parking be handled if you can go to the market and tell your car to orbit? Or go park a half-mile away and return to the curbside to load up your groceries.

How would motorcycles fit? Would anyone manually driving become viewed as a daredevil? Reckless, even?

Look at your current car and think to yourself, "this is a mule and the automobile is coming."
I think families would still want to own the 'entire' car for maximum flexibility/control. Taxis are supposed to work the same way, and so being driverless wouldn't make a difference.
+James Brown Traffic control is still necessary because you can't avoid intersecting paths, and you can't avoid traffic prioritization (e.g., EMS and police). Control devices may not take the same form as they do today, but they still will exist in spades. This means changing our infrastructure, which means a lot of heavy construction. Construction may or may not drop off, but it won't go away since most of it is to simply maintain and repair road damage that occurs simply because cars drive on them. Signage is a pretty small part of the overall infrastructure effort.

I don't see how trash pickup stops requiring humans unless we build trash collecting robots as well. I doubt in most areas the amount of police and EMS required will drop significantly even if the traffic accident rate went to zero (since it's about timeliness of coverage, not quantity of coverage--you can't/don't want to move the stations further away). Again, population density is a real killer and one of the real problems that needs solving.
A private car is not the same thing as a bus. Its a car and drives you straight to your home non-stop.

The future is here, it's unavoidable.

People will still keep their cars but the car companies will not be making low-end models which will not perform well through the long haul.

I think it's brilliant.
Add a comment...