The Future Of Cars
2nd August 2016

In this post, I muse about the future of cars – and with it, houses, cities and work. With the way things are interconnected, changing how cars function has big knock-on effects for other aspects of our societal life too.


In the relatively immediate future – I reckon about fifteen years from now – advances in car technology will have radically changed the society we live in (of course, some changes will happen sooner and others will carry on going). The changes I’m predicting here come from technology which already exists but has not yet developed enough, so perhaps further technological advances will change this even more. Or, for many other reasons, it might be different from this prediction: this post is meant to paint a picture of a potential world that is coming, I cannot actually predict the future that well.

We start with the self-driving car. These are pretty much ready to go, have already been driving around a lot and are about as safe as people-drivers. The technology is still improving, but that will be to get it to a point where they are almost flawless. But we are way beyond the prototype stage: this is not an ‘if’ but a when.

Before too long, all cars will be self-driving and human driving will be prohibited. This might take some time, as it’s difficult to overcome the human social intertia, but once people get used to having robot chauffers it’ll happen. It would be much more dangerous to allow human drivers alongside the autonomous cars, as the computer-driven ones will be much more predictable and able to interact with each other.

This is pretty cool in itself, but there’s much more to it than just ‘oh cool, robot chauffers’.


One change will be a shift – at least partial or for the majority of people – to sharing cars, instead of owning cars. There are a few parts to this.

Firstly, it’s a huge waste of resources to have a car parked doing nothing for ages: somebody drives to work at 0830 and leaves again at 1700, and in the time in the middle the car is redundant. Now that the car can drive itself, it can leave again. This doesn’t just solve parking problems: it also allows the car to go and drive other people around. Your car ownership also works as a personal taxi company, as your car spends the six hours it would have been in the office carpark earning money as a taxi.

Secondly, for those who aren’t the car owners, it will make more sense to not buy a car and to instead always get lifts from unused self-driving cars as it will be cheaper. The same way it’s cheaper to get taxis (or ‘ubers’, which are a different form of taxi) than to own a car if you don’t use a car much, but the trade-off will be much more in favour of getting the taxi instead of driving a car.

Thirdly, there is a greater possibility for carpooling. People might not like it because we’re all anti-social and want to avoid other humans as much as possible, but it will be much cheaper and function more like a shared small public transport carriage than a car. I’ll return to this point below with the ‘botnet’ discussion.

As well as individuals who own cars (or buy-to-own cars) to then rent them out, there will also be car-fleet-companies who buy cars just to rent them out. This should function somewhat like the actual market model used by economists, where supply and demand shift with prices and people buy cars or not depending on which is more price effective…

These changes will become possible because of the development of self-driving cars and because of the development of ‘artificial intelligence’. Speaking properly, it is not actually intelligence, it can’t think in the way that we actually think, but it can process really well; unfortunately, I don’t know a better term, so I’ll continue to call it ‘AI’ but with the quote marks to designate it isn’t really it.


The self-driving car is only the start; on it’s own, it would just be like having everybody with a chauffer, which wouldn’t change much – perhaps a bit cheaper, but not radically so, and with fewer possibilities. The bigger impact comes from what I’m going to call the CarNet: the ‘car network’ or ‘internet of cars’: all the cars will be communicating with each other, and also with some forms of ‘master intelligence’ run by one or many computer programmes. This will be sort-of like a military commander or the way an app like Uber works: it co-ordinates between the cars and the users to solve various problems, such as which car drives where, how the cars interact, what speed they go, and crucially, which people go in which cars.

This is a similar sort-of job to the central hub for a taxi company: people phone a central number to say ‘I want to go from X to Y’, the hub tells a particular driver to go and collect the person, etc. Except on a much large scale and dealing with much more complexity.

So, big change number one: the CarNet won’t need to be controlled by humans. This would have been hard to understand a short while ago, but now that we have apps like Uber we can see how it works. With Uber, the app provides the interface between customer-passengers and drivers; of course, the driver will be replaced by a self-driving car.

The CarNet will then decide which car should come and collect you, looking at the available pool of cars, which ones are free, what sort of capacity might be needed later, and which cars are doing similar journeys to pool together.

This is also where carpooling comes back in: the current limit to effective carpooling is the logistics of linking people up, which is solved by having an AI that can do it for you. You don’t know what the nearest 1000 people to you need cars for, so you can’t work out a good carpooling system, but the AI system will be able to do it easily enough. The user-passenger books in their journey to work, the CarNet can group people doing similar journeys together.It may well be more efficient to have a passanger change cars partway through their journey and put them in a car going to a closer destination. So the botnet of autonomous cars becomes a huge public transport network.

It will also control how the cars interact with each other: each car has its computer-brain as well as the overall CarNet. This will control all the usual driving things (braking, turning, etc) that humans do, but also be able to change more. As it will be controlled, the rules of the road can be changed too: a simple example would be traffic light timings, as one direction is busier than others they get a longer time for cars to go through. Really though, there will be no need for traffic lights: those are to tell humans what to do en masse, whereas the CarNet will be able to tell each car what to do. At some times of day there might be a one-way system, it might make sense to send huge swarms of cars together instead of individuals, or lots of other things I haven’t thought of.

This will hugely improve traffic flow: if you can eliminate human error and introduce a hive mentality, cars are able to travel much more quickly and interact far better with each other. There’s no need to leave much braking space if you know that nobody is going to brake! Cars won’t just have to watch each other via sensors to work out how closely to travel: they will be communicating, matching speeds and making space for each other to be more efficient overall. It’ll all be much more efficient and smoother rides without the stop-start of human driving, similar to what the variable speed limits on motorways (where it slows you down to 60/50/40 when it gets busy) but much better.

Kind of like this: 

RUSH HOUR from Black Sheep Films on Vimeo.


So cars will be much more efficient – in terms of journey time and energy used – for a number of reasons. Computers are much better drivers than people and the CarNet efficiencies will mean traffic flows with much less need to stop-start, which will make it quicker and mean less fuel (or energ) is needed. Carpooling will be much easier to do, which split prices. The greater utilisation of cars – in terms of making use of them instead of having them sat around doing nothing – will also help. It will also aid the shift to electric cars, as the cars can go and recharge themselves at the right time – and if you are doing a long journey, you can change car part-way. Electric cars should have lower operating costs too.

(Caveat: this depends on how how the market is regulated and functions… if it functions as a competitive market, or if it is run as a public service and not for-profit, then fine. But if it ends up being controlled by a few main companies then they can raise prices or artificially restrict supply to increase profit).


The time spent driving will be freed up for other use. People spend a lot of time commuting to work. If you are driving, you are a bit limited in what you can do – you can listen to things or have a conversation, but your hands are on the steering wheel and your focus on the road. The driver will be liberated and time now free to use.

The space in the car will also be free to use. Imagine you are sat in a black-cab (or in the rear seat of a bigger-than-average car). Remove the driver seat and partition (or the front two seats), and remove the middle seat too. There is a fair amount of space in the car.

(and remember also that with magical self-driving cars the driving is much smoother and they are much more safe, so seatbelts probably won’t be necessary)

There is enough space in the car to turn it into useful space. Two key uses come to mind: sleep-space and office-space. Time can be used for sleeping and working: commuting no longer wastes time but happens in the background while you do other things.

(You could also read, or write, or play candy crush saga, or scroll through a social media feed, or watch a film, you get the idea there’s a lot you can do. But I think the things with greater potential impact are work and sleep.)

This huge change does not come from the fact that your car is self-driving, but rather from the fact that your ‘car’ will actually be a self-driving room. For sleep-space, it’s a self-driving caravan. I’m not sure drivable offices currently exist, but there will be adequate space for a desk and chair (and we are already at a point where internet access it very quick without a landline, so you will have very good internet too).


The prediction of the self-driving-bedroom/office is a strong one, I think: it seems somewhat obvious to utilise the space and time for useful things, be that sleeping or working. The implications of this are much more guesswork: it’s a complex world with lots of interconnectedness, and nobody can understand this well enough to know with much certainty what the affect of this one change might be. So this is now even less about predictions and more about possibilities.

There is a long trend of moving into cities, especially into London. Housing costs have increased as a response and the city has expanded, with many able to live quite far from the centre while still commuting. It could be that in the future, it won’t be unusual to live about four hours drive away, with the day’s routine looking something like this:

  • Wake up at 3am, move to the car, fall asleep in the car-bed. Iit’s been said that in human sleep, we have two sleep cycles anyway, so this wouldn’t be too bad.
  • Wake up again at 7am in London, use bathroom facilities and eat breakfast and maybe even exercise or spend some time outside.
  • Begin office-based work at a usual time.
  • The morning of work would have the physical interaction things concentrated, primarily meetings.
  • Lunchtime, followed by a short amount of work.
  • Early-mid afternoon: move to an office-car for the commute home and continue work on the non-physical-interaction things of reading, writing, emails, phonecalls. As Virtual Reality technology develops, the use of physical meetings may continue to reduce (or maybe phonecalls and conference calls just get an upgrade).
  • Arrive home at 5pm, work complete. A reasonable time to arrive home.

The commute-sleep time works for all jobs, but the commute-office time only works for office based jobs. This will not apply for jobs that include physically doing things (though these are more likely to be improved by technological advances and AI in other ways).

With this possibility, a lot might change. This is based on assumptions about the cost of travel, but as I said above I think travel will become a lot cheaper. Though this is hoping for a transition to sustainable electricity generation instead of fossil-fuel based, which is somewhat optimistic at the moment… Also, if it’s possible to work while travelling, then that will cancel out the cost somewhat.

So the concept of the city should collapse somewhat, as it is no longer necessary to live so close to the place of work. It will still exist, it makes sense to have some concentrations, but the greater commuting potential should draw many out of it. Hooray! The (so-called) Housing Crisis may be averted!

Instead, I think village living will increase. I have a hunch that most people will prefer living near to nature instead of in densely urban places: it would be possible to live by the beach but still work in a city! Especially when many of the services we need will become more portable too: instead of visiting shops, home delivery will be even easier than it already is (partly due to self-driving cars again!).

This should throw the social life up in the air too, because with travel time effectively removed (by being turned into useful time) and self-driving caravans, it’s much easier to end up in the same place as your friends or family. This too could be organised by an ‘AI’ personal assistant, you just tell it ‘this month I want to spend 1 week with this person and a few days with that person and two weeks in this location’ and it can sort out what’s the best way of doing this. Add on top of that the way that housing can be rented (Air BNB has shaken up the housing market a bit, for example), and that house ownership may go in a similar way to car ownership described above, and we may end up with a population of nomadic workers who take frequent ‘holidays’ that still allow them to work…

So that’s something that might be nice.


As I say, just a possibility, not even a prediction: there are many other variables. This ‘vision’ has only changed car technology, it hasn’t taken into account many other possibilities. Some of these are good: the best path forward I can see includes greater robotisation of work, freeing humans up for more leisure and requiring less time to work, perhaps a 25-hour work-week, and a universal basic income to even this out. But perhaps we don’t manage any of this, we might have huge global conflicts that interrupt peaceful utopian development, perhaps disruption becomes very easy (hacking or otherwise) and this becomes infeasible…

More about the path of future development and whether technology will be used to our general benefit in a later post…

(It would also be foolish not to mention how fragile this would be, and fragile can be really dangerous. When the transport network of a city depends on a few bridges, if those bridges get blocked then the whole city cloggs up; when so much of our life depends upon electricity, when we lose electricity life because very difficult. If there is a mistake somewhere in the code, this would cause huge problems, much more than when the tubes aren’t running in London! Similarly, it would have to be very secure from hacking or from interference from unexpected things: cars can only respond to what they are programmed or self-taught to respond to, so anything outside that could be a huge danger.)