Joe Williams home

Objects of Power

Langdon Winner’s Do Artifacts have Politics? (synopsis of the essay) has been a bit of a touch stone for me in the last few months. The essay makes clear that the things humans make have both intended and unintended consequences. Some of these objects affect society by being built with expressly political design but others are innately political simply by existing and being used. The essay discusses a number of examples such as low overpasses Robert Moses built deliberately to limit public transit from reaching Long Island and hegemonic energy infrastructure such as nuclear vs more decentralized and democratic solar. However, I am interested in viewing this work through the lens of modern information technology and I think a lot of what Winner describes carries forward, the internet itself is a prime example.

Notions of centralization and decentralization are common when talking about the structure of the internet. While the underlying infrastructure of the internet started as a distributed network; today it is far more centralized due to the economics to running data centers and networks. This pressure towards centralization shouldn’t be too surprising given the military origins of the internet. Going back to Winner’s essay, Jerry Mander is quoted regarding nuclear power:

“… if you accept nuclear power plants, you also accept a techno-scientific-industrial military elite. Without these people in charge, you could not have nuclear power.”

During the creation of the ARPANET and the later internet it would not be a stretch to say if you accept the internet, you also accept a techno-scientific-industrial military elite. Without these people in charge, you would not have the internet. The same is true for many foundational technologies we use today, such as GPS. Additionally, it’s arguable that the “center” of the internet exists near the Dulles airport, it’s not a mistake that it happens to be in the US. While the internet looks a lot different today the vestiges of it’s military origins and centralized power remain.

On top of this physical infrastructure we have built the web. Just having infrastructure isn’t very useful, we need applications and code running in those data centers to do things we want. At the application layer we are faced with the economic challenges and complexity of running distributed systems. As a result internet services tend to become centralized. A visceral example is that basically no one runs their own email servers these days opting for a relatively small number of companies to do it for us. It simply doesn’t make sense for everyone to run a email server at their house. The result is a small number of companies dictating how we communicate, how we store all of our family photos, writing and anything else we want anyone, including our future selves, to experience or see.

What we end up with is the internet built with political design and innately political services running on top of it. Society set the conditions for the internet to exist as it does today and the internet now affects society. I venture that this feedback loop carries a “memory” of past decisions and biases with it, something along the lines of long-range dependence. I think sensitive dependence on initial conditions likely also plays a role, without the cold war the internet would not exist as we know it today. Conway’s Law suggests that organizations design things that mirror their organization structure, as a result the military built the internet in it’s own image. It’s clear that the physical manifestation of the early internet has tangible and lasting effects on what would later get built on top of the it and how society would use those services. The implications and politics of which now impact literally every facet of our lives.


Changing gears a little, I don’t think Winner could have anticipated how pervasive technology has become in the subsequent forty years since the essay was written but that doesn’t make the essay any less true. We all read articles reciting how our phones are always-on, networked supercomputers. Many of us feel “naked” without them, unable to connect with anyone or anything that matters. In Computational Thinking the authors describe technology, and more specifically information technology and computing, as a human multiplier allowing us to do more with less. In many cases this may be the same behaviors we have always done but now amplified in every way that matters, such as speed, effort and reach. I like this multiplier analogy because it drives home that technology is a tool that humans use for human purposes rather than it be technology only for technology’s sake. As a result technology is a powerful multiplier for our own biases, be it a bridge or a cloud service.

Additionally, the internet enables zero distribution and transaction costs and the power can’t be understated. The fact that any digital product can be created and then downloaded, installed and used by just about anyone is remarkable and carries huge implications for society. I would go so far to say that while the economics of running the cloud tend to create pressure towards centralization at the infrastructure level, zero distribution and transaction costs creates decentralization from the perspective of who can be involved and what gets created. We see this tension in the platform battles today. We have centralized platforms sandwiched between decentralized consumers and creators. Without the former the later would be hard to find, consume and pay, without the latter the former would be a vacant strip mall. While there are caveats to the current system, it goes without saying that it’s never been easier to create something and get it into the hands of someone who might use it.

So, what does all of this mean? The bottom line is biases we see in society, however egregious, can end up being reflected in the objects we build. Left unchecked these biases impact how and who use them, prolonging whatever biases and assumptions they were built with well into the future. For instance, overpasses are rarely replaced. We must also be aware that technology can affect society not only by what it does but how it’s used, banal technologies used in abhorrent ways are just as bad as abhorrent technologies used in banal ways. The combination of ubiquitous computing and zero distribution and transaction costs means we have more power than ever to multiply and distribute whatever we create. It’s important for everyone, as creators, designers and engineers, to be considerate of the impact our work on society and take an active role in checking our biases and be aware of how the things we build get used. As the old saying goes with great power comes great responsibility. If we are not considerate in our creation we will build the internet equivalent of low overpasses for future generations.

Fork me on GitHub