Mobile, ecosystems and the death of PCs

One of the ways that tech progresses is in generational changes in scale. We had mainframes, then minicomputers, then workstations and PCs, and now mobile, and each generation brings a step change in scale. That scale means that it becomes the new ecosystem and the new centre for innovation. iOS and Android smartphones alone are now outselling PCs 5:1, not even counting tablets, and that will rise to closer to 10:1 in the next few years. So, this is the new scale ecosystem. 

Each previous generation first gained scale and capability by creating a whole new market, and then, some time later, reached the point that it could take share from the previous generation. PCs were sold just as PCs for a decade or more before they could kill workstations and take over data centres. Does mobile do the same? Does mobile kill PCs, and if so how? Or will you always need a PC to do 'real work'?

Well, the future has always looked like a toy that can't be used for real work. In 1960, real work looked like this - a typewriter, telephone and electromechanical adding machine. 

Everyone in this image was a cell in a spreadsheet - the building was an Excel file and once a week someone on the top floor pressed F9 and the whole building recalculated. That was what real work needed. A decade later, of course, the company had bought a mainframe and the adding machines were on the scrap heap. Then, in our own time, every week, thousands of VPs in hundreds of companies pull data out of internal systems into Excel, make charts, paste them into Powerpoint, write bullets and email the PPT to a dozen stakeholders. But actually, their job isn't to make Powerpoint - that's just a tool. Their job is to do something, not make slides, and that PPT should really be a live SaaS dashboard. 

So, 'real work' changes with new tools - first the tools fit the work, and then the work changes to fit the tools (see this piece I wrote in the summer for a much longer discussion of this point). 

Meanwhile, the 'toys' get better. In 2000, if you'd been making a list of tasks that were unlikely to move to the web, CAD and video editing would probably have been at the top of the list. Yet our investment Onshape is now delivering CAD in a browser - and doing cloud data management and version tracking, making it not just as good as Windows but better. The founders of Onshape have been around for long enough to be told not just that you couldn't do CAD on the web, but also that you couldn't do it on Windows and, earlier, that that you couldn't do it on a PC, which was also useless for 'real work'. And in much the same way, is doing cloud video collaboration in the browser. The things you can only do on a PC, with a native PC application,  keep shrinking. Indeed, I've argued elsewhere that it's time to invert our mental model and think of the PC, not the smartphone, as the limited device. 

So, each new computing platform will never be used for real work, but the platform gets better and the work changes to fit the new platform. In tech, 'never' seems to be 5-10 years (so does 'soon'). 

But when we say that the tools change, and hence that mobile will change, what do we mean? What does 'mobile' mean? After all, 'PC' - 'personal computer' -  encompasses a whole range of devices. Setting aside things like ATMs and machine tools, how many of the 1.5bn or so PCs on earth are actually personal, and not shared amongst a family or owned by a corporation? (Perhaps a few hundred million.) In the same way, 'mobile' is a problematic term. 

First, 'mobile' does not really mean mobile. As we all know, most use of 'mobile' devices on the internet is actually over wifi. Most people use a smartphone when there's a laptop in arms' reach. And relatively little use of 'mobile' devices is when you're walking down the street or waiting for a coffee. 

If mobile is not where you are, or the network connection, what is it? 

There's another obvious way to define this - 'mobile' means no full-size physical keyboard, a small screen and an OS that doesn't do multitasking... well, except for the 'mobile' devices with keyboards, large screens or (now) multitasking. Steve Jobs told his team an iPad was just an iPhone with a big screen. Hence the problem: what exactly is the difference between an iPad Pro and a Surface Pro? Are they both 'PCs'? We know somehow that they are different, but what is the difference that's important? How are either of them different to a Macbook? To a Lumia running Windows 10 that's plugged into a keyboard and screen? Is that Lumia different to a Mac Mini plugged into a keyboard and screen? Why? You can't use the screen size or the keyboard to define 'mobile' as distinct from a 'PC'.

It certainly isn't the performance - at least, not for much longer. An iPhone 6S beats the Macbook on some benchmarks, an iPad Air beats Surfaces from prior years and it seems pretty likely that the iPad Pro will be close enough to a Surface for there not to be much point arguing about it. The lines on the Moore's Law chart are converging for anything with a battery. The same applies to the visible differences in the software. Saying that a PC is distinguished by multitasking and multiple windows is pretty short-sighted. It's just software. It changes. 

Meanwhile, most of the form factor differences seem to me to have little to do with the essence of the device. They're a matter of superglue. If I superglue a keyboard onto an iPad and install Office, have I made a laptop? How many of those 1.5bn PCs are running applications that I cannot run on this iPad - and, to my earlier point, how many people will always need to run those specific applications? If I hack Android onto a Surface, and install Office, what exactly have I lost, or gained? Is the difference between a 'smartphone' and 'tablet' any more meaningful than the difference between a PC with a large or small monitor? 

On this basis, instead of thinking of 'tablets and smartphones' as one category and 'PCs as another, we should think of larger screen and smaller screen devices. That is, you will have something you carry with you (a 'phone') and may or may not also have something with a larger screen that stays mostly at home or in your office. In the past you might have chosen between a laptop or desktop - today you choose between a laptop, desktop or tablet, depending on what you want to do with it. That is, perhaps we should think of tablets as being as much 'PCs' as desktops and laptops are. 

If you think that all these devices are moving to sit on a single spectrum with no hard barriers between them, just what screen size you pick and whether you clip on a keyboard or not (which aren't really differences at all), this can take you to Windows 10. If 'computer' is mutable across many form factors with the same fundamental capabilities - if 'mobile' and 'PC' have converged - then a single OS that can run on any of them can seem like a good idea: take the accumulated base of the old ecosystem and spread it onto the new form factor, adjusting the UI a bit to take advantage of touch. 

Well, perhaps.

This is the  IBM XT/370, from 1983. It was an IBM PC, and so it ran MicroSoft DOS, but it also, via an extra add-in card, ran S/370 mainframe instructions (from 1970). It sold well, apparently, and was a perfectly good idea with excited customers. But it wasn't the future. 

The comparison isn't perfect - a Lumia runs the same code as a Windows 10 'PC', in theory, rather than supporting two completely different sets. But the parallel with a Lumia or Surface running Windows 10 is in the ecosystem. You can put the old ecosystem on the new form factor. You'll probably sell some of them. But is that the future, or is it a new Chevy Camaro or Mustang - a product that your existing fanbase loves but that ignores the Teslas and self-driving cars on the way? 

Taking our historical analogues a little further back, this is the Republic Rainbow. 

Developed right at the end of WW2, it remains the fastest piston-powered aircraft of its size ever built. It took piston propulsion as far as it could go, and that was better than the jets available at the time, but jets were the future, and pistons were not able to deliver anything more. 

Each generation of technology goes through an S-curve of development - slow improvement of an impractical product, then explosively fast improvement once fundamental barriers are solved, and then slowing iteration and refinement as you solve every last issue and the curve flattens out. PCs are on that flattening part of the curve, just as the Rainbow was. They get perfect because you're debugging the big things you invented in the past, and now your innovation is in the extra little things (such as the Rainbow using exhaust for extra thrust), and there are no big new innovation to debug. But meanwhile, the new ecosystem is catching up, and the curve of development and innovation for that generation will flatten out way out of reach. The new curve is crossing the old one. This is why they look simliar - this is why a Surface Pro and an iPad Pro look similar. They both exist right at the point that those development curves cross. The iPad might still be a little below, but its curve is heading up.

That is, the point that you can start to do old ecosystem things on what look like new ecosystem devices is also the point that the new ecosystem can do those things too - but the new ecosystem has 10x the scale, and the new ecosystem is just starting down the innovation track where the old one is at its end.

But what is that ecosystem, if it's not a keyboard or a big screen or multitasking? What is 'mobile'? I think there are four components:

  1. First, iOS and Android are a step change in ease of use over Windows and MacOS. Microsoft has arguably matched this on phones, but on 'PCs' all the complexity to support the old way of doing things has to stay - including things like supporting interchangeable hardware. If you like tinkering with your computer this step change is bad (just as the move from command lines to GUIs was), but it enables far more people to use these things. 
  2. Second, iOS, ChromeOS and (debatably) Android have a fundamentally better security model. This comes with reduced openness, but now that the threat is not a floppy with a virus-infected copy of Leisure Suit Larry but 500 people in a foreign country hacking your Financial Controller's assistant's child's preschool to send spear-phishing emails, that's a much more valuable tradeoff.
  3. Third, the ARM ecosystem has a fundamental power-consumption advantage over x86 and a fundamental advantage in the scale of the industry investment around it
  4. And fourth, the broader scale advantage - the ARM/iOS/Android ecosystem is moving towards selling 10x more devices each year than the Wintel ecosystem. That's a similar disparity to that between PowerPC/Mac and Wintel 20 years ago. So this is where all the innovation is - in semiconductors, in components and in software. No-one is going to found a new company to make Win32 applications (though enterprise Windows apps will be worked on for a  long time,  just as mainframe apps were).

So, a simple answer to a complex question - when we say 'mobile' we don't mean mobile just as when we said 'PCs' we didn't mean personal. This isn't about the screen size or keyboard or location or use. Rather, the ecosystem of ARM, iOS and Android, with 10x the scale of Wintel, will become the new centre of gravity throughout computing. It will take over things like IoT and wearables in one direction and, in due course, the data centre in the other, and it will push onto the desktop. The smartphone is the new sun