The Ugly Truth About Self-Driving Cars

Driving

Great Expectations

Recently, I’ve been reading a lot of articles that are trying to temper the expectations related to autonomous vehicles, and with great satisfaction, I would like to say…it’s about time. If you bothered to read my article about VR being overhyped and underdelivered, you probably noticed I mentioned some other technologies that fall into that same category, and autonomous vehicles are one of them. It’s not that I’m skeptical about the benefits of the technology, I just understand that achieving those benefits are significantly further down the road than anyone wants to admit. If you don’t believe me, keep reading…

According to IHS Automotive, a leader in automobile industry statistics, at the beginning of 2016, “the average age of all light vehicles on the road in the U.S. had climbed slightly to 11.5 years.” Even if fully autonomous cars were available today, America wouldn’t see any significant market penetration for at least a decade, and most of it would be limited to higher socioeconomic areas. To everyone who thought self-driving cars were going to be bobbing and weaving down the streets of their local cities by 2020, you should probably prepare to be disappointed.

You may be asking yourself, why is the timeline is so important? It’s important because of one of the most significant benefit promised through the evolution of autonomous vehicles is related to safety, and achieving it can’t be accomplished until autonomous vehicles comprise ~90% of all cars on the road. Keep in mind that number is my personal calculation, but until self-driving cars make-up a significant portion of vehicles on the road, cities won’t see any significant decrease in the number of automotive accidents that occur every year.

Human Error

Do you know the most common cause of accidents for self-driving vehicles? It’s human error, the same thing it’s always been. Accidents have been happening for the same reason for as long as I can remember. Someone makes a wrong decision, puts other people’s lives at risk, and placing a computer at the helm of one of the vehicles won’t change this fact as long as there are humans on the road with them. Waymo, autonomous vehicle spinoff from Google, stopped reporting their accidents at the beginning of 2018, making it harder for interested parties to keep up with their efforts to remove human error from the roadways, but the good news is, California has archived all of the previous reports on their site if you’re interested in reading through them.

The most common accident type reported were humans rear-ending self-driving cars. Because computers don’t make decisions, they make calculations; autonomous vehicles will ALWAYS run the risk of being plowed into at a yellow light when there are humans in the cars surrounding them. If a computer controlled vehicle can safely stop before the intersection, it will do so, while its human counterparts can be expected to merely say the light was “pink” when they hit the intersection. Human behavior of this type is precisely why autonomous vehicles face such an uphill battle when it comes to public acceptance.

Humans expect the vehicles around them to make decisions in the manner they do, and that means running into the back of a lot of computer-driven vehicles. Running a yellow light is one of the riskiest human driving behaviors on the road, one that we take for granted as we’re driving with other humans, but it’s also one that computers won’t tolerate. Other behaviors, like coming to a complete stop at a stop sign (never happens in California…) and when turning right on a red light, will all lead to accidents between the computer and human-driven vehicles.

Winner Take All

Computers will always strive to provide an element of society that humans can never achieve, perfection, and their achievement of it will only further highlight human imperfections (more accidents). Ultimately, it will be a human that forces a self-driving car to have to choose between saving the lives of its passengers or taking the lives of other drivers. Right now, some engineer is sitting in a room evaluating a Kobayashi Maru scenario that forces a self-driving car to choose lesser of two evils in an unwinnable situation.

For example, a human driver falls asleep and crosses over into oncoming traffic, and someone has to die. Will your self-driving car decide to save its passengers or the passengers in another vehicle? You won’t know the answer to the question until it’s too late. Simply knowing that an engineer has to program a predetermined outcome into a computer for this scenario is already a scary enough thought. What I’m more afraid of is the method that needs to be employed to significantly decrease the chances of any unnecessary carnage happening as a result of these kinds of scenarios.

In a situation where a catastrophic event is inevitable, and death is an assured outcome, the best way to minimize the damage is to make sure all autonomous vehicles react to the situation in the same way. I’ll give you a second to digest that…

“To prevent additional cars from being involved in accidents, all autonomous vehicles on the road should be running the same system so they can anticipate the calculations of other vehicles in their proximity. Think of it as hive mind.”

If a car suddenly blows a tire on the freeway, every autonomous vehicle should avoid the car, in unison, at the same speed, in the same direction, to prevent any unnecessary collisions. If all the cars are running the same system, the other self-driving vehicles on the road won’t need to guess the calculations of the other vehicles involved; they’ll already know what’s going to happen. Instead of having a ten car pile up, the result is a two-car accident, saving more lives in the process.

This aspect of the technology isn’t frequently discussed, but we all know what it means, someone needs to have a monopoly on self-driving vehicle technology. Even though the United States has antitrust laws in place, to truly reach the pinnacle of efficiently concerning autonomous vehicles, only one technology should be implemented. So I’m putting everyone on notice…the self-driving car market is playing a winner take all game, and they should all know that winning is everything.

A Reality Check for Augmented Reality

Virtual Reality

The Next Big Thing

Augmented Reality

First, there was the IoT, then came wearables, and I can’t remember if virtual reality or self-driving cars came next, but I’m sad to say, none of these will pan out to be worthwhile technology investments (not actual investment advice). All of Silicon Valley’s latest technology flavors of the month have the same undeniable allure of base-metal alchemy. They all revolve around sound theories, like using energy to turn lead into gold, but the amount of energy it takes to make it happen isn’t worth the effort. I recently read an article on CNBC.com that alluded to the fact that industry outsiders are starting to pick on the fact the VR is struggling, and they’ve already counted Facebook’s acquisition of Oculus as a miss for Mark Zuckerberg.

I’m sure Oculus isn’t the only VR headset having a problem living up to the hype because most of us have already set our expectations of the technology somewhere in the upper stratosphere. Since the early 90’s, movies like Lawnmower Man have wowed audiences with the possibilities of a virtual world. If that title is a little bit too obscure for you, we won’t forgo mentioning The Matrix, and if you’re feeling really geeky, you’ll respect my name drop of Sword Art Online. If you’ve seen any of the previously mentioned virtual reality-based movies, you might recognize there is a common element in all of them – “the rig.”

The Rig

The “rig” is a pretty generic term for the contraptions the characters in these flicks strap themselves into while diving into the virtual world. The reason the rig is so crucial in these movies is that they provide a way for these characters to immerse themselves in those worlds without requiring physical movement, something that doesn’t exist today. The ability to effectively enable virtual movement without requiring physical effort is missing a key component to today’s headset based AR/VR units that will limit their commercial success. If you’ve ever used an Xbox Kinect you probably know where this conversation is headed.

For me, the Kinect was as close as any game manufacturer has actually come to producing a fully interactive experience [some us figured out you could still play Wii on the couch], and it opened my eyes to the fallacy that virtual reality represents to the general public. Virtual reality, as it perceived today, is not the next evolutionary step from where Nintendo’s Wii and the Xbox Kinect left us. Those systems were designed to have the appeal of adding physical movement to a traditionally sedentary activity, and if we’re honest, the marketing undertone of “get your potentially obese kid off the couch” was designed to get more parents on board with gaming. Ultimately, what those systems taught us was that we don’t want our virtual experience to require physical movement [not what they intended].

Signs of Exhaustion

Xbox Kinect

The first hint our new found love for immersion may not work out was when game developers had to start labeling how much physical exertion each Kinect game required. I’ve always been in reasonably good shape, but after an hour and a half of Kinect Adventures, I was ready to hit the showers. At the same time, I starting hearing rumors on the internet of people passing out while gaming, and don’t quote me, but I’m pretty sure at least one person died playing the Wii. Either way, we were all reminded why we wanted our virtual worlds to remain separate from our real ones. A lot of the activities we participate in while gaming, are things we are unable to do in real life, so if they start requiring physical movements, you’ll find a lot of us pressing the off button.

A secondary effect of all our newly immersive consoles was an increase in the amount of floor space gaming consumed. Until the Kinect arrived, there was no need to move the coffee table, notify my downstairs neighbors of potential noise, and put on slip-resistant footwear, but now all of those things had to happen before I put the disc in the console. A single player game on Kinect required approximately six feet of space to play, so for apartment dwellers, two-player gaming was mainly out of the question. If this was the kind of space required for the limited in-game movements these games offered, how much space is necessary to reproduce an entire virtual world?

1:1 Movement

One-to-one, say it with me, one-to-one. This ratio is the heart-breaking reality of why the current iterations of VR will never be a success. As of right now, virtual reality has a 1:1 movement ratio, requiring users to move one foot in the real world for every foot they would like to move in the virtual one. This situation compounds every negative aspect of virtual gaming I spoke about in the previous section. Imagine playing a first-person shooter in VR… How much running and jumping does the average avatar do in a single match? Are you planning on doing that in the real world too?

Some newer accessories to VR headsets have illustrated that developers are somewhat aware of the massive problem they are facing, but with each new addition, we get further away from the VR experience we’ve seen in the movies. Oculus and Samsung have both introduced controllers to help alleviate the problem, and I’ve also seen a few custom solutions floating around the internet, but these new accessories introduce a harsh reality that VR/AR may simply be traditional gaming with an expensive peripheral.

The moment a controller is added to the VR experience, gamers become conscious they’re just playing a regular video game while wearing a headset, and the gaming experience returns to people sitting on the couches with controllers in their hands. Until virtual reality develops the ability to “plug you in,” just like they do in The Matrix, I’m afraid the technology will continue to devolve, and I can already see Silicon Valley trying to lower expectations by marketing the experience as “augmented” instead of “virtual.” Without being able to fix the 1:1 problem, I hate to say it, but VR is not the next big thing.

Why The Internet of Things is Failing

Internet of Things

Reality Check!

Let’s take a second to step back into reality, suspending the influence of the Silicon Valley’s hype machine, and taking the time to analyze the current situation of the Internet of Things (IoT). If I had to give Silicon Valley a grade for how well they’ve influenced consumer’s awareness of the IoT, it would be an “F,” and I don’t think I would be the only person to deliver that evaluation. Overall, Silicon Valley has failed to maintain the excitement around the “The Internet of Things,” with consumers understanding very little about how these connected devices benefit them, and more importantly, not really caring. Before you can convince the world that a network of connected devices is the future of productivity, you first have to convince them of some smaller, more tangible points.

Networking Woes

If you’re [Silicon Valley] going to take on the task of connecting every device known to man, I think it would be a good idea to start with trying to make devices more accessible to connect to a network. My background is littered with networking horror stories from a variety of consumer electronics retailers, and right now, consumer’s frustration level with basic networking could potentially be the single most significant hurdle to a world of connected devices. The “Networking Equipment” category is consistently one of the most frustrating for retailers, with return rates always ranking among the highest in store, and it poses a customer service nightmare for every party involved.

You see, networking is one of the only categories in retail that enlist more than two parties to safely and securely create a home network. A best-case scenario limits the interactions to three entities: the retailer who sold the connected device, the internet service provider (ISP), and the manufacturer of the networking equipment. Any business transaction involving more than two parties opens itself up to a plethora of problems (I’m looking at you Uber), and in this case, three is definitely a crowd.

To alleviate this problem, manufacturers of networking equipment have decided the most natural thing to do is engineer themselves across the finish line (typical response from geeks). Innovation after innovation has been applied to networking equipment, starting with connection wizards, peeking with Wi-Fi Protected Setup (WPS), and sadly ending with auto-connecting mesh routers. Personally, I probably would have given up when consumers weren’t able to figure out one-touch connections with WPS, but the industry keeps trudging along.

Security Woes

When customers can successfully connect the latest generation of connected devices, things haven’t always gone as planned, as demonstrated by a string of highly publicized security breaches dominating the headlines. In October of 2016, one of the most significant internet outages ever witnessed was caused by hacked IoT devices. 2017 was ushered in by the release of Brickerbot, an attack specifically designed to permanently disable poorly designed IoT devices, a process known as “bricking.” All of these security snafus are linked to one specific manufacturing goal – maintaining the bottom line.

Manufacturing IoT devices happen in the manner as any other product, which means manufacturers adhere to the same priorities, with the goal being to manufacture these products at the lowest possible price point; This translates into lower memory capacity, less built-in security, and minimal investment devoted to human IT resources. The result has been a large quantity of these devices shipping with their default configuration and customers who don’t possess the knowledge or patience to change their configurations leaving them as is.

So what’s wrong with the default configuration? To put this into layman’s terms, it’s the equivalent of breaking the first rule of Fight Club. Everyone knows the rule, it should be easy to follow, but it continues to be broken. The first rule of internet security is “never leave your device’s username and password on the default settings,” as doing this creates opportunities for anyone who can read an instruction manual to access the device [you should be thinking about your security right now].

Talking Solutions

I’m not here to bitch-n-moan about the world without offering up some solutions. As a former corporate trainer in the consumer electronics space, I understand the importance of consumer education and how much better off a situation can become by applying a bit of knowledge to it. RTR Digital offers a Networking Basics course at our learning site, RTR Learning, but if you’re not going to enroll, we’ll still provide some basic tips.

Change the Default Settings – Every device is shipped with a default username and password (usually on a sticker on the device) as a way for users to access the setup menu, configure the device, and then change the password so no one else can access the administrator settings. Never leave the username or password set to “admin” or “password1234”.

Name Your Wireless Network Something Abstract – When configuring your wireless network, don’t include any personally identifiable information (e.g., name, street number, house color). If someone is determined to access a network, physical proximity is the key, and associating the network work with its location is giving up way too much information.

Use the Guest Network Feature – Setting up a guest network is a feature on almost all modern networking equipment – use it. The guest network feature enables you to hand out the password to your wireless network without exposing your personal information in the process. Devices on the guest network are given internet access but are in a separate part of the network from the devices connected to the primary network.

Create Unique Passwords – When creating passwords, every device should have a different password, which means you can’t use your kids’ birthdays every time. It may seem like a hassle, but to make it easier, you should come up with a system that enables you to remember them. You can try something like the first three letters of the manufacturer’s name, followed by the purchase month and year of the device.

Ultimately, it’s up to retailers, manufacturers, and consumers to take control of their own responsibilities when dealing with connected devices. There are three parties involved in the future of the IoT, and all three of them have to decide how bad they want the world that is promised to them. Feel free to leave comments or questions on this post, and we’ll be sure to respond.