Skip to main content

The four phases

A fictional mainframe

This post is part of February’s IndieWeb Carnival, in which Manuel Moreale prompts us to think about the various facets of digital relationships.

Our relationship to digital technology has been through a few different phases.

One: the census

In the first, computers were the realm of government and big business: vast databases that might be about us, but that we could never own or interrogate ourselves. Companies like IBM manufactured room-sized (and then cabinet-sized) machines that took a team of specialized technicians to operate. They were rare and a symbol of top-down power.

Punch cards were invented in the 1880s, and were machine-sortable even then, although not by anything we would recognize as a computer today. In the 1930s, a company called Dehomag, which was a 90%-owned subsidiary of IBM, used its punch card census technology to help the German Nazi party ethnically identify and sort the population. (Thomas Watson, IBM’s CEO at the time, even came to Germany to oversee the operation.)

The first general-purpose digital computer, ENIAC, was first put to use to determine the feasibility of the H bomb. Other mainframe computers were used by the US Navy for codebreaking, and by the US census bureau. By the sixties and seventies, though, they were commonplace in larger corporate offices and in universities for non-military, non-governmental applications.

Two: the desk

Personal computers decentralized computing power and put it in everybody’s hands. There was no overarching, always-on communications network for them to connect to, so every computer had its own copy of software that ran locally on it. There was no phoning home; no surveillance of our data; there were no ad-supported models. If you were lucky enough to have the not-insignificant sum of money needed to buy a computer, you could have one in your home. If you were lucky enough to have money left over for software, you could even do things with it.

The government and large institutions didn’t have a monopoly on computing power; theoretically, anyone could have it. Anyone could write a program, too, and (if you had yet more money to buy a modem) distribute it on bulletin board systems and online services. Your hardware was yours; your software was yours; once you’d paid your money, your relationship with the vendor was over.

For a while, you had a few options to connect with other people:

  • Prodigy, an online service operated as a joint venture between CBS, IBM, and Sears
  • CompuServe, which was owned and run by H&R Block
  • America Online, which was originally a way for Atari 2600 owners to download new games and store high scores
  • Independent bulletin boards, which were usually a single computer connected to a handful of direct phone lines for modems to connect to, run by an enthusiast

(My first after-school job was as a BBS system operator for Daily Information, a local information and classifieds sheet in my hometown.)

In 1992, in addition to bulletin board systems and online services, the internet was made commercially available. Whereas BBSes, AOL, etc were distinct walled gardens, any service that was connected to the internet could reach any other service. It changed everything. (In 1995, my BBS job expanded to running what became one of the first classifieds websites.)

But for a while, the decentralized, private nature of personal computing remained. For most private individuals, connecting to the internet was like visiting a PO box: you’d dial in, would upload and download any email you had pending, browse any websites you needed to, and then log off again. There was no way to constantly monitor people because internet users spent 23 hours of the day disconnected from the network.

Three: the cloud

Broadband, the iPhone, and wifi changed everything. Before the advent of broadband, most people needed to dial in to go online using their phone line. Before the iPhone, cell connections weren’t metered for data, and there was very little bandwidth to go around. Before wifi, a computer needed to physically be connected with a cable to go online.

With broadband and wifi, computers could be connected to the internet 24/7. With the iPhone, everyone had a computer in their pocket, that was permanently connected and could be constantly sending data back to online services — including your location and who was in your address book.

It was incredibly convenient and changed the world in hundreds of ways. The web in particular is a modern marvel; the iPhone is a feat of design and engineering. But what we lost was the decentralized self-ownership of our digital worlds. More than that, we lost an ability to be private that we’d had since the beginning of human civilization. It used to be that nobody needed to know where you were or what you were thinking about; that fundamental truth has gone the way of the dinosaur.

Almost immediately, our relationship to software changed in a few key ways:

  • We could access all of our data from anywhere, on any device.
  • Instead of buying a software package once, we were asked to subscribe to it.
  • Instead of downloading or installing software, the main bulk of it could be run in a server farm somewhere.
  • Every facet of our data was stored in one of these server farms.
  • More data was produced about us as we used our devices — or even as we walked through our cities, shopped at stores, and met with other people — than we created intentionally ourselves.

While computing became infinitely easier to use and the internet became a force that changed global society in ways that I still believe are a net positive, surveilling us also became infinitely easier. Companies wanted to know exactly what we were likely to buy; politicians wanted to know how we might vote; law enforcement wanted to know if we were dangerous. All paid online services to build profiles about us that could be used to sell advertising, could be mined by the right buyer, and could even be used to influence elections.

Four: the farm

Our relationship is now changing again.

Whereas in the cloud era we were surveilled in order to profile us, our data is now being gathered for another set of reasons. We’re used to online services ingesting our words and actions in order to predict our behaviors and influence us in certain directions. We’re used to Target, for example, wanting to know if we’re pregnant so they can be the first to sell us baby gear. We’re not used to those services ingesting our words and actions in order to learn how to be us.

In our new relationship, software isn’t just set up to surveil us to report on us; it’s also set up to be able to do our work. GitHub Copilot learns from software we write so that it can write software automatically. Midjourney builds stunning illustrations and near-photorealistic images. Facebook is learning from the text and photos we upload so it can create its own text and realistic imagery (unlike many models, from data it actually has the license to). Far more than us being profiled, our modes of human expression are now being farmed for the benefit of people who hope to no longer have to hire us for our unique skills.

In the first era, technology was here to catalogue us.

In the second, it was here to empower us.

In the third, it was here to observe us.

In the fourth, it is here to replace us.

We had a very brief window, somewhere between the inception of the homebrew computer club and the introduction of the iPhone, where digital technology heralded distributed empowerment. Even then, empowerment was hardly evenly distributed, and any return to decentralization must be far more equitable than it ever was. But we find ourselves in a world where our true relationship is with power.

Of course, it’s a matter of degrees, and everything is a spectrum: there are plenty of services that don’tuse your data to train generative AI models, and there are plenty that don’t surveil you at all. There are also lots of applications and organizations that are actively designed to protect us from being watched and subjugated. New regulations are being proposed all the time that would guarantee our right to privacy and our right to not be included in training data.

Those might seem like technical decisions, but they’re really about preserving our ownership and autonomy, and returning those things to us when they’ve already been lost. They’re human, democratic decisions that seek to enforce a relationship where we’re in charge. They’re becoming more and more important every day.

· Posts