2 June, 2015

Constellation Computers

The way we generally understand innovation is that technology gets smaller, faster, smarter, cheaper, as described by Moore’s Law. This is helpful, but also isolating. A perhaps more nuanced story is told by Metcalfe’s Law, that networks get more useful when there are more nodes connected to them. This explains everything from long distance phone networks to Facebook, but it’s also a useful framework for thinking about the next wave of holistic technological change1.

Why did the Newton, Palm Pilot, and Windows Tablet PCs become terminal branches in the evolutionary tree but iOS and Android lead to a Cambrian explosion? The simplistic, and not entirely wrong, answer is the one that fits with Moore’s Law: the component parts all got small and fast and cheap enough. A more nuanced answer is that WiFi wasn’t ubiquitous, cloud storage didn’t really exist yet, and apps weren’t even apps but software that came in boxes sold on shelves. The iPhone and iPad of course have vastly superior hardware but they also connect to an entire ecosystem that could scarcely be imagined when the Newton was unceremoniously killed.

As we transition beyond mobile, the Moore’s Law approach will obviously continue to be important in order to make these new, more intimate machines small and increasingly invisible. But the network of devices will be just as, and eventually, more important. Right now, we bifurcate these as “wearables” and “the internet of things”, but maybe it’s better to think of them as co-dependent “constellation computers”.

Much of the thinking about these early next generation devices, including my initial criticism of the Apple Watch, has been preoccupied with the hardware and features. Apple’s focus, too, is mostly about the artifact for obvious reasons: the rest of the pieces aren’t ready yet. The Watch has to be able to stand on its own to reach a critical mass, at which point it will become its own star in the constellation, alongside the iPhone, the Apple TV, and also lightbulbs, thermostats, door locks, cars, refrigerators (and the food in them!), medical devices, etc.

These connected objects are easy to mock, and are going to be slow and stumbling to roll out, but are critical asterisms to the constellation. I don’t want a watch to replace my phone, I want it to replace my wallet and my keys. That means a lot of chips and radios and software embedded in the real world.

If you were to imagine the tools for powering and connecting these points, they might look like Google’s just announced Brillo and Weave, an operating system and protocol for building constellation computers. It’s a typically Google approach, a big, open source idea without a product yet, but perhaps better suited to the challenge than top-down mandates like HomeKit. Of course, Apple’s more product-focused plans — start with the wrist, then expand to the place where people have the most control over their environment, then out to the world — makes some sense, too. Perhaps the competing designs for constellation computers will interoperate, like the web today, or maybe they’ll serve to lock us further to our chosen platforms.

A computer in isolation is useful. A computer that connects to every other computer an order of magnitude more. A pocket-sized computer that instantly connects to all the world’s information and services has been transformative. A constellation computer is as close as we’ve gotten yet to the promise of machines that enhance and empower our lives.

  1. I’m being a bit fast and loose here with Moore’s and Metcalfe’s Laws, which refer to fairly specific, observable phenomena. Moore’s Law is about the number of transistors that can fit on a chip doubling every 18 months, but it’s a useful shorthand for thinking about the pace of hardware innovation: hard drive capacities get bigger, memory gets denser, screens get sharper, radios get more powerful, it all gets cheaper. Metcalfe’s Law was an observation about nodes on a network but is similarly useful for thinking about the transformative effect of many pieces fitting together.