Apple is set to unveil the new Apple Watch Series 6 this week. It’s also rumored to be creating a new, lower-cost version of the Apple Watch instead of just dropping the price on an older model. The new Series 6 is rumored to add other health features like blood oxygen monitoring, while watchOS 7 will bring sleep tracking and yet more fitness options.
The Apple Watch is, without a doubt, a health device first. I’m curious to see how Apple navigates announcing a feature like blood oxygen monitoring — something that has been on everybody’s mind during the pandemic.
Apple has done a very good job being clear on what the Apple Watch is and is not. It is a nice health monitoring device that can serve as an early warning system for certain conditions. It is not a medical device and shouldn’t be used as your only health device if you are at risk for the things the Apple Watch looks for. Apple has never (and I believe never will) dissembled about the difference — but right now it’s more important than ever that consumers understand that difference.
Anyway, I’m glad the Apple Watch finally found its place. For the first couple years there, it wasn’t entirely clear what it was for and it was even less clear that Apple had a good answer. Instead, it had several answers. And as this year’s Apple Watch refresh approaches I’ve been thinking about one of the answers that hasn’t come to pass: ambient computing.
My friend Walt Mossberg’s final column for The Verge was titled “The Disappearing Computer” and it was about that idea, ambient computing. Back in 2017 we didn’t have a clear definition for it (and truthfully, it’s difficult to pin one down now), but Walt had a good working model for some of the signs that it has arrived:
The technology, the computer inside all these things, will fade into the background. In some cases, it may entirely disappear, waiting to be activated by a voice command, a person entering the room, a change in blood chemistry, a shift in temperature, a motion. Maybe even just a thought.
We’ll leave the mind reading to Elon Musk for now. As for the rest, it’s not that far from what the Apple Watch can do today. The Apple Watch is explicitly a computer designed to be on your body all the time, to blend into the background and become invisible, and to get your voice commands.
The Apple Watch, not the HomePod, ought to be Apple’s primary device for ambient computing. It is better suited to that task than the devices any other company sells right now. Smart speakers from Amazon and Google sit in your home, but they haven’t found an ambient computing foothold outside your home. Sure, you might have the Google Assistant on your phone or in your headphones, but it’s still too phone-centric.
By now you’ve probably already guessed the fly in this ointment: Siri. Apple’s digital assistant simply can’t be the platform that’s necessary to unlock ambient computing. Alexa and the Google Assistant aren’t quite ready either, to be fair, but they’re both much further along that path than Siri is.
The person presumably tasked with closing that gap is John Giannandrea, who led up search and artificial intelligence at Google until Apple snapped him up in 2018. Giannandrea recently spoke with Ars Technica and revealed he created the team that applied machine learning to the iPad’s Apple Pencil recognition algorithms so it could have lower latency and better recognize handwriting — something Google was already doing with Chrome OS and Samsung just started doing with the Galaxy Note.
On the iPad, it works: both latency and handwriting recognition are much better than I’ve seen before. The “scribble mode” in iPadOS 14 isn’t quite up to the task of replacing a keyboard entirely, but it’s great for short bits of text.
I bring it up not to draw a line from this application of ML to generalized ambient computing, but to point out that there is a lot that can be done with the ML and AI tools already in everybody’s tech workbench. They just need to be applied in new and clever ways. Using ML to improve handwriting and latency on the iPad isn’t a sea change, but instead, it’s a step in the right direction.
Stepping in the right direction is what Siri needs right now. Even on the new beta for watchOS 7, I still can’t ask Siri to do something basic like set multiple timers. It’s actually ridiculous! If you set a second timer, the first one invisibly gets canceled without any indication it’s gone. It’s the thing that keeps the Echo dot in my kitchen.
Harping on multiple timers in Siri — and harping on Siri in general — can be seen as making too much of small complaints. But on the flip side, Apple is well aware of this complaint and has been for some time, yet hasn’t fixed it.
That’s troubling, frankly. Apple could — and does — add great new capabilities to Siri on a regular basis. But nobody will ever discover those capabilities if Siri duffs the basics on an equally regular basis. And it still does.
The first Apple Watch was like the first beta of Siri: a mess. The Apple Watch never had a moment where it was “fixed,” but instead was fixed slowly over time via relentless iteration and improvement. In theory, the same should apply to Siri, but it hasn’t happened at the same pace.
Even though Siri isn’t where it ought to be, the Apple Watch is still the best smartwatch on the market by a wide margin. But if Siri could do more, the Apple Watch could be something more. The era of ambient computing is still coming. Will Siri be ready?
Towards the end of his interview with Ars, Giannandrea talked about hiring new talent for his team. “I guess the biggest problem I have is that many of our most ambitious products are the ones we can’t talk about and so it’s a bit of a sales challenge to tell somebody, ‘Come and work on the most ambitious thing ever but I can’t tell you what it is.’”
That sounds great, but my advice is to start talking sooner rather than later. And the “talk” I’d like to see is actually action: new features for Siri that arrive via the same relentless improvement of the things people are trying to use Siri for today. And hey, maybe start with timers.