Om Malik, writing for the New Yorker:
Porsche’s and Apple’s design philosophies are similar. Much like the 356, the original iPhone was about defining a foundation for the future. It was different from other phones on the market — it made a rectangular touchscreen the main way to interact, displacing buttons and keypads. Now the iPhone is the essence of a phone.
This is not a new argument, but it is a very good one. There is an expectation of what an iPhone should look like and, though it has morphed in form and materials since its debut, the iPhone 7 still looks like an iPhone, and that’s right. So is this:
The seamless interaction between the technologies hidden behind the screen, the software, and our services is good design. Apple thus far has made sure that it gets most of that experience right — especially the stuff under the hood. Perhaps the next time someone criticizes its designs, we should remember: good design means your phone doesn’t explode.
I’ve written a fair bit about Siri over the past month or so: in my iOS 10 review, in response to Walt Mossberg’s piece, and in response to a piece from Stephen Hackett. I think it’s important to keep bringing it up because I think Siri is currently fundamentally flawed in its design.
The way I see it, Siri requires three streams of improvement that can roughly be prioritized in terms of their complexity and perceived intelligence. At the highest level, it should be able to maintain context over the course of several requests. That is what a good human assistant would be able to do, and it’s a request for Siri that I’ve often seen expressed in tech circles.
Something slightly less complex but, arguably, of similar value is to improve the number of things Siri knows and can do. Frustrations with Siri’s limited knowledge have partially been alleviated through the introduction of SiriKit, but it is a limited set of APIs. Trivia and news items should be more frequently updated, and that’s something only Apple can do.
But there are usability concerns that run much deeper. For example, holding up my Apple Watch and saying “Hey Siri, text Michel” will return an onscreen button that must be tapped in order to dictate my text. I’ve mentioned this previously, but I’ll bring it up again because it grates on me for a couple of reasons. First, a task initiated by voice should continue using vocal interaction because the user has indicated that their hands are occupied. Second, this is something Apple already knows because the same command on an iPhone responds with an audible prompt for dictation. On the iPhone, this is an example of good design; on the Apple Watch, it’s poorly-designed in a pretty obvious way.
If Siri is to be the interaction mechanism of the future — as is indicated by bringing it to the Mac, using it as a primary user interface for the Apple TV, and the introduction of the AirPods which can’t even adjust the volume without depending on Siri — it ought to deserve an appropriate amount of attention to its design and functionality.
Apple’s most recent hires and acquisitions indicate that, behind the scenes, Siri is being given a high priority within the company. Yet, it’s hard to square those acquisitions with Siri’s age: Apple has had five years to work on this stuff. It would be ridiculous to argue that they blew their chance — not with over a billion Siri-capable devices in active use — but there’s definitely an impression that Apple isn’t yet good enough at augmented reality and machine learning. More worrying for me is that the user interface component of Siri — a field where Apple typically excels — simply isn’t good enough.