A quiet shift to a new architecture

It's fascinating to read between the lines of what's happening across the industry right now in regards to the types of processors we use every day. Intel has largely been the dominant CPU player for some two decades, but there's increasing discomfort with the company's lack of meaningful progress or iteration in the past few years from almost every company that relies on it.

Almost all of the chips out there that power smartphones are based on ARM architecture, which is fundamentally different from Intel's in many ways: it's more battery efficient, supports instant-on devices, produces far less heat, and so on. That's what makes it a good fit for the iPad, and other mobile devices, too.

If you've been waiting for all-day laptops without compromises or some mobile-only OS, this is essentially the step change that needs to happen to get there. Thinner, lighter and more powerful are currently oxymorons, shrouded by Intel's ever more confusing marketing of its processors, hiding the fact that it's struggling to make meaningful performance improvements.

A huge point of contention has been shifting desktop-based workflows to ARM architecture: it's no trivial job, but meaningful jumps forward are quickly arriving, even if there's no loud yelling about intent to move away yet: nobody wants to prod the Intel bear until they're sure the monopoly is no longer required.

First, there's signs that the upstream cloud industry is looking at a shift to ARM, with companies like Cloudflare declaring their intent to shift over entirely to the architecture within just a few years, already making meaningful progress toward making the jump.

Then there's the client-side: Apple has long been rumored to be shifting away from Intel in its Mac to in-house silicon, but Microsoft is even shipping devices today that experiment with the ability to run any software, without modification, on ARM-based devices. At first, performance wasn't very good because the software was emulated, but it's getting to the point where it doesn't matter much at all.

Generally speaking, if you want to shift to a new architecture like Apple did from PowerPC to Intel a decade ago, developers need to entirely re-write their applications to make them work (or suffer slow emulation of their products). That's a hard sell, particularly if a single platform is moving, and tends to take a long time. That's what makes the emulation play by Microsoft so impressive, while it works away in the background on selling developers.

Now, we're seeing some even bigger indications that the shift may be closer than expected: software vendors are actually getting onboard with it. Google is quietly working closely with Microsoft to port Chrome natively for ARM devices, bringing one of the most important apps required for success across, and others, such as Canonical, the creator of Ubuntu, have created native versions of their development tools.

It's really interesting to observe from afar: this is a complex game of 3D chess unfolding as all parties involved pretend like they aren't really doing this. Spooking Intel might lead to a disadvantage if done too soon, but suddenly shifting architectures would be incredibly painful too... unless the entire industry, from Mac to PC, did it at the same moment.

Microsoft is trying to quietly ease developers into it, but it seems closer than expected based on these new developments, particularly if the company is able to gather a few flagship apps like Chrome. I suspect in the next year, we'll see much more aggressive pushes to make this the mainstream, and very loud moaning from Intel as soon as it happens.

Why you should care: all-day battery life, modern operating system architecture, and maybe a renaissance in the form factors of the devices on the market will arrive when ARM does (or, alternatively, if Intel is able to suddenly start delivering). Best of all: no more fans screaming when you open a few Chrome tabs.

Tab Dump

No surprises here: Zuckerberg won't step down as chairman of Facebook
I don't know why everyone keeps insisting he should do this when he literally controls so much of the company that it would be a stupid move to do so. This literally isn't happening, unless he's forced out by a law change.

Snap is planning a new version of Spectacles with two cameras
Yes, the company plans another new version of its wearable glasses, despite the fact that its social platform appears to be imploding.

Instagram plans to kill off fake followers
Facebook has promised the magic of machine learning is going to start tackling the problem of fake followers and likes on the platform. It's an issue that's become hard to reel in, because what defines fake anyway? How do you detect that?

Bitcoin-rigging investigation is focusing in on tie to Tether coins

Good long read: the tiny town experiment that could bring internet to millions