One last thing…
All of this talk about the rise of ARM processors in laptops and desktops, and the spectacular capabilities of the Apple M1 platform, has got me thinking. There has been a significant period, some might say well over a decade, during which stagnation has been the norm. That’s not quite fair – performance has improved, battery life is better, screens are much better, Wi-Fi is vastly superior – but the fundamentals haven’t changed. The handle has been cranked and each year has brought more of the same, with small but useful increments.
In contrast, the smartphone platform, with its tablet cousin, has gone from nothing to total dominance in the same period. It remains to be seen whether the PC market can move to ultralow-power operation in a way that will succeed for the next decade and beyond, or whether Intel, AMD and Microsoft are just rearranging the deck chairs on the Titanic.
Eight months ago, I declared that the iPad
Pro with Magic Keyboard ( see issue 310, p52) was the best computer I’d bought in over a decade, and I stand by this statement. I have the staple necessities of Office, albeit in limited cut-down versions, and access to a massive list of highly capable tools covering almost every area I need.
One of the key points of this platform is that it joins together all of the devices in my life into one experience. It isn’t seamless, but it does a good job. The old mantra of “pour your stuff into our cloud file storage” works up to a point, but we need to move away from the idea of devices being endpoints, lashed together with some hacky code to try to keep things in sync. The integration of iPhone, Apple Watch, macOS and Apple TV is streets ahead of rivals, and is testament to the power of owning the entire stack.
And it’s clear that Intel can’t dig itself out of its PC chip hole without broadening the scope of its own platform. Creating Intel-badged ARM processors isn’t the solution. The same is true for Windows – it’s parked itself at the end of a profitable and stable cul-de-sac from which there is now no escape, no matter how hard Microsoft tries to tweak the user interface.
What’s needed is a true revolution. The good news for Intel is that the timing is right. Pat Gelsinger left Intel in 2009 when he was CTO and has spent the past decade broadening his horizons, first at EMC and then as CEO of VMware. He returns to Intel as CEO and surely has radical change in mind. It’s time to take on the entire computing fabric space – from tiny IoT devices up to userfacing computational devices – and look forward at least ten years.
Today’s IoT devices encapsulate everything that’s horrible in modern computing: nasty user configuration, buggy OS platforms, unknown support regimes, poor experience and almost non-existent manageability. We’ve moved on from the first exciting days of network connected devices and it’s time the entire stack grew up.
However, just making a better IoT platform is simply not enough. At what point does an IoT device become a “user thing”? Is an Alexa device with screen a computer or an IoT device? The smartphone revolution has shifted the centre of people’s computing world from the Windows desktop or laptop to something they carry with them at all times.
You might think that the solution is to create a patchwork of interconnections, based on a proprietary cloud back-end service, that allows devices that have been blessed for that particular platform to talk together. This will not be good enough. It’s now that we need an open-source, defined platform for device-to-device communication, for data, event handling and manageability – and within a properly specified security framework that keeps our data safe. I don’t care whether the architecture is ARM, x86 or something else – I’m tired of having to think about such things in the higher-performing end of the computing world. And I’m tired of IoT devices acting in an untrustworthy fashion and making me seriously worry about whether they should be on my network or not.
There’s no doubt that small, lightweight, low-power, definedfunction and cheap compute devices are the future. And the future for Intel is to get ahead in this space, but to do so in a way that doesn’t lock anybody out. Provide the hardware platform, the development tools and frameworks, along with the security and management interfaces, and allow everyone to play. Such a move would be brave, but, after a decade of being out in the cold, it would surely make Gelsinger proud to see “Intel Inside” once again on a product.
There’s no doubt that small, lightweight, lowpower, defined-function and cheap compute devices are the future