StenoWiki

I'm writing this prior to WWDC 2020 where Apple is widely rumored to be introducing one or more ARM-based Macs, or introducing (as they did at WWDC 2005) the eventual switch of some or all Mac models to ARM-based processors.

This page is not organized at the moment.

This idea has been floating around for the better part of a decade, and I've spent most of the time both believing that it would be silly to suggest Apple didn't have Mac OS X on ARM-based test harnesses or iPads in the lab already, and that it wouldn't benefit them much to do it.

I think Apple has plenty of reasons to do it and I don't think that the technological feat is all that major at this point, as much as that I don't think they care enough about the Mac enough to spend the time and effort on it.

Further, I believe that if they do, they will fail to pass on the benefits people are touting of this move to their customers. By this what I mean is that we'll see chips capable of powering Macs with lots less energy get put into smaller Macs with smaller batteries, or we'll see a Mac that might cost $200 less to mac with its new processor carry the same price tag. I also fear it might not solve the age-old problem of Macs that sit for years on end with the same hardware.

So, Apple gains a lot of choices and chooses, hypothetically, the least interesting options, or the ones that are least advantageous to its end customers.

Further, it was pointed out that Apple may end up reducing choice within its own existing product framework. That's probably fine. Apple has already played with this idea for the 2018/2019 MacBook Airs, but an amount of choice has since come back to the Air, and, more importantly, the iMac is available in configurations from 2 to 18 cores and $1099 to whatever the highest end iMac Pro costs. (The high end 27-inch 5k iMac has an 8-core CPU and a good GPU and costs around $3200-4000 or so.)

In most of Apple's product lines, the people buying at the entry level have different needs than the people buying at the top. The reason the iMac family is split into four groups of discrete systems on three unique platforms is really because what someone who only uses facebook and mail has different needs than someone running an iMac Pro. The question is exactly how to denote these options and how you make this differentiation. Apple notes processors by saying a speed, a number of cores, the Intel family it comes from, and then lists the particular kind of graphics the machine has. (Intel HD/UHD [number] or Radeon [number].)

In the iPhone world, what happens is that Apple says “here's the new phone!”, they give each processor a name and then don't differentiate between performance levels in different phones that use that CPU. They list the CPU, disk storage size, and nothing else. That wouldn't go over well with the Mac set who often differentiates between Macs based on the amount of RAM and have diffrent CPU and graphics needs. It doesn't, for me, make sense to introduce a fresh model for each of these potential configurations. That's not, strictly speaking, my problem at the moment, though.

The rumors kicked up in February or so, to approximately the same level they typically do every couple years, and then have ramped up significantly over the past few weeks.

Leading up to a leaker posting a somewhat detailed, if not blindingly obvious playbook on reddit. (Disclaimer: as of this particular edit on 6/14 I haven't read it yet but my skim of it is like “yup” especially since some of the things they point out were, well, pointed out when they happened five years ago.)

The environment of the past handful of years, especially as Intel has a little bit of a slowdown in major updates to its processors, and AMD has more or less caught up with the past few years of Intel's IPC performance, and has been selling processors less expensively, and adding cores to appeal to the enthusiast market.

However, AMD is supply constrained, and they've yet to prove themselves again in workstation and server markets, and only with the third generation of Ryzens are they suitable for

Relevant Blog: 2020-03-30 ARM on Macs Isn't That Important

There's also this idea that Intel is holding Apple back that has gotten popular over the past handful of years, and, to be honest, I would believe it even a little bit if Apple rushed to put the newest CPUs in its machines every time one became available, instead of waiting months, often years on end. Even machines with really obvious upgrade paths, like the 2014 mac mini, which sat with Haswell CPUs for four years, even though Broadwell (and then skylake and then kaby lake) had become available.

To this very day, Apple sells an iMac for $1099 with a 5400RPM hard disk, 8GB of RAM, and an old 7-th gen dual core processor – 2016's finest. Though, to add to the concept that Apple doesn't care too much about pushing the performance of its low end products, the $999 MacBook Air configuration now features a dual core i3 CPU, which is reputed to do about as well as the old Broadwell dual from the 2015-17 MacBook Air, which itself wasn't a huge upgrade from the dual Haswell in the 2013 MacBook Air, which itself was a boost from previous Airs and is basically neck-and-neck with the 35-watt M-series CPUs in the 2011 and 2012 13-inch MacBook Pros.

I think the single biggest reason for Apple to do this is because they're control enthusiasts. I question, a lot, whether or not we'll actually get tangible benefits like cheaper Macs or a return to things “we” (the Mac community at large) want, like the MMMM or the mini being both cheap and powerful again just because the CPUs changed.

I also strongly believe that Apple outright doesn't care about the Mac any more. It's nowhere near the volume and profit that various iOS devices earn, so I almost don't blame them for just sticking with Intel because at this point, it's easiest.

That said, if it comes to fruition, cheaper Macs, a 10-12-inch laptop, and a more “entry” level cheap Mac mini are high on my list of desires. I'd also be super interested in a cheap 20-23-inch iMac with an ARM processor. I'll probably make it a priority to budget for and buy an ARM-based Mac if they get announced, both because I kind of want a Mac anyway (my current thought would be to get a 21.5-inch iMac, but you have to spend nearly $2000 to get what I'd consider a “good” 21.5-inch iMac.)

Implementation Details: Gruber points out that lots of people are probably wrong on what the introductory development platform will be. The popular, and lazy, theory here is iPad Pro, but iPad Pros only have 6GB of RAM and the Mac will probably not become a touchscreen laptop as part of this, so including it on the developer hardware makes little sense. The probable answer Gruber proposes here is mini, iMac, or MacBook but with ARM bits inside.

Also, looming questions about virtualization.

Osborne Effect: JLG has a blog post speculating whether or not Apple might Osborne themselves a little bit. However, this doesn't matter because Apple is the most valuable corporation on the face of the planet and is more valuable and has more money than many, if not “most” sovereign nations.

Battery Life: Apple has a track record of not bothering to use efficiency advances to boost battery life, but rather to make machines thinner or lighter, whether or not there's a demand for that.

Thinness: Apple focuses above all else on thinness, but hasn't meaningfully reduced the footprint of the mainstream consumer notebook since boosting it from the 12/14-inch iBooks to the MacBook in 2006. Small-footprint notebooks are beloved by everyone who buys them, and most of those people say they don't care if the machine has to be a bit fatter to achieve it small footprint. (But, it probably wouldn't have to be that much thicker, especially if Apple gave up on the wedge shape in order to fit more volume for compute/cooling/battery within the maximum thickness of the machine, because maximum thickness is still the measurable component.)