NYT
When a smartphone app wants to alert you to a coming appointment, a text message or some bit of breaking news from your social network, it sends a tiny flare that lights up your phone’s lock screen. Depending on how you’ve set it up, the app might then buzz your pocket like a manic bee, sound a citywide panic alarm or begin playing “La Cucaracha.”
When they were first introduced, these so-called push notifications were a creative way to let our phones get our attention; now that they buzz every few seconds, they are a constant annoyance. And the glut of notifications is just one example of a growing problem with our smartphones: They are not smart enough.
Like a bumbling concierge, your phone often tries to assist you without pausing to consider any of the basic information it collects about your life. For instance, your phone has access to your calendar, and it also knows your physical location. So why isn’t it smarter about sending you the right notification at the right time — for instance, not during a first date? Why can’t it prioritize alerts from your wife and your boss over notifications for tweets from your high school pals?
And why, long after you’ve stopped caring about Candy Crush Saga, does it keep alerting you to new developments in the game — especially at times it should know you’re in bed?
Your smartphone is the information clearinghouse of your life; it knows more about you than your spouse, your dermatologist and even your favorite national intelligence agency. Yet your phone often behaves as if it knows nothing, with each app, and the entire operating system, blithely disregarding information that should be useful in determining how to help.
Once you begin thinking about your phone’s stupidity, it’s hard to stop. Look at its address book. Why are the names arranged alphabetically rather than in order of the people you interact with most frequently, or the people you’re likely to interact with today?
Or consider the calendar. Given all that your phone knows about how you spend your day — where you are, where you’re going next and how you’re procrastinating — why can’t it suggest the best time for you to tackle each item on your to-do list?
The good news is that some of this seems to be happening, slowly. As my colleague Molly Wood wrote in May, start-ups have lately been creating a new breed of programs known as contextual apps. These aim to process information about how you use your phone to improve how it works — for instance, by arranging the icons on your home screen according to which apps you use most often, or at certain times of the day.
The Humin app sorts your contact list according to a variety of factors, including how frequently you connect with people, how well you know them and your location.
One example is Humin, a much-buzzed-about contextual program that’s making its debut in Apple’s App Store on Thursday. Humin is an intelligent replacement for your iPhone’s dialer and contacts program. The app sorts your contact list according to a variety of factors, including how frequently you connect with people, how well you know them and your location.
If you’ve just landed in New York for a business trip, you can load up Humin to see pictures of the people you’ll be meeting on your trip. You can also search your contacts as flexibly as you might search the web; look for “people I met yesterday,” and the app instantly pulls up a list.
Most of this is done automatically. Humin mines the data already on your phone and on various social networks to create its intelligent contacts, so it can determine, without your having to instruct it, how well you know each of your friends, and where you met this person or that. The result is an intriguing glimpse of the future of the smartphone; it’s as if your phone suddenly sprouted common sense.
Yet while contextual apps like Humin can be useful, I suspect that they aren’t enough to combat the routine, generalized idiocy displayed by our phones. The more I used these apps, the more I yearned for Apple, Google and Microsoft to build intelligent features directly into their smartphone operating systems, so that the entire phone can make use of the data it gathers.
Video | Apps That Predict What You WantMolly Wood tests out contextual apps that can predict what you may want from your phone depending on the time of day or the activity you’re doing.
If a phone’s operating system were intelligent, it could apply data-filtering algorithms to more parts of the user interface than any single app could. In particular, an intelligent OS could smartly filter the barrage of notifications generated by your phone’s many apps.
The phone could decide, based on its assessment of each notification and what it knows about you, whether and how to present an alert: to give it priority to ring your phone immediately, to save it for later or to dismiss it outright.
You might be wary of that idea, especially with regard to the mining of your personal information. But an intelligent operating system would not need to invade your privacy any more than your phone already does.
Ankur Jain, the co-founder and chief executive of Humin, says that the information that Humin deduces about you lives in a walled garden on your phone. “What we do is build the index local on your phone, for each person’s phone,” Mr. Jain said. “So you don’t even have to have a data signal to use it.”
Will anyone build such an intelligent OS? Of the tech giants, Google is the likeliest candidate. The company’s mission is to sort through information, and it already invests heavily in artificial intelligence, including context-aware programs like Google Now.
David Singleton, the director of engineering on Google’s Android team, said in an interview that the company was working on giving app developers ways to add more contextual information, and that eventually it might apply some intelligence to how it presents notifications. “We do see it as a core part of our mission to make this perfect,” he said.
He likened the problem to email. Once, not long ago, we were all hopelessly drowning in messages. But eventually intelligent tools like spam detection and Google’s Priority Inbox began to filter our messages, making email slightly more manageable.
Mr. Singleton cautioned that intelligently sorting notifications might be more difficult than filtering email. There are many kinds of notifications, and there’s a higher penalty if an algorithm guesses wrong; you’d be very angry if your phone didn’t show you an important alert just when you needed it (“your bus is arriving!”). It isn’t clear when artificial intelligence algorithms will be up to that mission-critical sorting task.
But I took the conversation with Mr. Singleton as a reason for hope.
As our phones merge with our cars, houses and other connected devices, we could well drown in data. There will be more apps, contacts, messages and other digital bits than any sane human has the time or cognitive capacity to make sense of.
As that occurs, the phone will have to morph into an intelligent filter; it will need to be able to figure out who needs you, and why, and decide when to demand your attention. In other words, the smartphone will have to finally start living up to its name.
