State of Ubiquitous computing 2016

Image credit: Black Mirror on Netflix

What are some key learnings this year about wearables and connected things? Here are my key learnings within ten areas of Ubiquitous Computing that keeps me excited and concerned some of the time.

The promise of Ubiquitous Computing is counter intuitive, when simply defined, it sounds like a future resembling Netflix’s show Black Mirror where computing as we know it is pervasive, in the way, and overwhelming. In reality, ubiquitous computing helps make computers more useful. Pervasive or ubiquitous computing means more connected things with compute power that is secure, private, contextually aware, extensible resulting within i/o that blend in, get out of the way, and maximize utility for the user without complexity.

For me, a world where my “things” knows who I am, securely, and presents information to me when I need it, through a natural interface, means less time staring at the black mirror in my palm or fumbling through remote options for my TV. This singular idea has been my guidepost for things I work on professionally and personally. This year, I have been curating my interest in this space as collections of posts on Google+. To date, these posts have amassed a decent number of fellow enthusiasts and readers. Here’s how the numbers breakdown for followers in each collection:

• Connected Things — 34,415
• Connected Mobility — 90,356
• Wearables Weekly — 75,994
• Connected Health & Fitness Innovation — 43,588

So many internet points

With a quarter million followers across the various topics I post, it comes as no surprise then, that in 2016, I’ve had the opportunity to give quite a few talks and shared a ton about connected cars, wearables, and IoT. If you follow my posts, attend my talks or listen to my podcast, thank you! You fuel my passion in this space.

10 things learned about the State of Ubicomp, 2016:

This year we heard a lot about machine learning, self driving cars, VR, AR, and Amazon finally shipped an assistant in our living rooms that we actually want to use. A feat Microsoft has been trying to achieve since the mid-2000’s with Windows Media Center.

It’s an understatement to say a lot happening in ubiquitous computing falls within the ‘emerging technologies’ world. As far as maturity is concerned, it’s good to look at Gartner’s hype cycle to notice that most of the tech in this space is just getting over the hump of the “peak of inflated expectations”. I think this means 2017 is ripe for continued industry consolidation and the trend of startup exits and failure will increase. There is lots to learn.

Gartner’s hype cycle

1. State of the market: Special Purpose vs. General Purpose

We learned this year that there can’t be one market leader across the ubiquitous computing space. It’s just too broad. It must be stated that even within the various verticals that make up ubiquitous computing, there won’t be the zero sum conglomerates as we’ve had in the past. Being a cog in the wheel, is not just a challenge for larger companies, the innovators dilemma affects startups now. Winners may earn linchpin status, but there’s an entire wheel moving so rapidly, constant adjustments need to be made to keep up. There are more technology startups than we’ve ever seen creatively working on how to make computers more aware, personal while keeping utility clear and simple. Industries within Ubicomp have accepted that the pace of innovation is lightspeed. Large corporations are adapting with smaller teams, bigger spend on R&D and a copy of Jake Knapp’s Design Sprint book on a mantle.

The rule this year is to keep up fueling that innovation center or partner. This just how the law of accelerating returns works.

The singularity is near, deal with it

The concept of one platform to rule them all, is being left behind with the PC era. Within each technology sector, I find it difficult to believe one company, alone, will dominate as IBM, Apple, or Microsoft once did.

A specific example: Intel’s role as a general purpose chip maker does not translate well in the special purpose computing era. Intel is a general purpose chips maker, this year are struggling to gain a foothold across the many sectors that dominate the Ubicomp space; Wearables, IoT, self driving cars etc. Meanwhile, ARM continues to crush it as it’s market share, licensing business model, and industry position in special purpose computing grows year over year. ARM’s business model has spurred many leaders like Broadcom, Apple, NVIDIA and others in the world of Ubicomp.

Special purpose won the Ubicomp market 2016. I think this is a trend.

2. The world is changing

To realize a world where our silicon powered things help us proactively, that is, without too much input from us, computer input and output as we know it, has to change. There has been a spike in discussion about the dangers to the lives of middle class, this rapid pace of change will have. It’s not all bad though, some changes will also improve the lives of users, while drastically blurring the physical and digital world, as well as transform industries and the way we live and work. As computers become more pervasive, the more our economies will change. This is not something I learned in 2016 necessarily, just something I’ve been reminded of frequently.

See also: Humans Need Not Apply, by CGP Gray.

People losing their jobs to robots is no laughing matter, sorry.

3. There's no place like ::1/128

The smart home is the front page of ubiquitous computing. Many have tried over the years. We keep learning over and over that Hollywood-style interfaces look great on the big screen but don’t translate well in the home. The trick with a seamless connected thing is a contextual experience that maximizes utility for the user without complexity. Amazon, Samsung and droves of startups dominated home consumer electronics. This may have been expected to some, however the winning interface for the home has been voice. The success of Amazon’s Alexa, with over 1,000 skills and counting, was due to the platforms near-flawless execution of natural language understanding (NLU) and automatic speech recognition (ASR). How awesome would it be if your car infotainment system worked that flawlessly? Funny you should ask, Samsung hedged it’s bets with a lot of Intelligence of Connected Things acquisitions. With an increased pace of acquisitions of startups like Viv, Harman you now have Amazon Echo-like conversational services, powered by Samsung when heading home.

We are still not there yet; brands like the Wink Hub haven’t yet made it into general lexicon but Apple has made some strategic partnership bets to make the millions of IoS devices and connected things out there secure and extensible under a simple, familiar user experience.

Image credit: Internet of Shit by Motherboard

4. Wearables balancing special vs. single purpose

Users overwhelmingly voted with their wallets on what type of wrist worn wearables they preferred this year. When it comes to market share, Fitbit proved that users want special or single purpose wearables as it continued to dominate Apple Watch and Android Wear for wrist real estate. The realization that consumers want specialized wearable experiences also brings me back to my earlier point about Intel. Smartwatch shipments are falling 51.6% YoY as of 2016 Q3 the only logical explanation based on data, is that users more utility for the battery life we get. 2016 has been promising though with the first of many special purpose W1 hardware chips for wearables shipped this year with Apple’s Airpods. In time, industry will continue this trend, building upon specialized low-power chips vs. repurposing miniaturized chips that haven’t been optimized for wearables.

Also, specialized vs general purpose in the wearables space is my reason why Snap may succeed where Google Glass failed. Now, here is Robert Scoble in the shower, again.

Image credit: Business Insider

5. Pervasive Computing, brought to you by the cloud.

Microsoft, IBM, Google, NVidia for leadership Cloud, Machine Learning/Car Intelligence even Firefox when it comes to the rapid pace of the WebVR platform. Microsoft’s cloud business led the tech juggernaut’s earnings and their efforts in AI is always a welcome opportunity for independent developers, academics, and startups alike. Standing on the shoulders of giants never felt so good. As a matter of fact, there has been a spike in partnerships across industries focused on machine learning, and connected mobility. In order to keep up with the pace of innovation, car manufacturers are adopting new business models, and accelerating the rate of partnerships with technology companies like Microsoft, IBM, Google, NVidia and others for example.

For companies like Tesla, owning the bits that make your car see, hear, and move autonomously makes more strategic sense as it severed its relationship with MobileEye this past summer. I think the former will be the norm, where companies will “stick to their knitting” and leverage the strengths of others where it makes sense. I should mention, Tesla still maintains many strategic partnerships like it’s energy partnership with Panasonic.

6. Security:

Now more than ever, it is important for industry to rally around IoT standards. It is ever so critical for devices to be built from the ground up with security and privacy in mind. If the hardware and platform has to have network access then the standards or legislation should ship with a capability of digitally signed updates. OTA’s on connected things in the home or enterprise should not be an afterthought.

Friends don’t let friends buy cheap internet toasters

An unfortunate demonstration of the exponential damage un-patched ‘things’ can cause, happened with Dyn and the Mirai botnet. If you’re reading this, you were more than likely affected when in October, this singular event exposed many device manufacturers weren’t implementing standard security practices like the enforcement of personalizing factory-default usernames and passwords during new device activation. Most of the devices affected were off the shelf web cams, DVR cable boxes and every day connected devices. We all learned that as an OEM, if you intend to ship a device that needs internet access, then the device needs a level of security that gives the user comfort that their information is private, is intuitive to secure and interoperable to update. This reminds me, I need to disconnect my toaster.

It is going to difficult “put the toothpaste back in the tube” as it pertains to the the state of security in the “connected things space.” One reason many devices weren’t built with security in mind is due to the costs related to development and maintenance through continuous updates. I think, many will agree that solutions like Google’s Weave may take the lead or heavily influence what may become industry standard.

Privacy? What privacy?

7. On privacy

I have more questions than suggestions or answers with privacy pertaining to connected things. Gartner predicted there’d be 6 billion connected devices in 2016. That’s a lot of smart things gathering what you may consider private information within your home. In my view, recent world events regress efforts in the internet privacy space and this affects ubiquitous computing, bigly. From a macro lens, the U.S. and U.K bought into fear uncertainty and doubt. Some of this FUD was championed by rhetoric, populism and the economic transformation this very industry brings.

American citizens decided to elect Trump, a candidate who sides with the government’s justification to defeat encryption. Across the pond, the U.K. exited from the EU, and brought in a Prime Minister who just passed the so called Snoopers Charter, a bill that makes the Patriot Act look like child’s play. With this, I say, bend over folks, our privacy, as it pertains to the millions of connected things with our data, is royally fucked!

Before our liberty burns, however, my hope is that solutions like Apple HomeKit or Google’s Brillo platform, may set the stage for developers to easily solve problems with privacy in mind.

8/9. Interoperability and Extensibility

The security scares we suffered this year exposed an even more concerning flaw; the more “things” get connected, the more vulnerable we are. In recent years, the market has pushed so called smart TV’s and DVRs with Netflix apps that never get updated. Building interoperable things that talk to other things is expensive and the underlying platform architecture is even harder. OEM’s have failed to date with a secure unified standard. It’s like everytime I decide: “I’m choosing with Zigbee” to manage my network of things, another standard comes along.

XKCD: How standards poliferate — https://xkcd.com/927/

Even Google can’t get their head around it. Google acquired Nest who decided none of these protocols worked and so, they invented Weave. After acquisition, Google starts working on Weave. Sounds sane, until you realize the Nest Weave is different from Google Weave and will continue on different development tracks.

Everyone wants their own standard and these ecosystem fiefdoms will continue to make it difficult for interoperable things. Until we figure out the security stuff, maybe that’s a good thing.

10. The best interface is no interface

One of my favourite thought pieces by Golden Krishna on interfaces actually advocates no interfaces. This year, I got use to interfaces based on natural metaphors. The screen on my Android Wear watch for example, only turns on when I lift my wrist towards my face. How awesome is that?

The only intuitive interface is the nipple

Humans are slow to learn new interfaces, we think we want Minority report but in the home and at work, we are practical and rely on existing metaphors and behaviors we’ve come to accept as natural.

Keep the buttons, give me Alexa for my car.

It’s hard to create interfaces that are easy to use and easy to create interfaces hard to figure out. Even Tesla’s voice actions are shitty, and no, a 17 inch touchscreen doesn’t make it better. Keep the buttons, physical or digital, give me Alexa for my car. Samsung, you better make that Harman deal worth it.

If Tesla, can’t execute and Apple can’t either, maybe Google will help us? This year Google kickstarted 2016 with the Ubiquity Dev Summit at the Strand in San Francisco CA. I was fortunate to attend and to be inspired with not just what Google was working on but where the industry as a whole is heading. Maybe I’ll do a review of how they are doing, but here is a look at some of the ubiquitous computing platforms I’m interested in seeing evolve next year:

• In car: Android Auto, Apple Carkit, Blackberry’s QNX
• Wearable Platform: Samsung Tizen, Android Wear, GarminIQ, WatchKit
• TV: Apple TV, Android TV, Amazon Fire
• Sensors: iBeacons, Physical Web, Eddystone (UUID/URL)
• Developer platforms/Context:
• Google Awareness API, Nearby API,
• Project Tango, Windows 10 (Hololens),
• Amazon Alexa, Google Home/ChromeCast
• Interfaces:
• In-ear assistants like the Moto Hint+, products from Vuzix, Epson, HiMax, devices and interfaces like Google Glass (there is still hope) MagicLeap, Oculus, HTC Vive, Hololens, MetaAR, Cardboard, Daydream/Tango

Lost Explorer Podcast with Chris Jenkins

So where do we go from here? The mission of ubiquitous computing is counter intuitive. It requires infrastructure, platforms, ecosystems, stacks, frameworks and hardware to work in perfect harmony just to deliver things we need when we need it. To sum up we accomplished a lot but there is still a lot to be done next year with:

• Security and privacy in a Trump’ian and Brexit world
• Context, awareness and the intelligence of connected things
• Interoperability and extensibility
• Simpler, more useful interfaces that delight and the fall of apps.

For good or ill, these are a few of the many reasons I continue to be geeked out about emerging tech and ubiquitous computing. It is also why I try to inspire more minorities to get into science, tech engineering and design.

Resources

My Lost Explorer PodcastNotable thought leaders on TwitterFavourite tech podcasts, pretty much anything on TWiT.tvhttps://medium.com/@nobleackerson/ode-to-ubiquitous-computing-5eb772459ca1#.fnvlh7vfs