Somewhere around 1995, a manager, Bill, at work commented how he was not looking forward to the day when he had to pull over to the side of the road and reboot his car. To be clear, we are already there and have been for some time. We were really there already in 1995.
Every new car has some form of the rather unhelpful "Check Engine Soon" light. These lights rarely lie, although faults that they signal may be intermittent or transitory. Whether the identified fault will cause imminent and catastrophic damage, or it is merely an aged sensor beginning to slow, the driver is free to continue merrily on the way - or nervously on the way depending on his or her level of chronic worrying.
This situation IS the "pulling over the car to reboot" - but the default state is to allow the driver to continue. The default to continue does not work in a driverless car future; the fail-safe is to be immediately off the road until all systems are go. Without the driver, the job of many of the sensors and associated software takes on a more life-altering role beyond the life of the automobile and beyond the life of those in that immediate automobile. Until driverless technology becomes mundane and invisible, this paints a very frustrating future vision.
Luckily robots are not programmed with a sense of fairness, but reliability of "things" is expected to be much higher than of people. Brakes are expected to work 100% of the time, even if a human doesn't use them appropriately and takes on a telephone pole. Computers are supposed to boot up every time even if someone clicks on a phishing link and opens the computer up to external control. Cars are supposed to start even if basic maintenance is sometimes ignored. So it will be with autonomous vehicles; every situation encountered will be expected to be handled expertly, even if that isn't the standard for humans. If a deer runs in front of a driverless car, while the car is driving through an oil patch with winds dropping a tree across the road and the car crashes, it will be interpreted as a vehicle/software failure. Developing vehicles to drive autonomously through 75% driving situations is probably feasible right now. Expand that to 95% and it gets difficult, 99% harder. That last 0.001%, becomes nearly impossible...
Cars are also being developed for the globe, bringing efficiency to manufacturing reducing development costs. Creating autonomous cars that will work as well in Minot, North Dakota as they will in New Delhi, India seems daunting, unless global development reverts to regional vehicles - this is economically unlikely and increases the technical difficulty even further.
I actually enjoy driving. No, getting stuck in heavy gridlock isn't fun, but I've organized my life so that is rare. I even often enjoy the daily commute, although the destination may not be ideal.
Listening to the media would suggest I'm a dinosaur in this, but I don't buy it. I think this is partially due to the growing urban/not-urban divide, which itself is painted with a city-scape lens by the largely urban media.
Uber - basically a less unsexy taxi cab - and driverless cars are now prompting clairvoyant wizards to foresee the death of the personal automobile. Even douche bags are telling us cars are sooooo 2012 (I've never met this douche bag, but reading what he has written certainly suggests he's quite comfortable in his douche-bag-ness).
We've seen this before. Dean Kamen told us the Segway was going to revolutionize cities. Cities were going to be torn down and rebuilt all around his mall-cop-mobile scooter; cars were going to be melted down and turned into Segways. Dean Kamen may be a smart guy, but he must live in a very narrow bubble.
The one thing we can predict, is that we can not predict the future. Technology can be envisioned, but how it is implemented is something different entirely. And how it eventually gets used once it is in the grubby paws of the average person is different still.
Because the future can't be predicted, I'm going to go ahead and predict the future.
1) Driverless roadways are decades away.
There are a lot of old cars out there. Vehicle quality continues to improve, and older cars are remaining reliable for longer and longer. The often-cited average age is 11 years. And that is average, meaning half are older than that. Not only that, but how driverless technology will interact with drivers is evolving slowly. Driverless cars can be programmed to respond in a very predictable manner, but the remaining analog interface (the human) can not - whether in a driverless car or not. There are also a lot of other interests that play a role in addition to millennials who don't want to drive (the difference between an interest and a special interest is whether one agrees with it or not): motorcycles, bicycles, antique and vintage automobiles, roadkill.
Much like the houses of the future are only slowly showing up due to a lack of infrastructure and because nobody is tearing down all the old houses to build new ones, driverless roadways will need infrastructure improvements that don't currently exist - we can't even currently keep our bridges safe and roads (relatively) free of potholes, adding in new infrastructure (and ongoing maintenance) seems even more unlikely.
2) Driver aiding technology is here and will continue to be grow.
Driver aiding technology is currently at the curiosity stage. Tesla's successes and failures are widely and loudly reported. Fairly or not, this is true whether the technology was being used as intended or not. Quoted from above: "And how it eventually gets used once it is in the grubby paws of the average person is different still."
Every manufacturer is looking at some form of driver aided technology, and calling it driverless. Regulations are slowly evolving, which is probably a good thing.
Some of this will be good. Some will be bad.
3) Driver aiding technology will be expensive and frustrating - at first.
Cruise control, fuel injection, anti-lock brakes, throttle by wire, touch screens, dual-clutch transmissions, etc... These are just a few examples of technology that was first available on (primarily) expensive vehicles, but worked its way down to plebeian cars. Once in average-joe-mobiles, these things usually work, but still have levels of evolution to go through until they reach nirvana - aka nobody thinks about them anymore. Until that point is reached, it can be frustrating. So it will be with driverless aids.
But advances will be made. Nobody thinks about cruise control, fuel injection and anti-lock brakes anymore (and few would want to go back to carburetors or drum brakes on all four corners). Touch screens and dual-clutch automatic transmissions are getting pretty good, with some warts still showing.
There is a fundamental difference, fuel injection was and is happy existing within the confines of the individual vehicle. If fuel injection required all (or most) other cars on the road to also be fuel injected, or a complicated set of rules for the fuel injection to work and for the carbureted vehicles to aspirate fuel, the evolution of fuel injection would have been very slow.
I certainly hope we are not setting ourselves up for a late-1970s re-reality, where the mandates for fuel efficiency and emissions got so far ahead of the technology that the result was a lot of really malaiseful cars.
4) Somehow, driver aiding technology will be used for porn.
Every technological advancement always seems to be used for porn. Glossy magazines, VHS, DVD, HD-DVD, online streaming, cell phones, smart phone, personal computers, real-time streaming... So somehow I'm sure driverless cars will be used for porn too.
Sadly, I may actually be a dinosaur. I'm not really looking forward to the current vision of the dystopian driverless future. I enjoy driving. I even still prefer a manual transmission. I'd be lost without a motorcycle. What I'm encouraged by is that this change will happen much slower than what visionaries are sure of - and it will look a lot different than they see as well.
No comments:
Post a Comment