More than any prior Consumer Electronics Show, CES 2019 was all about smart homes and the digital assistants inside them. I lost count of how many connected doorbells, sensors, smoke alarms, TVs and cameras I saw on the show floor this year. The same could be said of products that have Amazon’s Alexa or Google’s Assistant integrated inside them. In fact, it was harder to find products that didn’t have these voice assistants in them already, or coming soon.
So CES 2019 was a big win for the smart home then, right? Not really. At least not the way I was hoping for.
Steven Sinofsky, former President of the Windows Division at Microsoft, wrote up a detailed summary of what he saw at CES that’s well worth the read, but I want to highlight a relevant section that summarizes how I feel as well about the “smart” home:
“The core question for all of the products as I walked the floor is “are the smarts in the right place”? Does every device need all of the smarts? For example, does every device need a microphone AND a network connection? How many cameras (microphones) does one need in a home? When does privacy risk outweigh benefits? How many different times in a home do I log on to a streaming service and from how many devices? Where will I get the optimal experience for any given service and how do I even know that?”
Sinofsky alludes to something I recently mentioned when thinking out loud about smart speakers and microphones that follow us around the house; at some point, we’ll have too many input and output devices for digital assistants.
Frankly, I already do and upon my return from CES, my wife said she was starting to feel “creeped out” by microphones in practically every room of the house.
That’s why I think some type of mobile digital assistant is in the near future and also why I like my Vector robot with Alexa integration far more than most of the speakers in my home: If I want privacy, I can simply walk away from Vector. In the future with greater range, a super-sized Vector-like device could be nearby wherever I am in the house or I can tell it to go away for a little while to create a “microphone-free zone”.
And then there’s smarts part of the smart home equation. This is what I didn’t see at CES.
Today we have voice control over most of our devices, which simply enables control from wherever we are. That’s nice, but it’s still just a control function. We can program routines and scenes, but that requires a manual setup which some smart home owners simply don’t want to do. So where do the “smarts” actually come in?
They should arrive as smart homes learn our habits, can adjust the environment based on our ever-changing schedules and evolve based on our preferences. Isn’t that what AI, ML, BigData, and processing at the edge are all for? Haven’t we been hearing about those technologies for a few years now? The answer is yes and yes.
But that really isn’t happening today. And it really didn’t happen at CES 2019 either, suggesting that we’re at least another year off before it does, pending a few one-off announcements this year.
Samsung did say that it’s evolving its AI and chips because it believes that 70% of data will be processed at the edge within the next few years. Let’s see if Samsung puts its AI where it’s mouth is when the Galaxy Home speaker with Bixby finally goes on sale. As it stands now, most of the processing in our digital assistants — the heart of the smart home for those without hubs — is done in the cloud, rending the devices useless without an internet connection. In other words: They’re not very smart on their own.
Google too showed me a glimmer of hope with its Smart Clock from hardware partner Lenovo.
This $79 4-inch touchscreen with speakers and a microphone is like a junior version of Google’s other smart displays. Although it didn’t demonstrate the function, Google said the clock will suggest alarm times based on your morning and evening routines, possibly in conjunction with events on your calendar.
That’s smart, but it’s really just the tip of the iceberg in the future I envision.
At some point, we need devices that require less user interaction from a voice or programming standpoint to make them smart. Instead, by algorithmically learning our personal preferences and schedules, a smart home with predictive intelligence should optionally suggest what to turn on, watch, or listen to as well as when, where and in the best lighting/temperature climate for a truly personalized and smart home experience. Maybe that will start to happen at CES 2020.
I still say that a truly smart home will require a personal computer based master hub. It needs to talk to all of your devices, directly or indirectly; and it needs to run the sophisticated algorithms that make the House really smart. Eg. when the shower turns on early on a workday it should trigger a whole ‘getting ready for work’ routine , turning on the coffee maker, display weather and traffic data on you mirror, set the thermostat, pause and the tell the microwave to warm a bagel, and other things you do every workday morning. That takes more processing power than you find in an IoT device.
With my current DIY setup; I have to manually write all these algorithms. An AI that figures out all that will take even more processing power.
Look at the AtmosControl which will be shipping in Q2. http://www.atmoshome.com
Love your podcast, I am an avid, long time listener. In your last newsletter about your devices and upcoming move you said about your Delta connected faucet “ The Delta Alexa-enabled faucet stays because changing it is a pain, although if someone doesn’t have an Alexa it’s useless to them. This really bothers me. I hate having a connected device that no one uses. It’s not just a waste of time and space, but also a potential vulnerability in the home that offers only risk and no reward.“. I disagree, it is still a quality faucet, has the touch to turn on/off feature that ruins you for faucets that don’t. Yes, the voice functionality is lost without a $39 dot but who that would buy a smart home wouldn’t make the tiny investment in a dot to be able to say “Ask Delta to dispense 1 cup” and watch the magic! I posted here because I didn’t see a place to comment in the newsletter. Thanks for all I have learned for my smart home from you. Sorry you are leaving Texas! You are always welcome back!
Nice post Kevin. I’m hoping we’ll soon see truly intelligent and *private* smart homes. I want the benefits of IoT but I don’t want to have to worry about any of my audio, video or metada leaking or being misused. Give me a few core AI/ML modules that I can run locally on a PC or Mac or dedicated smart hub, and have that connect to the internet selectively, and only share anonymized as needed.
I work in the Smart Home industry and feel that what is really missing is some piece of identifying hardware or key that we can carry on our keyfob that allows our Smart Homes to work out exactly who’s in a room or at home.
So for example when I walk into a room, my Smart Home can start playing music I like – If it notices that my wife is in the room also then it alters the playlist to music we BOTH like – If the kids are also in the room it doesn’t play songs with explicit content.