cars in the future

I will warn everyone ahead of time that this post is basically my prediction for the future of transportation. I can see that interesting some people and severely boring others. If you are bored by speculation regarding the future of cars, then don't bother reading this post. Don't you wish everything I wrote came with this disclaimer?

Everyone knows that Toyota has been in the news lately for problems with breaking. For people who own a Toyota vehicle this must be a bit nerve-wracking. For some people adamantly against the idea of buying a car assembled in the United States by a non-unionized Japanese company, I think there is a bit of restrained glee (I decided not to use the word "schadenfreude" because the word seems pretentious). While a lot of people have gotten the idea that this proves the reputation of Toyota's manufacturing excellence a farce, I have to disagree.

Since the issues sound like they are due to very quirky behaviors with the software, this has very little to do with Toyota's ability to assemble a quality car. Software development is a whole different world from car development. Because of this, situations like this one with Toyota and future similar incidents are almost inevitable. As long as we rely on large amounts of software to automate functionality in our vehicles, and we will more and more in future years for economic reasons, inevitable defects will appear. Just due to random chance, some will impact important functions such as the car's ability to stop. This is just speculation, but it sounds like automotive software issues are more difficult to diagnose and fix as well because Toyota sounds to me like it is stalling until the software issue is identified and addressed.

As someone who has spent several years devoted to supporting software and interacting with the support staffs of other software companies I feel I have some qualification to compare expected standards for software compared to automotives. Software is typically held to a lower standard. For example, if a defect makes your browser crash and you have to reboot that is irritating, but it is not entirely unexpected. If your car steering fails to function one day while you are on the highway because of a defect that is much more unexpected and serious. While a Windows defect could cause Microsoft customers some pain a defect in a Prius could cost a customer his or her life.

As an example of what I am talking about, part of the reason that the things NASA builds are so expensive is that there is almost no tolerance for defects. It is very difficult to fix something that is millions of miles away, so it has to be nearly perfect. Even with these standards, NASA has had many serious failures over the years with both hardware and software, and I would argue that as stupid as many of the defects were, having them was quite nearly unavoidable. It is likewise unavoidable that some software glitch is going to cause some serious automotive issues that take lives.

I have considered this quite a bit. It only makes sense that the next big advance in automobiles will be self-driving vehicles. Technically, it is close to possible to build a car that can drive itself already. Prototypes have been made, but they are not anywhere near safe enough to use on the road. The technology simply is not mature enough, but most of the remaining research involves details rather than undeveloped technology.

I expect that the technology will start with semi trucks since the drivers can still take over the vehicle from auto-pilot if there are problems. Eventually, though, the technology will become advanced enough that no truck driver will be needed in the vehicle even as an emergency co-pilot. The day that the technology is mature enough to allow for self-driven cars unmonitored by a human will be a good day for consumers because of cheaper shipping, but it will be a very bad day for anyone who makes a living largely based on their having a CDL. On that day school buses will not have drivers, but rather just an adult who keeps order. Domestic airlines will lose a lot of business because it will be easier and cheaper to rent a car equipped with a bed overnight and sleep while the car does the driving than it will be to deal with the hassle of a flight. Hotels that are not destination hotels will be hurt as well for the same reason: That no one will need to stop for a rest midway through their trip. This should also reduce the need to expand roads and lessen our dependence on oil because an automated car uses less gas than a manually driven car does and can drive closer to other vehicles (reducing its road footprint). All of this is almost technically possible now (or at least within five years), but it will probably not happen for the next twenty because of the problems that Toyota is having now.

What I expect will happen is that small bits of automated functionality will be introduced into cars slowly, such as Lexus' automated parallel parking. Even though the current pace of adding automation to cars is relatively slow, it will get even slower because sooner or later another software issue like what Toyota is seeing will appear and another CEO will be called before Congress to get chewed out regarding his inability to fix some issue that no one entirely understands but that is causing random accidents. This will happen a handful of times and each time it does it will slow the deployment of new automated functionality to a standstill.

Eventually, though, the technology will mature and the next generation will tell their kids about the good old days when cars came with steering wheels, gas pedals, and bucket seats. Not only that, we also had to drive barefoot through the snow uphill both ways...