The Internet of Things, Machine Learning, and Virtual/Augmented Reality will dominate the tech landscape in 2017

As the new year looms closer, I wanted to take some time to discuss three key technological thrusts that will be important in 2017 and should continue to expand exponentially in both market size and capabilities. These technologies do not work in silos, but rather form an ecosystem of hardware and software that will completely transform the way we interact with and understand the world around us.

Internet of Things

Lets start with the Internet of Things. As the cost of making micro-controllers continues to decrease, companies will want to attach sensors and communication devices to everything that has the ability to measure or has a quantifiable element. The idea that a particular object is “smart” or “intelligent” is not well defined, but I think of it more as an “add on” or feature than a core element of a particular object. For example, if you added a sensor to a rock that could tell you the outside temperature and humidity, the rock hasn’t become “smart” – instead it has been added as a node to a network that can be tracked, traced, and programmed as needed. One of the key questions that arises when we discuss IoT, is the density of sensors or devices that will emerge over the next several decades. With smartphones, we know that one per person on earth is a reasonable estimate – and the trends have shown that we are quickly reaching that number. But what is a reasonable measuring tool for smart devices? Theoretically, every square meter on earth could have its own smart device. I suppose this will depend on what the key use case is for these devices. If wireless electricity ever becomes mainstream, I wouldn’t be surprised if there were an IoT device where every electrical pole used to be. What if we start transporting objects on tubes that are placed on every street – like a never ending conveyor belt? In that case, you would need switching stations that would determine where each package needs to go. What if the need for ubiquitous wireless internet becomes clear and it cannot be provided by satellite or other creative flying objects like drones? We would need to have tiny wireless routers all over the place. In other words, its very hard to predict the density of “smart” things, although we can think of the use cases that might require a very dense “smart” environment.

Machine Learning

From the Internet of Things we can jump to the capabilities of machines – a topic we can generally call artificial intelligence or more specifically, machine learning. Recently I was using the app Blippar to show my little cousin how to use a smartphone to identify objects with computer vision - a key field in machine learning. At first she was astounded by the ability to name household objects with her phone. As she tried the app on various tables, pictures, and sofas she became a bit disappointed – realizing that many of the guesses were wrong. In general, I think that our experience with machine learning is not too dissimilar from my cousings experience with Blippar: somewhere on the “excited but slightly disappointed” side of the new technology adoption spectrum. It is thrilling to see a computer, whether it be a raspberry pi robot or a smartphone, hear us, talk to us, and explain the world around us. Yet, there seems to be something missing in most of these interactions. They are still too predictable and unable to grasp every nuance of a situation. In some areas, the problem is that we, as humans, are not sure what we want the computer, or artificial intelligence to do. Do we want it to guess where we are going in our car, or would we rather know what route has less traffic? It would take infinitely long to program an AI that knows what every human would like from it. Thus, the final bit of programming has to be done by the end user. In other words, the final data set that informs the machine is our own preferences. How do you create a generalized machine that is still nimble enough to adapt to millions of possible use cases? I can imagine a future where we send all of our photos and emails in the cloud to an AI who scans them and learns to recognize our family, friends, and coworkers both physically and by the personality traits they display in their emails. If given a large enough data set, an AI could pretend to be a human for one or two interactions. The question which Alan Turing posed, which is still relevant today, is – could we distinguish an AI from a human? Its becoming harder and harder every day – witness the rise of chat bots for customer service; virtual assistants like Alexa, Cortana, Siri, and Google Home; and AIs that control trading desks for hedge funds with no human interaction. Strangely enough, we might not be able to predict what a sophisticated AI will look like 30 years from now, given that much of its learning will be independent and subsequent to its original programming and “test” data sets. There are many futurists who worry that a purely logical AI will come to the conclusion that in order to save the planet, it must destroy the human race. Thus, the need to be especially careful when providing machine learning capabilities to an AI that could potentially gain access to weapons – whether they be traditional military armaments or cyberweapons.

Virtual and Augmented Reality

Moving from the physical world of smart objects and computers that can mimic humans, lets explore ways in which humans can enhance their experiences with the use of such objects. We can safely say 2016 represented the re-birth of virtual reality as a consumer experience. What caught many by surprise was the ascension of augmented reality – namely in the form of the “Pokemon Go” phenomenon. This was a great example of augmented reality merging with the Internet of Things. What was unexpected here is that there were no sensors or devices at any of the Poke-Stops or Poke-Gyms. These had been designated based on geo-coordinates established in a previously released Niantic game called “Ingress”. Of course, the Ingress players had no idea they were creating Poke Stops. They just thought they were creating portals for their augmented reality spy game! Talk about unintended consequences…When it comes to Virtual and Augmented reality, the key measurement will be market penetration, and thus units sold. At this point, the sale of virtual reality head mounted displays has been somewhat disappointing, but that should change as more engaging content becomes widely available. So far, the most amazing title I have tried in virtual reality is “Google Earth VR” for the HTC Vive. Playing this game (if you can call it that) gives you the impression of literally flying through the real world. In this case the “real world” is approximated quite well by an incredibly realistically rendered virtual world - and the graphics are only getting better. The first thing most people do (kind of egotistically) is fly above their own house, which is quite a surreal experience. I spent some time in Rio, Hong Kong, and Tokyo. The weird thing is I have memories of having been there now, even though it was only in VR. It seems that if VR is realistic enough, the human mind has a hard time distinguishing it from real life when it creates memories. If anything, this should lead us to be cautious in choosing the content that we immerse ourselves in. I for one, am not a fan of being scared by a VR alien or monster – but the number of horror titles out there shows that I might be in the minority. I suspect that in 5 to 10 years we will wear VR/AR headsets that present an exact replica of our physical world with “augmented” elements superimposed on real elements. These displays will provide an opaque setting, where we see the world as a virtual representation of reality, or a clear setting where we see through the lenses into our “real” environment with holographic overlays. Thus, vitual reality is just the opaque end of the mixed reality spectrum and augmented reality is the clear end. Where you decide to be on that spectrum, will be a matter of preference. Whatever you choose, the idea of a plain old non-enhanced reality will seem quaint and incredibly boring as soon as VR/AR reaches the use levels of smartphones. And as we saw with smartphones, it will be very difficult to get anyone to focus on anything that is not enhanced, augmented, or improved in some way.