Everybody knows now, we built the wrong internet. Instead of a democratic, participatory playground we seem to have built something predatory and exploitive. The solution—bear with me—is more awkwardness, more friction—Bluetooth!
JANUARY 2023 UPDATE: Since I wrote this I’ve been reading a lot about techno-feudalism and the disagreement between those who believe the network economy is actually not capitalism in the classical sense but a kind of revived feudalism and those who think it can still be classified as old-fashioned exploitive capitalism even though it’s highly-dependent on rents. Evgeny Morozov has a really good overview.
ANOTHER JANUARY 2023 UPDATE: in this article I propose that we dump the full-service internet and rebuild it using a parallel technology – Bluetooth. But, of course, there’s another parallel technology that might be a better candidate for the simpler, more passive internet that I propose – the home automation standard Matter. It’s an industry-wide standard, now accepted by essentially everyone (even Apple) that allows every device in your life to communicate. If you bought your washing machine in the last year or two it’s probably got Matter in it. Might be a better option than Bluetooth. Here’s an explainer from The Verge.
Here’s the problem, back then (I’m talking about the 80s and 90s—but the period of intenet hopefulness is much longer than that and really started at the end of the 60s), when we had the choice, we (techies, utopians, stupid dreamers) wanted the net to be flat, open and ‘democratic’. We wanted no hierarchies of access or quality. We didn’t want to recreate the phone network or a pre-Internet computer system—with dumb terminals dependent on powerful central machines.
We wanted proper Internet nodes everywhere—machines with IP addresses for everyone—not isolated clients hooked onto highly-connected servers concentrating bandwidth, CPU and data. We scorned centralised networks like X500, Minitel, the dial-up communities like Compuserve and Prodigy.
So we (the hopeful internet people) excitedly pushed full-service hardware out to the fringes of the network—into people’s offices, homes, pockets and cars. We wanted everyone to be properly on the net—with IP addresses, a real peer-to-peer set-up. It’s clear now, of course, that the effect of creating this ultra-flat network wasn’t the democratic one we hoped for.
The effect (and this is a very long and contested story that I won’t get into here) was to put, alongside that gorgeous, rich, full-service IP stack, a full-service exploitation stack, right there on the computer in your spare room and then later in your laptop and your mobile, then in your TV and your car and your bike and your toaster…
We lit up most of the homes on the planet, every smartphone in every settlement on earth and a trillion other devices and—let’s face it, while we weren’t really paying attention—permitted the corporations building the thing to load these devices with hardware and code optimised to own and monetise our behaviours. A happy social experiment that’s now a grim, spiralling disaster.
So, here’s the thesis: It’s much too late to re-engineer our beloved full-IP internet in a less predatory way. The die is cast, the chipsets and protocols and architectures are too deeply embedded, the miserable, exploitive business models are baked in to the hardware irreversibly. The Internet has been subtly converted from utopian playground to dystopian data mill.
But, alongside the complex, integrated, end-to-end Internet—the one we now know to be broken, salted with surveillance and extraction—there was another architecture, another arrangement of resources and access permissions, a humble standard called Bluetooth.
You’re going to have to bear with me here: Bluetooth—funny, dysfunctional, bleeping and disconnecting Bluetooth—the tech you use to connect your phone to the speaker in the kitchen—should be our model for the next, non-predatory internet.
Why? Because it’s in almost every respect less evil than those other ways of connecting devices. It’s a thinner, less ambitious protocol than, say, wifi, which wants to be transparent, frictionless, complete.
Bluetooth was designed to solve a smaller set of problems (allowing you to shout stock market trades into your phone while driving your Audi Quattro around Lower Manhattan really). And Bluetooth’s lower status and more limited goals mean that devices have, even now, never acquired the status of full network nodes. They hang off of full network nodes, servicing them, adding features (they’re like those Pilot Fish servicing sharks).
Bluetooth devices are not peers—they’re secondary nodes, symbionts—they even have a ‘master-slave’ relationship with actual network nodes. Hierarchy is visible, explicit, in the Bluetooth world.
So here’s the idea: let’s swallow our pride and rebuild the Internet on the Bluetooth model—at least the user end of it, the edges—moving users from the full-service internet out there to a more limited, less predatory platform. Let’s retain the fast backbone and the full-service bridges and switches and hubs and servers and cabinets and massive, stupid, monolithic platform applications but let’s strip out the perfect, end-to-end functionality that turns your PC into a data collection node, a surveillance device.
Let’s be clear: Bluetooth’s not ideal, not risk-free (there’s a history of exploits and there are major limitations to the architecture) but its weaknesses actually make it a good model for the next, non-predatory Internet. It’s simpler and definitely less inherently exploitable, nodes don’t automatically connect, don’t fire identifying date out to web sites and apps in the background, usually don’t have file system access, don’t share personal information and don’t expose basic hardware services to others. Apps running over Bluetooth can’t crawl freely up the stack into your address book or your social graph. They’re sandboxed, contained (I know, I know—there are big exceptions to all this).
Bluetooth is, by design, a gloriously limited technology. A Bluetooth Internet would be 50% more awkward, 50% less end-to-end, 50% less frictionless. But it wouldn’t automatically share your location, it wouldn’t pass through cookies and tokens without asking. It would make fingerprinting and triangulation harder. And it wouldn’t require the surrender of your fragile subjectivity in exchange for allowing you to share photos with your friends.
But the risks that flow from the exploitable Internet—predatory apps and platforms, ad networks that trade in user intentions and desires, your permanent absorption into corporate data trees, exploitation, ownership, control—none of these would go away, but moving the user internet to a simpler, more awkward, more constrained technology would slow it all down, make it all much harder. We’ll fix the net by embracing the old evils—friction, awkwardness, latency.
How to achieve this? How to actually flip the architecture, throw sand in the gears of surveillance capitalism? Sorry, no idea. Maybe another post.