State of the Apple Watch

I am very pleased with the of­fi­cial an­nounce­ment of the watchOS 11. Be­cause every sin­gle soft­ware up­date from here on­wards is go­ing to be more im­pact­ful to con­sumers than ever be­fore. Ap­ple knows ex­actly what they are do­ing when they roll out an up­date. From the keynote you might think that the watchOS 11 is barely an up­date. How­ever in re­al­ity it is ac­tu­ally quite the op­po­site. Since the ini­tial re­lease of the Ap­ple Watch, Ap­ple has been work­ing on core func­tion­al­ity of the watch and mak­ing it less re­liant on the iPhone for its core func­tion­al­ity, not to men­tion set­ting it up and con­fig­ur­ing it through the iPhone app. Ac­cord­ing to leaks and pub­lic doc­u­ments, Ap­ple once dis­cussed bring­ing the Ap­ple Watch app to An­droid. I am will­ing to bet they dis­cussed bring­ing to other Ap­ple de­vices as well. How­ever the con­clu­sion that they came to is that it is not worth while in the long run to main­tain a dis­tri­b­u­tion of this app on many plat­forms when even­tu­ally they aim to make the Ap­ple Watch a fully in­de­pen­dent prod­uct.

No mat­ter what new Ap­ple Watch comes out they have to work with the cur­rent and pre­vi­ous sup­ported gen­er­a­tions of Ap­ple Watches. This means choos­ing what fea­tures come and don’t come to each gen­er­a­tion. It is re­ally great that Ap­ple cares so much for main­tain­ing older gen­er­a­tions, but does it hin­der the po­ten­tial of the cur­rent gen­er­a­tion be­fore a new one is re­leased? Over­all watchOS 11 is the strongest re­lease of the soft­ware thus far. At the mo­ment of writ­ing this I read that Siri face has been dis­con­tin­ued, which hurts, be­cause that is what I have been us­ing since I got Ap­ple Watch se­ries 2 as a hand down dur­ing the pan­demic to re­place the aw­ful Fit­bit that I had. In short the ex­pe­ri­ence I had with that watch got to me to wait and up­grade to the Ap­ple Watch Se­ries 7 Cel­lu­lar. I was so blown away by the func­tion­al­ity I de­cided to try us­ing it in­stead of my iPhone and it proved to be re­ally fun. I will write more about this in an­other post as it would make this post too long.

The biggest fac­tor at play is Hard­ware. Es­pe­cially from a soft­ware per­spec­tive a tech­nol­ogy be­comes only pos­si­ble when we get in­no­va­tions in the hard­ware. What hap­pens when you have lim­i­ta­tions in the hard­ware? You have to make de­ci­sions that look at the main ob­jec­tives of the prod­uct and a de­ci­sion has to an­swer a few ques­tions no­tably:

• Will this compromise core features/functionality?
• Will this be useful to the user?
• How will this combine with other established features?

Let’s be­gin with Hard­ware. Ap­ple Watch is lim­ited in terms of hard­ware space that it can use due to its in­ter­nal case space. Firstly, the biggest is­sue stems from the watch size, if you scale a prod­ucts size, you will scale the in­ter­nal size along side it. Sec­ondly, the space taken up by the strap mech­a­nism when you see that it is on both sides of the case. Un­til Ap­ple puts dou­ble sided logic boards into pro­duc­tion on the Ap­ple Watch we will be con­strained by the length and width of the in­ter­nal space.

Cur­rently Ap­ple Watch lineup con­sists of 3 sizes which means it is hard to make one SiP (Sys­tem in Pack­age) for all of them that takes ad­van­tage of the in­ter­nal space of the big­ger watches.

Why is the size of the SiP so important?

The big­ger the SiP dye the more ad­di­tional tran­sis­tors you can fit on it. This al­lows you to add more cores on the CPU, GPU and the Neural en­gine. Ad­di­tion­ally, you have more space for RAM and SSD, which means they can be big­ger and store more things. Size of RAM is more im­por­tant as it con­tributes to mul­ti­task­ing abil­i­ties and dic­tates what the data bot­tle­neck is, in other terms how much data can be processed all at once.

What do the extra cores / transistors allow SiP to do?

Cur­rently Ap­ple Watch has a two core CPU and a sin­gle core GPU, which means that it has a very lim­ited num­ber of par­al­lel processes that can hap­pen at once given its ar­chi­tec­ture of RISC as li­censed by ARM, while be­ing de­signed by Ap­ple and pro­duced by TSMC. The Ap­ple Watch dye is based of a cut down iPhone SoC that they fit all the com­po­nents as close to­gether as pos­si­ble to fit it into such a minia­ture in­ter­nal space. Mul­ti­task­ing fea­tures come from hav­ing mul­ti­ple cores to rely on. This is why Ap­ple Watch cur­rently has lim­ited mul­ti­task­ing fea­tures.

### How will Ap­ple solve this prob­lem? Ap­ple when un­veil­ing their new line up of Ap­ple Watches in the Sep­tem­ber of 2024, will show that they have elim­i­nated the small­est model size, while ad­di­tion­ally re­work­ing the strap at­tach­ment mech­a­nism. This will al­low it to cut down less on the iPhone SoC due to more avail­able space, which will al­low it to have more cores thereby be­ing able to per­form more op­er­a­tions all at once. This all con­tributes to more func­tion­al­ity in the soft­ware.

Are there any alternatives?

They could make SiPs that are dou­ble sided but that would be too ex­pen­sive for com­mer­cial pro­duc­tion and will not yield the quan­ti­ties re­sult that are ex­pected for the Ap­ple Watch sales.