What’s in Store for 2016: A Few Predictions

Share via Twitter Share via Facebook Share via Linkedin Share via Reddit

Volviendo / Coming Back...

Every so often, it’s worth taking a step back to survey the wider technical landscape. As analysts, we spend the majority of our time a few levels up from practitioners in an attempt to gain a certain level of perspective, but it’s still worth zooming out even further. To look not just at the current technical landscape, but to extrapolate from it to imagine what the present means for the future.

For six years running, then, I’ve conducted this exercise at the start of the new year. Or at least plausibly close to it. From the initial run in 2010, here is how my predictions have scored annually:

  • 2010: 67%
  • 2011: 82%
  • 2012: 70%
  • 2013: 55%
  • 2014: 50%
  • 2015: 50%

You may note the steep downward trajectory in the success rate. While rightly considered a reflection of my abilities as a forecaster, it is worth noting that the aggressiveness of the predictions was increased in the year 2013. This has led to possibly more interesting but provably less reliable predictions since; you may factor the adjustment in as you will.

Before we continue, a brief introduction to how these predictions are formed, and the weight you may assign to them. The forecast here is based, variously, on both quantitative and qualitative assessments of products, projects and markets, based on everything from hard data to off hand conversations. For the sake of handicapping, the predictions are delivered in groups by probability; beginning with the most likely, concluding with the most volatile.

With that explanation out of the way, the predictions for the year ahead:


  • Bots are the New UI:
    There are two dangers to delaying the release of your annual predictions until well into the new year. First, they can be proven correct before you publish, meaning that your prediction is no longer, technically speaking, a prediction. Second, someone else can make a similar prediction, which can – depending on the novelty of what you forecast – steal your thunder.

    Both of these have unfortunately transpired over the past month. First, Google’s Cloud Functions and IBM’s OpenWhisk obviated the need for my doubling-down on a bullish forecast for serverless architectures. And just a few weeks earlier, Tomasz Tunguz – who is always worth reading, incidentally – unknowingly stole major elements from my prediction regarding bots in a piece entitled The New UI For SaaS – The Question.

    One of the most surprising conversations I have today is with enterprise vendors who dismiss Slack as a messaging vendor, or with engineers who view it as little more than an IRC-implementation for muggles. Both miss the point, in my view. First, because they miss the platform implications, which I’ll get to, but just as importantly because they obscure the reality that bots are the new UI.

    Consider the universal problem of a user interface. If you’re implementing a GUI, you face increasingly difficult decisions about how to shoehorn a continually expanding featureset into the limited real estate of a front end. Making matters worse, aesthetic expectations have been fundamentally reset by the incursion of consumer-oriented applications. And while you’re trying to deliver a clean elegant user interface with too many features, the reality of mobile is that you’ll probably need to do so with even more limited screen real estate.

    Those whose users primary or sole interface is the command line have it easier to some degree, but their lives are also complicated by rampant fragmentation. Gone are the days when you could expect developers to memorize every option or flag on every command because there are simply too many commands. Too many developers today are reduced to Google or Stack Overflow as an interface because they’re not using a given tool quite enough to have completely internalized its command structure and options.

    Attempts to solve these user interface problems to date have essentially been delaying actions, because the physics of the problem are difficult to address. Complexity can only be simplified in so many ways so many times before it’s complex again.

    Enter the bot, which is essentially a CLI with some artificial intelligence baked in. Deployed at present at relatively narrow, discrete functional areas, their ultimate promise – as Tunguz discusses – is much broader. But for now text-based AI’s such as X.ai’s Amy or the Slack-based Howdy or Meekan point the way towards an entirely new brand of user interface. One in which there is no user interface, at least as we are typically acquainted with that term. If I want to schedule a meeting with someone via Amy, I don’t log in to a new UI and look at schedules, I use the same user interface I always have: email. Amy the artificial assistant parses the language, has contextual awareness of my calendar and then coordinates with the third party much as a human would. Or if I’m booking with one of us internally, I no longer have to open Google Calendar: I ask Meekan to pick a time and a date and turn it loose.

    And bots are not just for scheduling meetings – or ordering cars from Uber. Within the coming year we’re going to see tools extensively ported to bots. Why can’t I start and stop machines via a bot as I would the CLI? Or ask questions about my operations performance? Or, elsewhere, my run rate or cashflow? Some of our clients are working on things like this as we speak, and Slack’s December Platform launch, including the botkit Howdy, will speed this along.

    We’ve all had the experience at one point or another – particularly if you’ve ever used Google Analytics – of paging endlessly through a user interface for something we know an application can do, but can’t figure out how. What if you could skip that, and simply ask a bot in plain English (or the language of your choice) to do what you want?

    Folks who have been using things like Hubot for years already know the answer to this. As platforms like Slack expand, more of us will begin to realize the advantages to this in 2016, as bots become the New UI.

  • Slack is (One of) The New Platform(s):
    Based in large part on the absurd success, both in terms of marketshare and revenue, of Microsoft’s twin platforms, Office and Windows, software businesses ever since have attempted to become platforms. Most of these efforts historically have ended in failure. Becoming a platform, as it turns out, is both expensive and entirely dependent on something that is intensely difficult to predict: volume traction. Even for well capitalized would be players with platform ambitions, the dynamics that lead to the annointment as a platform are difficult to navigate.

    Few, particularly those who still regard Slack as a jumped up instant messaging client, would have anticipated that Slack would become such a platform, but it’s well on its way. We have had persistent group chat clients and capabilities for decades, at this point, and for all of their immense user traction, even the most popular IM networks never made the jump to platform. Domestically, at least: China’s networks are materially distinct here.

    Most obviously, Slack’s growing its userbase: it essentially quadrupled over the past calendar year from around 500,000 users to over 2 million. But the important jump was in its app catalog. From 150 apps in the catalog at launch, Slack has almost doubled that number to 280 at the moment. And we’re seeing significant interest and traction from third parties who’d like to add themselves to that number, because Slack is checking an increasing number of the boxes first class platforms have to to be taken seriously.

    When we look back on 2016, then, it will be regarded as the year that Slack became a platform.

  • Newsletters are the New Blogs:

    Whether you attribute the decline in RSS and its client applications to the rise of social media like Facebook and Twitter is, to some degree, academic. Whether they were the cause or simply the beneficiary, the fact is that a great many whose consumption of content used to depend on RSS readers now look to the social networks to fill a similar need.

    Similar is not same, however. As Facebook’s algorithmic feed and Twitter’s much excoriated dalliances with something similar have demonstrated, one of the difficulties with social networks is that they’re difficult to scale. With an RSS reader, you don’t miss a post from an author you’ve subscribed to. With Facebook or Twitter, the more you friends you have, the more difficult it is not to.

    Enter newsletters. Well, technically that’s not accurate, as they’ve been around since well before RSS readers or social networks. But since the demise of the former and the rise of the latter, newsletters are increasingly becoming the de facto alternative, as Paul Ford suggests above. If you want to be sure readers don’t miss your content, and readers are similarly interested, newsletters have been pressed into service as the solution.

    In 2016, we’ll see this trend go mainstream, and authoring tools designed for actual authors rather than, say, marketers, will emerge.

    All of which means I probably need to start a newsletter already.


  • Open Source is the Future, and It Will Become Evenly Distributed:
    The rise of open source at this point has been well chronicled. While the most efficient mechanisms for commercializing open source software remain hotly debated, the sustainability of open source itself is no longer in question. In an increasing number of scenarios, open source is viewed even by staunchly capitalistic businesses as a logical strategic choice.

    Even so, we haven’t yet hit the tipping point where it’s the default software development model. There are still many more scenarios in which open source is an exception, a mere science experiment, rather than the most logical choice for a given piece of software.

    There were signs in 2015 that this was changing, and this will accelerate in 2016. Google, for example, has typically guarded its infrastructure software closely. It published the details that made building Hadoop possible, but kept its actual implementation closed. With Microsoft’s CNTK or Google’s TensorFlow and arguably Kubernetes (it’s not Borg, but a reimplementation of it), this pattern has begun to shift. Apple’s decision to make its Swift runtime open source is another example of an organization which has historically been protective of its software assets recognizing that the benefits to open source outweigh the costs of proprietary protections. Even in industry, enterprises are beginning to see the advantages – whether in developer marketing/recruitment/retention, cost amortization, etc – and make strides towards either releasing their own internal software as open source (see Capital One’s Hygieia) or easing restrictions on contributing back to existing projects.

    Open source will become evenly distributed, then, in 2016.

  • SaaS is the New Proprietary…But Will Lead to More Open Source:
    As I have argued previously, SaaS is on several levels a clear and present danger to open source sofware. First, questions about access to source are deemphasized in off premise implementations in ways they are not in on prem alternatives. Second, many SaaS offerings have incorporated the embrace, extend and extinguish model by building attractive proprietary extensions onto open source foundations. Lastly, just as open source enjoyed massive advantages in convenience and accessibility over proprietary alternatives, so too is SaaS more convenient than OSS.

    Where many OSS advocates still consider traditional proprietary software the threat, then, they would do better to shift their attention to SaaS alternatives.

    All of that being said, SaaS is counterintuitively a potential benefactor to open source in important ways. As described above, important SaaS vendors are both investing heavily in software development to tackle very difficult, unique problems and realizing that the benefits to making some or all of this software available as open source outweigh the costs.

    The Platform-as-a-Service market is perhaps the industry’s best evidence of this. The initial implementations in early 2007 – Force.com and Google App Engine – massively lagged IaaS alternatives in adoption not because of technical limitations, but because of their proprietary nature. The technical promise of PaaS – focus on the application, not the infrastructure it runs on – was intriguing from a developer standpoint. But no one wanted to write applications that would never run anywhere else.

    Fast forward six years and the PaaS market is a promising, growing category. Why? Because customer concerns about lock-in have been mitigated via the use of open source software. As ever, developers and the enterprises they work for are more likely to walk through a door they know they can walk back out of.

    AWS’ Lambda is a more recent indication of this phenomenon at work. Technically innovative, it underperformed from a visibility and adoption perspective largely because of concerns around lock-in. These may or may not be lessened by the release of similar server-less services from Google and IBM, but if history is any guide, the simplest path towards dramatically accelerating Lambda adoption would be for AWS to release an open source implementation of the product.

    Whether the famously private Amazon will take such a step is unknown, but on an industry-wide basis the growth of SaaS will lead to the release of more open source software in 2016.

  • Winter is Coming:
    We may be less than a week from the end of meterological winter, but the metaphorical kind is still looming. The obvious signs of a market correction are there: an increasingly challenging funding environment, systemic writedowns of existing investments, a renewed skepticism of the sustainability and funding models of startups, and existential crises for multiple large incumbents. The less obvious signs are the private conversations, subtle pattern shifts in job hunting trends and so on.

    How deep or prolonged the next dip will be is difficult to predict at this time, but what seems inevitable is that it will start this year.


  • Google Releases an iMessage Competitor at I/O:
    Google’s strategy with respect to messaging has been perplexing of late. While products like HipChat are correctly regarded as the primary competition for Slack, it is nevertheless true that a good portion of the latter’s traction has come at the expense of Google Talk – a product which has seriously languished in recent years. Towards the SMS end of the messsaging spectrum, meanwhile, Google’s general response to the rapid growth and popularity of Apple’s iMessage has been apathy and indifference. Which makes sense if the only business you care about is search. If, however, the enterprise collaboration and mobile markets are of some importance – as Google’s actions on paper suggest they are – this inaction is baffling.

    More to the point, for every quarter they delay a response, they’re that much further behind from an adoption standpoint. Even if they were able to roll out a viable iMessage competitor for Android tomorrow, for example, they’d be facing a protracted battle to win users back from competiting services.

    Perhaps Google has come to regard the messaging market as akin to the old IM networks; superficially useful, but limited in their long term value. Or maybe they’re pessimistic about the opportunity to compete with multiple closed, defensible networks and are planning the strategic equivalent of an island hop. The difficulty with either strategy is that if the first prediction above is true, and bots are the new UI, Google’s lack of a visible, well adopted chat vector to their users is a serious problem.

    Which is why I expect Google to attempt to remedy this in 2016, the logical release for which would be at the I/O conference. Google is undoubtedly behind, but not insurmountably so. Yet. Slack is still in low single digit millions from an adoption standpoint, and Apple has artificially created vulnerabilities with its single platform approach – an iMessage that worked seamlessly across platforms and, importantly, had legitimate (i.e. not Mac’s Messages) desktop clients for a variety of desktop operating systems would generate interest, at least.

  • 2016 Isn’t the Year of VR, the Rift/Vine/etc Notwithstanding:
    A little while back I had the opportunity to demo the latest build of Oculus’ VR software and hardware. It was legitimately mindblowing. I haven’t had too many experiences like it in my time in this industry. The last portion of the demo placed you on a city street in the midst of an alien attack. Action was slowed dramatically, so you could turn your head and watch a bullet float by, or watch the car next to you detonate and lift into the air as if it were underwater, but still on fire. Insane.

    But 2016 isn’t going to be the year of VR.

    Most importantly, the equipment is too expensive. As Wired says, the problem isn’t necessarily with the cost of the unit itself, in spite of the $600 price tag (or $800, if you want an HTC Vive): it’s the total cost of ownership, to borrow the enterprise term. First, the $600 doesn’t include higher end controllers. But more importantly, it doesn’t factor in the cost of the associated PC hardware – specifically the graphics card:

    True, you can bring that cost down by going with a desktop, but how many people will buy a desktop over a laptop these days? Even if cost is addressed, it will take time to populate the kind of software catalogs buyers will need to see to justify the expense and the equipment.

    Based on the few times I’ve used VR, I’m bullish on the technology long term. But my expectations for it in 2016 are modest.


  • “Boot” projects Will Become a Thing:
    Better than ten years removed from the initial release of Rails, it seems strange to be writing about the “new” emphasis on projects intended to simplify the boostrapping process. But in spite of the more recent successes of projects like Bootstrap and Spring Boot, such projects are not the first priority for most technical communities. Perhaps because of the tendency to admire elegant solutions to problems of great difficulty, frameworks and on ramps to new community entrants tend to be an afterthought. In spite of this, they frequently prove to be immensely popular, because in any given community the people who know little about it far outnumber the experts. Even in situations, then, when the boot-oriented project and its opinions are outgrown, boot-style projects can have immense value simply because they widen the funnel of incoming developers. Which is, as we tell our customers are RedMonk every day, one of the single most important actions a project can take.

    Based in part of the recent successes mentioned as well as a growing awareness of this type of project’s value, we’re going to see boot-style projects become a focus in the year ahead, because every project should have one.

  • Open Source Hardware Becomes a Measureable Player:
    We’ve known for some time that the largest internet providers have been heavily vertically integrated, more so by the year. From Google’s custom servers to Facebook’s custom networking gear to Amazon’s custom racks and custom chips built with Intel, the web pioneers have little reliance today on external integrated products. For all that traditional incumbents have attempted to portray themselves as arms suppliers to the world’s biggest and fastest web properties, the reality is that they at best have been relegated to niche suppliers and at worst have been cut out of the supply chain entirely. Initiatives like Facebook’s Open Compute project have only helped accelerate this trend, by democratizing access to hard-won insights in high-scale compute, network, storage problems.

    Vendors have sprung up around these and other efforts – Cumulus Networks, for example – and this will inevitably continue, as the same forces that sought to excise the margin on first software and then compute continue towards networking and storage. Call it the fulfillment of the disruption that began as far back as 2014, but in the year ahead we’ll see hard impacts from open source hardware on large existing incumbents.


  • AI Will Be Turned Loose on Crime:
    For anyone who’s listened to the first season of Serial, one of the things that hits you is just how much data there is to process. From verbal statements to timelines to maps to cell tower records to email threads, it’s an immense amount of information to keep track of, even for a single victim crime. With each offense, the complexity goes up commensurately.

    Complexity and synthesis of multiple forms of disparate information – particularly tedious, numerical information – is not something that people in general do particularly well. Computers, on the other hand, are exceptional at it. With the accompanying improvements in natural language processing, additionally, it’s possible to envision Philip K Dick-like AI-detectives that can process thousands of streams of information quickly and dispassionately, rendering judgements on outcomes.

    We’re a little ways off from Blade Runner, of course – Moravec’s Paradox still holds, even if yesterday’s Atlas videos are terrifying. But purely from an analysis perspective, we’re clearly at the point where an AI could assist in at least some investigatory elements.

    What would the interest be from the AI side? Clearly not financial, because even if the system worked perfectly it would likely take a decade or more to address law enforcement and legal concerns. No, the primary benefit would be marketing value. IBM didn’t have Watson play Jeopardy for the prize money; the benefit was instead marketing, introducing the first computer to play and beat humans at a spoken language game.

    With that in mind, it’s difficult to imagine a higher profile potential marketing opportunity than true crime. Consider the transcendent success of Serial and the more recent popularity of Netflix’s Making a Murderer. What if an AI project could be a primary factor behind the discovery of a miscarriage of justice?

    It would be very interesting indeed, which is why we might see it in 2016.


  • Silicon Valley Continues to Follow in Wall Street’s Footsteps:
    My Dad worked on Wall Street for forty years, the entirety of his career. When I was growing up, this fact could be cited with something like pride. If nothing else, Wall Street was a fiercely competitive market that attracted intelligent participants. Whatever else might be said about this flag bearer for capitalism, it meant you knew how to work hard and compete.

    Today, Wall Street is a ruined term, having become synonymous with a spectacular tonedeafness, outrageous excesses of compensation and uncontrolled greed. I’m still proud of my Dad, but in spite of his time on Wall Street rather than because of it. He was, fortunately for us, the antithesis of Wall Street rather than the embodiment of it. He was never corrupted by that business, and that fact did him no favors over the course of his career.

    When I got into technology a few decades ago, I had a lot of pride in my industry, much as I’m sure my Dad did. He probably felt about Wall Street the way I did about Silicon Valley. At least initially.

    Looking around the technology industry today I am regularly dismayed by what I see. From calls for the secession of California to arguments in favor of increasing inequality to literally unbelievable insensitivity to those less fortunate, the term Silicon Valley is – in the circles I travel in, at least – becoming synonymous with…a spectacular tonedeafness, outrageous excesses of compensation and uncontrolled greed. For the first time in my career, I am occasionally embarrassed to tell someone I work in technology.

    The overwhelming majority of the people in this industry, of course, are regular, good people. It is undoubtedly a case of a few bad apples ruining the bunch. But unfortunately much the same is true of Wall Street: most of the people who work there are not members of the 1%, just people trying to get by. That distinction, however, gets lost quickly.

    We in the technology industry are running the same risk, in my opinion. Unless the excesses are widely condemned, and unless we can collectively articulate a vision that isn’t something like “we always know best” or “the homeless should just learn a computer language”, I fear Silicon Valley is headed the way of Wall Street. That most of us aren’t responsible for the appalling lack of empathy won’t matter: we’ll all be tarred with the same brush.

    I don’t expect any progress in this department in 2016, which is why it’s listed here. Alas.


  1. HTC Vive, not Vine

  2. @Danno: fixed, thanks

  3. “the sustainability of open source itself is no longer in question”

    Someone didn’t read “Who Own the Future” by Jaron Lanier. Or do you really think anyone but Google et.al. are benefiting from the open source movement?

  4. “Do you really think anyone but Google et.al. are benefiting from the open source movement?”


Leave a Reply

Your email address will not be published.