Futures in Computing

Editor's Notes

• #2: Brand new talk, first time I’m giving it. And like all good talks it was finished an hour ago. You can find these slides up on slideshare.net
• #3: Im 36 years old. In my lifetime, I’ve seen, among other things; the computer go mainstream, the birth of the internet, the invention of the smartphone, the beginnings of the smart home, and the inception cyberwar.
• #4: We’ve taken nearly the entirety of handheld consumer gadgets and the desktop computer, and combined them into a single device, and connected it to every other device on earth.
• #5: In the next 30 minutes, I’ll do a review of the state of the art and future possibilities the following three segments of technology, Mobile, Internet of Things, and Artificial Intelligence And by the end of this talk, hopefully we’ll be using some different terms for the latter two.
• #6: First of all, it’s important to answer; why do we care. Not only is it fun to speculate on the nature of the future, but it’s actually quite useful in our day to day lives. A while back I saw this clickbait article flying around facebook about how successful entrepreneurs take the time to daydream. Now, for me, I can tell you two awesome things about daydreaming; first, it’s really enjoyable, it’s really, really enjoyable to sit down and relax and just think about things to come. As the fantastic Mercury Rev song goes, “Plans and schemes. Hopes and dreams.” Second, the output of daydreaming is actionable; it gives goals, and a framework in which to execute them. Much of my career has been spent building on the latest tech, and trying to invent tomorrow’s. Having an eye on the future helps to me to target my efforts so that I can maximize my impact by leveraging the momentum of the zeitgeist. It also helps you manage risk because you have an idea of where things are going. But truthfully, probably the best part is that it’s enjoyable. I love sci-fi. That’s where our entire livelihoods have come from, whether we realize it or not. So if I walk today and you’re just a little inspired to think about these things, for me, it’ll have been a successful experience.
• #7: So why do prediction fail? Just because something is technically feasible doesn’t mean it will happen. Flying cars are technically feasible, and practically unfeasbile. Technology adoption has a lot to do with our social and cultural customs. Not just the technology. And that’s great, because it means we’re surprised by how things actually are used. One of the greatest things about being a library or a platform creator, is seeing it used in ways I never imagined.
• #8: So let’s get into it shall we? Mobile futures.
• #9: The world of mobile looks much different today than when we started Xamarin 4 years ago; when I used to go around and tell people that mobile was the next big thing. Well, now mobile is the big thing. In fact, it’s the biggest thing. By the way, if you haven’t see Benedict Evans’ Mobile is Eating the World talk, you should. Benedict is a partner at the $4B VC firm Andreessen Horowitz and he’s brilliant. He goes into much more detail about just how big mobile is and how fast it’s growing, but I’m going to refer to pieces of his talk here. ----, though I’m not convinced he and John Oliver are different people; he looks and sounds just like him.
• #10: So just how big is mobile? Well, take a look at the following graph I pinched from Benedict’s talk. We’re used to looking at graphs that hockey stick up to the right, so this is a little different but bear with me; this is a graph that illustrates how few adults in the world are unconnected; that is, those that don’t have a reliable connection to the internet. Soon, being unconnected will be like Polio; it happens, but its rare. One thing you’ll notice here is that we’re basically at a point where folks are connected via smartphones. Of course, if we looked at US data, the graph is even more dramatic; only about 15% of the population doesn’t have access.  
• #11: In fact, more time is spent online via apps than all other media combined in the US. (source, p14). And to use a smartphone, you essentially have to buy or acquire one ;so smartphones and tablets now account for half of the global consumer electronics industry. (s: mobile is eating the world) 
• #12: This is because 4B people on earth buy a new phone every two years, vs computers which have about a 5 year life span, and less and less people are buying them. Sometime around September 2011, the PC sales business started declining. Think about that. 4 years ago, the PC business began to shrink. So not only have mobile phones been adopted like no other technology in the history of tech, but people are actually turning away from traditional desktop computing. 
• #13: In less than 5 years (2020), it’s projected that 80% of all adults on earth will have a smartphone. That’s 4 billion smart phones. To contrast that, there are perhaps only 700MM consumer PCs out there. In Sub-Saharan Africa, 70% of the population has access to cellular coverage, more than have access to plumbing or grid electricity, and 35% have smartphones. And of course with cheap Android, that’s growing everyday and feature phones are having a die-off. (source). 
• #14: And remember; a smartphone is basically a pocket supercomputer. An iPhone 6 has 625x the number of transistors as a pentium from 1995. They’re also packed with sensors and hardware features that have replaced other tech appliances. 
• #15: Do you remember the platform wars? Well, they’re over. Ok, maybe not over, but they’ve reached an armistice. There are really only two players left; iOS and Android. Between the two, they account for nearly 100% of store revenues, and nearly all device sales. Windows is essentially a non-player in mobile. The app-gap has paralyzed it. And I don’t expect any real turnaround here soon. One of Windows’ biggest value points is that if you develop for Windows Phone, it’s easy to take it to the desktop, but with PC sales taking a dive, and desktop computing being less relevant, that’s less and less important.
• #16: So we’re left with the iOS and Android Club. Apple owns the high-end market, and Android wins in penetration. (Source: VisionMobile) Apple has somewhere around 20% of the device market share, Android has nearly 80% (s: Gartner, May ‘15), but in terms of app store revenue, 2/3rds of it is generated under iOS, and 1/3rd of under Android (s: Benedict Evans). So even though 4x as many Android devices were sold in 2015 than iOS, each Apple device generates around 8x the revenue. And where is Microsoft here? Well in terms of app-store revenues, they’re vying for scraps. 
• #17: Microsoft’s play here is to cover the app-gap by running Android and iOS on windows phone, as well as unify the phone/tablet/desktop operating system design patterns. They now have a sizable share of ownership in CyanogenMod, which is the most-popular custom flavor of Android these days. Their intentions here are clear; be able to run Android apps on Windows natively. There is also some cultural indications that the enterprise may choose Windows 10 as their enterprise strategy, with the unified tablet/desktop play here being an advantage, and security and device management being a known quantity with Windows. But a lot of enterprises are beginning to standard on Android and iOS; so unless Microsoft an ace up their sleeve, they may be facing a struggle for a while.
• #18: Going back to the app-store, something else that’s happened here is that it’s matured. There’s now an app for everything and consumers are getting app-fatigue. 90% of all downloaded apps are run one time and never opened again (source). And so the barrier to entry is much higher. Most of the common app scenarios have lots of options for apps, and consumers have high expectations around those app experiences. Consequently, the price to acquire customers is high. The first 1/3rd of your app store revenue is cut off the top because it’s given to the app store itself, and typically half of what’s leftover goes to marketing and ad spends. (s: visionmobile) So we’re looking at high margins here. Certainly not traditional software margins.
• #19: But despite the app store frenzy dying down, there is still a gold rush on, and it’s in the enterprise, where they’re spending billions. Microsoft bought Acompli for outlook on mobile for $200MM (source) and slack, who has invested heavily in the mobile experience got to a $1B valuation in less than a year.
• #20: In fact, 43% of Enterprise App Developers make > $10k/month (source: Vision Mobile) While only 19% of Consumer App Developers make that.
• #21: You know what’s interesting? We spend nearly all of our time talking about the app store, but that conversation is already stale. This year, 2.5x the amount of money will be made via ecommerce, than all other sources combined. (s: visionmobile p:39) That means App Store revenues are a distant second. 
• #22: And yet, only 9% of developers are developing apps that have ecommerce in them! (s:vm p:40) That means that developers are seriously under allocated here! Many app developers are still chasing the old dream. But you know what, now it’s hard to get enough of an audience in which to leverage them to make the ecommerce money. So it’s a bit of a chicken and egg problem. That’s not to say you can’t make money in the app store, but you really have to find an underserved niche market.
• #23: So where do you find the money? Today, developers are best served by either finding good paying jobs in existing businesses, or seeking new opportunities. So let’s take a look at those.
• #24: The interesting plays are now in emerging technology and experiences. It’s a really good time to be daring, take risks, and try new things. Established ideas have a high barrier to entry, but daring, new ideas are often built by small teams or even single developers. There are a number of entries into the Unicorn Club, that is, those companies with a valuation of $1B or more, that have been started by single engineers. The future of the mobile is more than mobile; it’s an integrated, contextual experience. Which segues perfectly into: 
• #26: So what is the “The Internet of Things?” Sounds so wieldy. Well, it turns out, IoT is kind of a softly defined blanket term for cloud or internet connected devices, and they fall into a broad set of categories, including wearables, connected cars, etc. There’s a lot of talk about IoT being the next big thing, but thinking about them in terms of broad categories of devices, or the technology that powers them, we’re missing the point; because it’s not the technology, or the device itself that’s important, it’s the experience that you create with it. IoT comes down to two things. Connected slash Integrated devices, and Contextual Computing. Computing that knows where you’re at, what you’re doing, who you’re talking to, what you’re seeing, etc., and giving you an experience that is contextual to all those things. Let’s look at an example of a successful IoT experience.
• #27: Nest. By nearly any metric, Nest has been successful. It reinvented the thermostat. Phenemonal reviews, incredible adoption, and let’s not forget, a $3.2B acquisition by Google. And what is Nest? It’s just a thermostat, it’s really the beginning of IoT. A thermostat that is smart, connected, and a great experience. And that’s really what matters; the experience. Nest is so simple, it replaces your existing thermostat without any complicated installs, it senses when you’re home, when you’re away, it learns from your habits, and it has a beautiful integrated experience with your mobile device.
• #28: So how was Nest made possible. Let’s break it down. First of all, there’s no magic hardware there. The magic is in the experience, I keep saying that, because it’s so important. Nest is built with off the shelf components that are now commoditized. It has an LCD screen, wifi and bluetooth radios, a modern battery, etc. If you think about it, Nest is essentially a reconfigured mobile phone. And it’s the mobile phone supply chain that has made it possible. (s:mobile eating world, s:21) Which is leading to...
• #29: A hardware renaissance. Because now, hardware can be developed at the speed of software. Because it’s software that runs them. Think about it; you can now program hardware using a number of high level languages, including C#, Java, and others. By a show of hands, who here has played with Netduino or Arduino? Someone in this room might build the next Nest. Someone in this room might be $3.2B richer in a few years.
• #30: So mobile gave us pocket supercomputers and a supply chain to reconfigure those any way we can imagine. The promise of IoT is the promise of integrated, contextual computing. But really, the term “integrated” is already passé. That’s the thing about technology, as soon as it becomes ubiquitous we stop talking about it. Context, that’s what is king. Contextual experiences are going to be the next big thing.
• #31: So what’s next for Contextual Computing? Well, for one, there’s going to be a lot more devices out there. And nearly every device is an experience multiplier, because while they may be designed for a particular experience, their usage will be as wide and varied as the developer’s creativity that build for them. And because there will be so many new devices and experiences (remember, they’re cheap to create now), they will become much more ubiquitous. And not only will nearly all of them will be connected, but we’re developing secure, standard protocols for these devices to talk to each other and communicate with mobile. They’ll also be smarter. It’s one thing to come home and have your house heat up. It’s another thing altogether to come home, and have the house heat up, the lights come on, the garage open, start the music playing, notify you of any calendar obligations, and have your evening cup of espresso begin to brew. And how will these experiences get smarter? Artificial Intelligence.
• #33: When we talk and about AI, it’s often in the context of futuristic dystopias in which androids are better, smarter, and faster than humans. It pervades our scifi literature, and has been the subject of numerous films.
• #34: Hell, even I made a movie about it. Shameless plug, you can watch it on Vimeo, the link is there. Elon Musk and Stephen Hawking, in very famous blog posts, have come out against AI and are saying that will be humanity's undoing.
• #35: But I don’t think we should be afraid just yet. I think that the boundary between humans and AI will change so much, that we cannot possibly predict it’s outcome. So let’s break things down a bit. While the term Artificial Intelligence is still widely used, many people have started using the term Cognitive Computing in place of AI. I think probably because it doesn’t have some of the weight of things it’s not. In its strictest sense, cognitive computing is a way of making computers process more like their biological counterparts. And while most people think of human-like prowess of intelligence when we say Artificial Intelligence, Cognitive Computing evokes a sort of “small-task” problem solving. And it’s all around us. So; before we buy, lock, stock, into the dystopian futures dreamed up by sci-fi authors, that wind up as simplified fear-filled 24 hour news headlines; we, as engineers, and critical minded folks need to understand that we’re the ones that will create and influence what AI becomes. Think about that; to some degree we, and our peers are in charge here. Now as a storyteller, I can tell you that part of the reason we dream up these scenarios is that we can warn the world of risks and create a conversation. We try to awaken awareness within the communities that embrace these stories, and those communities, in turn, interpret and inform wider audiences. They have a place and a value in society, but narratives have the distinct ability to present a single side of an argument disguised as two. Because the narrator owns the debate. 
• #36: As I mentioned, we interact with AI every day, even if we don’t know it. On our devices and in our homes, we use Microsoft’s Cortana, Google’s Now, Apple’s Siri, and now even Facebook is getting in the game with it’s recently announced M. We get product recommendations on Amazon, we get listening recommendations from RDIO, and in both cases, these are backed by machine learning algorithms that get better the more you use them because they learn from you.
• #37: Of course, no discussion of AI would be complete without IBM’s Watson. Here it is destroying two of humanities finest recall devices. IBM’s Watson might be humanity’s best cognitive computing yet. Watson is a better dr. than House MD. Input your symptoms and get a diagnosis that is much more accurate than a traditional dr., and for exotic diseases where regional doctors may not have any expertise, Watson really shines. And now, you can use Watson in your own apps in a variety of ways.
• #38: Let’s look at the options available to us as developers, for integrating cognitive computing into our apps. There has been a tremendous amount of R&D dollars put into cognitive computing. Between Google, IBM, Microsoft, and Yahoo alone, there have been billions spent on AI. (forrester). Plus there’s a ton of VC money out there for AI. Combine this with the computing power and storage of the modern cloud (AWS, Azure, IBM) and you've got incalculable computing power over nearly limitless data. Cognitive computing generally falls into a few categories including Assistance, Prediction, Accessibility, Analysis, and Communication. Now this is certainly not all of the AI possibilities; but what’s available to us today, might surprise. 
• #39: Of course, probably most of you have head of Azure, but did you know that Azure has a general purpose Machine Learning service? They do, and it’s fantastic. It’s incredibly easy to learn, has a GUI interface to create new “experiments,” built-in big datasets, and you can convert them into a RESTful service interface to use in your app at the click of a button. Along with the hololens, this is probably one of the most interesting things Microsoft is doing right now, and I would argue, this is far more useful than the hololens. At least today. 
• #40: There are also a ton of other Cognitive Computing offerings that today, you can do file > new Xamarin app, and put these things to work in your code. There are natural language generation services that analyze your data or input and automatically write reports for you. We have call center and customer service optimization services that match people up with the right folks to talk to and offer solutions on the fly. There are ad and media services that use AI to measure and adjust marketing efforts for you.
• #41: And there are a ton of other services as well that do everything from medical diagnosis to threat detection, some of which are powered Watson, itself. 
• #42: Now remember how I mentioned that the boundary between AI and humans has yet to be defined? Well, it’s because the boundary between computers and humans has yet to be defined. I believe we’ve made, and I’m guilty of this as well; a false dichotomy between “us” and “them.” Thus far, we’ve spoken about AI as a wholly external, non-human thing that exists in a computer, largely performed over large datasets. But this is where we might be entirely missing the point in our collective cultural consciousness. I believe that with the help of our friends down the metaphorical hall in bioengineering, the human/computer interface will begin to blur. We’re already building biologically based cognitive computing devices, and our nervous system interfaces are getting better. We now have ear and retinal implants, active prosthetics and more.
• #43: And this frightening picture was created just 3 hours north of us at the University of Washington, in which the man on the left, controlled the man on the right’s finger, with his mind, over the internet. Ladies and gentlemen, I am not making this up. This is far more sophisticated than the morse-code brain message that was sent over the internet a couple years back.
• #44: So what is the future of AI? So where does that leave us, and most importantly, should I be afraid of skynet? I believe; no, we shouldn't be afraid, and yes, via more sophisticated human/computing interfaces, AI may be the end of humanity as we know it today. But I think that’s a long way off. Right now the #4 supercomputer in the world, running flat out, still only has 1% of human brain power. I think it’s likely, that by the time human-powerful AI is here, our relation TO AI, will have changed drastically. It’s true, computers may become us, but we are becoming computers too. We’ve already offloaded a lot of our tradition knowledge and procesing infrastructure to external sources. From a calculator to Google, we even have a term for it. It’s called external brain. The fact that the term isn't in latin or some other pretentious naming that you'd find in psychology today should give indication as to just how mainstream it is. Millennials can't get around without GPS, I couldn't live without wikipedia. but that's the tip of the iceberg. 
• #45: Whether we want it or not, AI is progressing quickly. It will get smarter, it will faster, and it will get more human like. And in nearly each place that it gets integrated, the experience will not be altogether unexpected; meaning that we likely won’t even realize the experience we’re having is powered by AI. The real problem here is that if we don't progress our ideas of capitalism and economics, the real challenge will be socioeconomic well before we ask ourselves the Philip K Dick question of what is humanity. I _highly_ recommend you watch “humans need not apply” on youtube. It’s short. [why: AI will put us out of jobs]
• #47: Before I leave, I wanna remind you that just because something is possible, it doesn’t mean it’ll happen. We’re probably not going to walk around everywhere with a hololens being fed ads from box stores because one, we’d look like a jerk, and two, it would probably not be fun. Not even Brad Pitt could make Bluetooth cool. Imagine walking around like a fighter pilot with a Microsoft heads up display. This likely won't happen. But what does happen, is up to us.
• #48: Ultimately, the future is in our hands.