Apple’s recent MacBook Pro updates and “courageous” iPhone 7 have proven to be controversial, and a recent pair of news items highlights how much pent-up frustration is aimed at the device manufacturer. First, Tim Cook made several posts to an Apple employee bulletin board, detailing Apple’s thinking on thinking on both desktops and the intersection between hardware and software design. Cook writes:
The desktop is very strategic for us. It’s unique compared to the notebook because you can pack a lot more performance in a desktop — the largest screens, the most memory and storage, a greater variety of I/O, and fastest performance. So there are many different reasons why desktops are really important, and in some cases critical, to people. The current generation iMac is the best desktop we have ever made and its beautiful Retina 5K display is the best desktop display in the world. Some folks in the media have raised the question about whether we’re committed to desktops. If there’s any doubt about that with our teams, let me be very clear: we have great desktops in our roadmap. Nobody should worry about that…
From a strategic point of view, we also focus on things where software, hardware and services all come together and bring out the magic that only Apple can. That’s our secret sauce. It shows up in a lot of different places, and it’s something that we look for in new employees.
A brand new report from Bloomberg offers an interesting counterweight to Cook’s confidence. According to that investigation, Apple has made a number of strategic decisions that took emphasis away from the Mac design teams. The software team has been reorganized, such that a dedicated macOS team no longer exists and most of the software engineers at Apple are iOS first, Mac second. Hardware engineers are now asked to work on multiple projects at once, or to develop multiple versions of a potential product simultaneously.
For example, Apple engineers working on the 2016 MacBook update reportedly wanted to add a second USB-C port and Touch ID support, according to the report. Instead, Apple pushed out a slightly iteration of the same platform and a new “rose gold” color. Engineers working on Mac hardware no longer command the same attention from Jony Ive, the report said, and the appetite for leapfrogging the competition with a more daring approach to technology doesn’t always pay dividends — apparently Apple intended to introduce new battery technology with the MacBook Pro, but unspecified problems killed that option.
The Bloomberg report concludes on this note: “Mac fans shouldn’t hold their breath for radical new designs in 2017 though. Instead, the company is preparing modest updates: USB-C ports and a new Advanced Micro Devices Inc. graphics processor for the iMac, and minor bumps in processing power for the 12-inch MacBook and MacBook Pro.”
Part of the problem Apple is facing is straightforward: CPUs haven’t really scaled since 2013. I discussed this last August, but it bears repeating. The image below shows the CPUs Apple currently sells in the three-year-old Mac Pro as compared with the most recent chips available from Intel:
Now, the one thing Apple could do is introduce some high core-count chips that weren’t really available in 2013. Back then, the Xeon E5v2 family topped out at 12 cores; modern CPUs in the E5v4 family top out at 22 cores within the same 135W TDP. But that option carries a hefty frequency penalty: The 12-core E5-2697v2 has a base clock of 2.7GHz and a boost clock of 3.5GHz. The 22-core E5-4699v4 has a 2.2GHz base frequency and a 3GHz boost clock. In other words, if your workload is embarrassingly parallel and can easily scale above 12 threads, the E5-4699v4 is a better option. If it doesn’t — or if you can’t afford the $7,000 price tag Intel has hung on the chip — you won’t see benefits.
The GPU market is a markedly different story. AMD’s Polaris refresh would be a good fit for the iMac or even potentially lower-end variants of the Mac Pro, while Vega should address the Mac Pro market head-on (and of course, Nvidia has its own lineup of GPUs, should Apple be interested). Either way, the GPUs baked into current Mac systems are quite old. Apple has never put graphics front-and-center, and it failed to follow up on the graphics-centric argument it made for the Mac Pro. Instead of investing in OpenGL or OpenCL (macOS still only supports OCL 1.2, despite OCL 2.0 now being widely available), Apple has doubled-down on Metal, its own custom iOS API that it also backported to macOS. But major companies like Adobe, which do make use of GPU acceleration, have only begun to add Metal support and it’s not fully baked yet.
But I don’t think the issues Apple is facing with macOS and its various Mac systems are simply a matter of lackluster hardware cycles. Apple used to present buyers with a very simple argument: “It just works.” Those three words, however inaccurate or accurate you feel they happen to be, captured most of how Apple marketed itself to customers. MacOS was easier to use. The iPhone and iMac were designed to work together. Apple hardware was a consolidated, discrete ecosystem, with fewer choices but better integration of its various components. And while you may not have found this argument particularly appealing, there was truth to it. Apple built routers and peripherals that were designed to interface with its own software and platforms in ways that other companies rarely matched.
That argument seems much less potent now than it did in the past. Apple is leaving the peripheral business (apart from dongles, natch) and its clearly not putting the resources into macOS development that it once did. The latest MacBook and MacBook Pro don’t ‘just work’ unless you’ve got a varying number of dongles to hook up. In the event the company does refresh the Mac Pro, everyone who shucked out for Thunderbolt 2 adapters so they could plug non-Thunderbolt peripherals into the system is going to need to buy Thunderbolt 3 adapters for their Thunderbolt 2-to-USB (or what have you) adapters so they can continue to use their hardware. It’s not very Mac-like, when you get right down to it. In fact, it’s precisely the kind of inelegant-but-functional arrangement I associate with PCs.
There’s another answer to this issue, though — one that has nothing to do with Tim Cook or Steve Jobs and everything to do with the evolution of the PC and smartphone ecosystems. In retrospect, it’s clear that the iPod, iPhone, and touchscreen interfaces arrived during the last great hurrah of Moore’s Law, at a time when process node scaling and multi-core products were still delivering rapid improvements. The original iPhone in 2007 happened because it became possible for a capacitive touchscreen device to happen. Steve Jobs and Apple still deserve credit for the idea, but the idea only worked because technology had advanced to the point that it could be made to work.
Such inflection points are rare — one could argue that we’ve only had a handful throughout the entire history of computing. The first computers relied on a non-realtime batch interface, followed by command-line interfaces that incorporated video terminals and keyboards. The addition of the mouse and GUI are significant enough to deserve their own separate classification, followed by the rise of capacitive touchscreens since 2007. In 72 years (1945 – 2016) we’ve had four of these transitions, and Steve Jobs’ career happened to fall at a time when he could make significant contributions to both.
In the end, it scarcely matters. However loyal Apple customers may be in the short term, they’ll inevitably leave the platform if Tim Cook doesn’t give them reason to stay. 2016 was a distinct let-down in this regard, and 2017 isn’t looking much better if the Bloomberg report is right.