The idea behind Miranda July's Somebody[1] isn't necessarily a new one, but it's one of those ideas whose time is just about coming. Basically, a sort of real-life implementation of Bruce Sterling's Maneki Neko[2], the realisation that when you have a sufficient density of network connected people, you can treat them like, well, nodes in a connected network. And you can use them to route packets from one to another. It's the kind of thing that people would've had the idea for years ago - the writing's been on the wall for ages for this kind of thing - but that tech people get a bit too excited about, or only really works as an art project (which, to be fair, is what July's project is) because you need that Sufficient Density. But, we kind of have it now. It's a sort of phase change. You see it in Bluetooth-powered mesh networking chat apps like Firechat[3] - no reliance upon existing network infrastructure, just on the individual nodes having sufficient density to throw up a network. Again, mesh networks have been on the weak-signal radar for years, but they've always relied upon having some sort of extant infrastructure that can be repurposed. In Cory Doctorow's Little Brother, the in-universe equivalent of the Xbox 360 is repurposed through live boot CDs that bring up a custom, unauthorised operating system that creates an ad-hoc mesh network using the built-in wireless adapters in consumer gaming hardware.
But now - what's the density of smartphones that can run a Bluetooth stack in the background, act as a little anonymous identifier and pass off packets to each other in the wild? Or, think about it this way: given what state actors did with Stuxnet, and how it got to where it needed to go, and how many zero-day vulnerabilities went into it, what could you do with a massive network of rootable phones?
This is partly what's meant by software as a material - there's a computing substrate that's been deployed in our cities, and whether you're Team Android or Team iOS, there are enough of them out there for you to think of this as a computing substrate, a nascent infrastructure that's just lying in wait. Until recently, we haven't had the battery life, hardware and OS maturity to have persistent, low-power connections on devices, but now we do thanks to protocols like Bluetooth LE. And that's just the malware - what could Transport for London do with a smartphone based mesh network?