do we really need self-driving cars?

Self-driving cars won't solve congestion in major cities. They will be the congestion in major cities.

For as long as we've envisioned the future, flying cars were part of it. We've been trying for create them for decades and there are some prototypes that are basically cars converted into light helicopters or planes. But as awesome as they look in the movies, flying cars would be much riskier than everyday driving and once millions of people are using the same air routes, landing pads, and hangars to get to their destinations, it will get old quick. Imagine multi-layer traffic jams of hovering helicopters trying to avoid each other's turbulence, and that's what LA, NYC, and the entire Bay Area would look like with flying cars in the real world. And to complete the mental image, it would be loud as hell because rotors and jet engines make a lot of noise.

Now, should we ever come up with the technology to float our cars in the air with nothing more than a low puttering sound as they pass by, like the cars from The Jetsons, they could solve the problem of not having enough roads in a dense hyper-city of the future, and effectively stack traffic in layers while requiring very little additional infrastructure. As reliant on technology we're not sure is even possible as they are, they do represent a solution to a very real issue. Self-driving cars, however, seem like more of a solution that's in need of a problem . We have a good idea how to build them and a decent rundown of the limitations that need to be addressed before they're widely deployed, but even if we get all the tech right, what will we gain?

Don't get me wrong, an autonomous car would have its uses. It can turn a long, boring commute into a productive or relaxing one. It could transport goods without needing anything more than the occasional pit stop without a human driver, cutting transportation costs by a healthy margin. It could talk to other cars and working together with them pack traffic more efficiently to speed up the overall flow of vehicles. However, if you take to the streets of, say LA, you'll rather quickly notice that even if you eliminated much of the so-called shockwave traffic, and greatly reduced accidents, the biggest and most pressing problem is that there are too many cars on the same road at the same time, and there's little evidence that just building more roads to accommodate more cars helps on a macro scale.

Aside from the countless security concerns that will come from having to allow autonomous cars to talk to each other, all we're more than likely to get in the cities that will benefit most from them is more efficient traffic jams. But at least you can watch your favorite movie while you're stuck in them as a consolation prize of sorts. So we're going to go through all the trouble of securing autonomous cars, figuring out their algorithms when problems pop up, train humans how to pounce and recover from catastrophic failures, and pass legislation to smooth their introduction to… what exactly? The problem here is that people, for economic reasons, are moving into large, dense cities more and more, where we need to figure out how to have fewer cars on the road, not more with enough LIDARs, 3D maps, and automated steering. But hold on, why are people moving? Because small towns are dying and cities are more and more where the jobs are.

The numbers don't lie. Just 20 cities host 52% of the American GDP, so while throngs of pundits and politicians deride cities as socialist citadels of crime, drugs, sin, and sex, while the real America consists of Norman Rockwell-esque small outposts, today's economy is driving ever more people into megacities to shorten supply lines for business hubs, and boost efficiency. For now, it's cheaper to build more roads to try and keep up with the flow of new residents' cars, but as I just noted, that's a losing strategy over the long term. Ultimately, city planners will be worried about efficiency, and your fairly typical two person, two car, one driver per car household is extremely inefficient when it comes to moving people around. Simply making the cars autonomous doesn't solve this, but a concept like the Next Future Transport is a step in the right direction.

Imagine these little, on-demand cubes that drive themselves and interlock into useful shapes for the most efficient people routing, but on their own, purpose built magnetic tracks, functioning like a just-in-time train system. Their large, container sized, industrial big siblings could be navigating the inter-city highways delivering goods across continents, also in specialized lanes. In the meantime, while cars will be a necessity in rural areas, in the city, they will be a luxury used every once in a while while your typical, day to day commute will take a Next-like system to and from work, and to run your errands. For sacrificing a little of your personal space once in a while, you would be able to get around a giant megacity quickly and efficiently, and you'll save thousands in car and insurance payments per year.

Here's the bottom line. Suburban sprawl is extremely expensive and in the economy of the future, driven by automation, consolidation of supply lines, and increasingly centralized in hubs, it's too costly to maintain as anything other than a luxury. As much as we may not want to admit it, the future is on demand automated public transport, not self-driving cars that entertain us through our ever longer commute on yet another highway. At some point it will become too costly to keep sprawling and building more highways, and the suburbs will become an endangered species. Building robots to navigate the tedious, daily monotony of that sprawl in one of the most inefficient forms of transport we have, seems like a waste of time since the technology would be best put to use to something simpler and more long term.