Where the Ap­ple Watch could go from here

The Se­ries 4 is a gen­er­a­tional leap, but the Ap­ple Watch still has plenty of room to grow.

Macworld (USA) - - Contents - BY JA­SON CROSS

Ap­ple seems to have fi­nally fig­ured out ex­actly what the Ap­ple Watch is for, and built just the right hard­ware to do it. The Ap­ple Watch isn’t a lit­tle smart­phonelite on your wrist, but is now more nar­rowly fo­cused on be­ing a health and fit­ness ac­ces­sory that pro­vides lots of quick, glance­able in­for­ma­tion.

The Se­ries 4 hard­ware is the first true “gen­er­a­tional” up­date since the orig­i­nal Ap­ple Watch, and it’s a great ex­pres­sion of that new fo­cus. The much faster pro­ces­sor makes in­ter­ac­tions in­stan­ta­neous, smooth, and pre­cise. The up­graded sen­sors have the kind of fidelity and range to en­able all kinds of new health and fit­ness track­ing. The larger, higher-res dis­play can show

more com­pli­ca­tions and makes text and icons eas­ier to read. And the up­graded mic and speak­ers make voice in­ter­ac­tion truly use­ful.

I think we can as­sume Ap­ple will in­tro­duce a Se­ries 5 watch next year, and that it’s far too early to ex­pect an­other Se­ries 4–like over­haul. New gen­er­a­tions of mo­tion sen­sors and op­ti­cal heart rate sen­sors don’t come along that quickly, and it doesn’t make sense to change up the dis­play size and res­o­lu­tion so soon.

So where can the Ap­ple Watch go from here? I don’t know what Ap­ple’s plans are, but I have some ideas.


I hope that the fu­ture S5 SIP (sil­i­con in pack­age) will be zero per­cent faster than the cur­rent S4. The speed im­prove­ment found in Se­ries 4 is huge, and has fi­nally reached that magic thresh­old where tap­ping, swip­ing, scrolling, and open­ing apps feels in­stan­ta­neous. Ap­ple should be fo­cus­ing 100 per­cent of its ef­forts for the next pro­ces­sor on achiev­ing the same per­for­mance with lower power drain.

The ul­ti­mate goal? Build a smart­watch that per­forms like the Se­ries 4 with week­long bat­tery life. It might take a cou­ple gen­er­a­tions to get there, but each step would be a worth­while im­prove­ment. If Ap­ple could add, say, 50 per­cent longer bat­tery life ev­ery year for the next three years, each leap would en­able all new Watch-us­ing sce­nar­ios. It would be more im­pact­ful to the over­all ex­pe­ri­ence than higher per­for­mance.


It will take more than just clever sil­i­con en­gi­neer­ing to take bat­tery life from “all day” to “all week.” Ap­ple will need to dra­mat­i­cally lower power use from the dis­play, with­out sac­ri­fic­ing the cur­rent OLED dis­play’s lovely true blacks or the 1,000 nits bright­ness that makes it easy to

read out­doors. The Se­ries 4 al­ready made the switch to LTPO (lowtem­per­a­ture poly­crys­talline ox­ide) as a way to lower OLED power con­sump­tion, so more-dra­matic changes are needed to make big im­prove­ments. Work­ing out a way to dy­nam­i­cally ad­just re­fresh rate could help, but to get a re­ally big boost, Ap­ple is go­ing to need to move to a new type of dis­play tech­nol­ogy.

Ru­mors are that Ap­ple is in­vest­ing heav­ily in Mi­croled tech­nol­ogy ( go. mac­world.com/mcld). A Mi­croled dis­play would have all the ad­van­tages of OLED— per­fect blacks, great color gamut, su­per­fast re­sponse times, thin struc­ture, high max­i­mum bright­ness—but use a frac­tion of the power. It’s just too dif­fi­cult and ex­pen­sive to mass-pro­duce right now. The fact that the Ap­ple Watch dis­plays are so small makes it the per­fect prod­uct to in­tro­duce Mi­croleds on, while scal­ing up pro­duc­tion to iphone-like sizes and quan­ti­ties.


Open the Health app on your iphone, go to the Health Data tab, and you’ll see four big bold cat­e­gories high­lighted up top: Ac­tiv­ity, Mind­ful­ness, Nutri­tion, and Sleep.

The Ap­ple Watch has built-in fea­tures for the first two, but doesn’t re­ally do much for the se­cond two—those cat­e­gories are only ser­viced by apps that in­te­grate with the Health app. But Ap­ple could re­ally com­plete its mis­sion in mak­ing the Ap­ple Watch the ul­ti­mate health and fit­ness tool by com­plet­ing the square, as it were.

A New York Times ar­ti­cle from 2017 ( go.mac­world.com/frip) men­tions that

Ap­ple has been re­search­ing a way to per­form non-in­va­sive blood glu­cose mon­i­tor­ing. Cur­rent meth­ods re­quire tak­ing small blood sam­ples by pierc­ing your finger, or plac­ing sen­sors un­der the skin (cal­i­brated with oc­ca­sional finger-prick read­ings). It’s the­o­ret­i­cally pos­si­ble to use op­ti­cal sen­sors and lights—hard­ware very much like the op­ti­cal heart rate sen­sors al­ready on Ap­ple Watches—to mea­sure blood glu­cose. Pos­si­ble, but not easy; mak­ing such a thing re­li­able and ac­cu­rate is said to be “years away.” Then again, no­body ex­pected Ap­ple to stuff an ECG into the Ap­ple Watch by now, ei­ther.

Mea­sur­ing blood glu­cose is a big help to di­a­bet­ics, of course, but its use­ful­ness ex­tends way be­yond that. Hav­ing your blood glu­cose mea­sured all the time could be in­cred­i­bly im­por­tant for ev­ery­one try­ing to eat bet­ter, im­prove ath­letic per­for­mance, or lose/gain weight. Imag­ine get­ting a no­ti­fi­ca­tion at 11:30 a.m. suggest­ing you eat a low-carb lunch be­cause your blood su­gar is above av­er­age and you’re try­ing to lose weight. Or a pop-up one hour be­fore you typ­i­cally go to the gym suggest­ing that you eat a high-carb snack to give you en­ergy be­cause your blood su­gar is a bit low.

There’s no short­age of apps to try to help you eat bet­ter, but all they have to go on is the data you bother to in­put—food you eat, weight, body fat, and so on. Most of us can’t be both­ered to do much more than weigh our­selves ev­ery day, so re­ly­ing on a bunch of user-en­tered data is a real weak link. If diet apps could trans­par­ently see what your cur­rent blood su­gar is, and how it changes through­out the day and in re­sponse to ex­er­cise, it would to­tally rev­o­lu­tion­ize nutri­tion mon­i­tor­ing and rec­om­men­da­tions.

Blood glu­cose mon­i­tor­ing is go­ing to be tough to get right. Be­ing a lit­tle off on a heart rate mea­sure­ment won’t af­fect most

peo­ple, but mil­lions of di­a­bet­ics could suf­fer se­ri­ous con­se­quences if the data they rely on is a lit­tle spotty.


You can al­ready use your Ap­ple Watch for sleep track­ing. Apps like Au­tosleep ( go. mac­world.com/atsl) and Pil­low ( go. mac­world.com/pllw) track how long you sleep, mea­sure deep sleep ver­sus rest­less sleep, and more. They’re al­most en­tirely au­to­matic, but not quite as sim­ple as an in­te­grated Ap­ple so­lu­tion should be. I imag­ine that Ap­ple is un­doubt­edly, at this very mo­ment, gath­er­ing data from hun­dreds of vol­un­teers around the world who wear an Ap­ple Watch while they sleep, while us­ing more com­pli­cated and so­phis­ti­cated sleep-track­ing meth­ods as a com­par­i­son. All that data is prob­a­bly be­ing used to train a so­phis­ti­cated ma­chine learn­ing al­go­rithms.

Even­tu­ally, Ap­ple will re­lease the fea­ture with much fan­fare, say­ing you never have to set or con­fig­ure it, just wear your watch when you sleep and get awe­some and re­li­able data right there in the Health app. Maybe Siri will sug­gest you go to bed ear­lier if you haven’t been get­ting enough sleep, or could wake you up dur­ing the light­est part of your sleep cy­cle (as long as it’s be­fore your alarm time).

If cur­rent Ap­ple Watches al­ready have the hard­ware to do all this, what’s the holdup? My guess is that it’s one of two things: ei­ther the test­ing and al­go­rithm train­ing is just tak­ing longer than ex­pected

and sleep track­ing will come to all our Ap­ple Watches in watchos 6, or Ap­ple is wait­ing to make it ex­clu­sive to fu­ture hard­ware with longer bat­tery life.

I’ve had no prob­lem charg­ing my watch for about 90 min­utes in the evening and then wear­ing it all night and the whole next day, but those that do more out­door ac­tiv­ity track­ing might run out of power be­fore the day is done. If Ap­ple could ex­tend bat­tery life by just a few hours, it would prob­a­bly be enough to make sure nearly ev­ery­one could re­li­ably wear their Watch all day and all night on a sin­gle charge.


The most im­por­tant im­prove­ments to the Ap­ple Watch will come as watchos up­dates. Cur­rent hard­ware (Se­ries 4, at least) is ca­pa­ble enough to run most of these up­dates, but Ap­ple may want to make them ex­clu­sive to fu­ture hard­ware any­way. It wouldn’t be the first time it has done such a thing.

Giv­ing watch faces some sort of dim, monochro­matic, al­ways-on face so the watch doesn’t look like a black slab on your wrist most of the time—that feels like a no-brainer. To­day’s watches could do it, but I wouldn’t be sur­prised if Ap­ple was try­ing to get bat­tery life up be­fore adding a bat­tery-drain­ing fea­ture like that.

Cus­tom watch faces may never hap­pen, but it cer­tainly wouldn’t be dif­fi­cult for Ap­ple to re­lease “Facekit” tools that al­low de­vel­op­ers to de­sign their own watch faces while mak­ing sure com­pli­ca­tions work prop­erly and Ap­ple’s de­sign sen­si­bil­i­ties aren’t com­pletely aban­doned.

One of the most re­quested soft­ware

fea­tures—an­droid sup­port—is just never go­ing to hap­pen. Ap­ple has the most pop­u­lar watch in the world and still has many mil­lions of iphone users to sell to.

For the most part, I think we can ex­pect ma­jor new soft­ware fea­tures to lean heav­ily on ma­chine learn­ing and AI. The Siri watch face got a lot bet­ter in watchos 5, and as Siri im­proves in gen­eral, so too will the use­ful­ness of hav­ing it on your wrist. But more than that, fea­tures like fall de­tec­tion show how much Ap­ple can do with ma­chine learn­ing. The new ac­celerom­e­ter plays a part, but it was train­ing a ma­chine learn­ing al­go­rithm on lots of slips and falls that al­lowed Ap­ple to build very ac­cu­rate and hard to fool move­ment pro­files to iden­tify falls. The new run­ning ca­dence track­ing fea­ture is an­other good ex­am­ple.

Think about tak­ing that gen­eral prin­ci­ple—train­ing ma­chine learn­ing on moun­tains of sen­sor data to iden­tify ac­tions— and ex­tend­ing it through­out the Watch ex­pe­ri­ence. Maybe the work­out app could au­to­mat­i­cally iden­tify com­mon ex­er­cises and count reps, then use a smart al­go­rithm to watch your heart rate, fac­tor in your age, and let you know ex­actly how long to rest be­tween sets. Per­haps

Ap­ple Mu­sic could build a bet­ter work­out playlist for you by watch­ing calo­rie burn and move­ment in­ten­sity, as­so­ci­at­ing it with the songs you were lis­ten­ing to at the time, and fig­ur­ing out which tunes keep you en­er­gized.

I sus­pect that, two years from now, most of the rea­sons we love our Ap­ple Watches will have more to do with new broadly-ap­pli­ca­ble soft­ware fea­tures fo­cused around ever-smarter ma­chine learn­ing and Siri stuff, and less to do with amaz­ing hard­ware ad­vances. There will be new hard­ware ev­ery year, but with the pos­si­ble ex­cep­tion of longer bat­tery life, we’re prob­a­bly still a few years away from the kind of dra­matic hard­ware im­prove­ments we got this year with the Se­ries 4. ■

The S4 is so fast that the S5 doesn’t need to be faster. It just needs to last longer.

Mi­croled dis­plays would be thin­ner, brighter, and far more pow­er­ef­fi­cient than OLED.

The Health app pri­or­i­tizes nutri­tion and sleep, but the Ap­ple Watch doesn’t do any­thing for you there.

Fit­bit’s sleep track­ing is an ad­van­tage, but one we can’t imag­ine Ap­ple will al­low for long.

Fall De­tec­tion is the re­sult of ma­chine learn­ing, lots of test­ing, and new sen­sors. Ex­pect a lot of new and ex­panded fea­tures to come from us­ing ma­chine learn­ing on huge data sets.

Newspapers in English

Newspapers from USA

© PressReader. All rights reserved.