Digital and rationality

Digital technology, often wrongly associated with rationality, is in reality an imperfect and material technology. Our choices of digital tools are political and determine the world we collectively build.

A Pseudo-Critical Spirit

In debates, there is an increasing conflation of terms: digital, rationalization, artificial intelligence, mass surveillance, Big Tech, capitalism, social media, ecological disaster, etc. It is obviously essential to cultivate a critical spirit—that is, an informed and singular way of thinking about the world we live in. But to do this, we must precisely inform ourselves about this world to shape our vision, which is not the same as an opinion. An opinion is an allegiance to a “camp of thought,” justified by nothing more than intuition or influence, often based on power dynamics rather than conditions for autonomous thinking. I therefore wish to provide an informed perspective on the link between digital technology and rationality, to also define what is political about the choices we make with digital tools, and, hopefully, to support better choices in how we use them.

The Choice of Digital Tools Is a Political Choice

Digital technology, which has been in development for centuries, is now a set of technologies indispensable to human life for roughly the past 40 years. It has thus changed the conditions of human existence. The choices we make in using digital tools are political (in the etymological sense of the term): we do not contribute to building the same world if we use tools created by Microsoft, for example—undoubtedly one of the worst capitalist conglomerates, and this from its very origin, foundation, and core mission (this is well-documented in Walter Isaacson’s book The Innovators, 2015). We contribute to a different world if we use, for writing and communication, tools developed collaboratively by mutual aid communities instead.

Our mere use of a technology has an influence around us, even if we are unaware of it. The more a technology is used, the more, through a snowball effect, it will be adopted by others—what Cory Doctorow calls the “network effect” (in Internet Capture: A Manual for Deconstructing Big Tech, 2025, C&F éditions). Gradually, this traps users in co-dependency via these tools, reinforced by what he terms “exit costs”—what we lose if we leave the network: our photos, social connections, parts of our professional and personal reality, our digital heritage, etc.

The Influence of Digital Services on Social Spaces

The politico-economic principles underlying these technologies impact the societies where they are deployed and which they help shape. Thus, yes, our choices in using one technology or another should absolutely be informed, consciously and as much as possible, about the stakes and strategic postures of the organizations behind them. Are they capitalist organizations? In whose interest do they develop these technologies? Or are they libertarian, anarchist groups—in the etymological sense, meaning without centralized power? These approaches build entirely different worlds, and this diversity exists in the digital realm far more than we are led to believe. For example, the vast majority of the internet runs on free, collaborative, community-driven software (the Linux system), ensuring cooperation detached from commercial interests, which provides unparalleled reliability and makes this shared digital commons so robust and enduring.

The Myth of Digital Rationality

Now, let’s address the common confusion between digital technology and rationality. The widespread use of digital tools is often framed as a rationalization of humanity, a dilution of our critical thinking into algorithmic processes that erode what makes us human. This perspective is based on certain observations, but they are highly partial and biased. It is more of an opinion that cherry-picks elements from reality to justify itself.

Let’s look closer. Digital technology is a set of technologies with two aspects:

1. Hardware,
2. Software.

This is what distinguishes it from all other technologies: since the dawn of computing in 1822 with Charles Babbage’s Difference Engine, we have had fixed mechanical machines that can perform diverse operations via their own languages, as Ada Lovelace brilliantly described in her 1842 text. Digital technology thus functions through languages—multiple ones. Machine languages allow us to program them to perform useful tasks, just as trains, power plants, or plows are useful. It’s simply another level of utility, not solely tied to language but connected to the languages digital technology operates with.

Take accounting and administration, the first two fields to fund digital tools after World War II because they enabled productivity gains and greater organizational efficiency. These are not inherently harmful to humanity. Information itself is not bad. A library, for example, is a good thing—provided access is free and content is diverse.

Thus, digital technology is a set of technical tools operated by languages. And since these machines are human-made tools to assist our endeavors, there has always been reflection on human-machine interfaces: how machines can better understand us to more effectively support our needs—needs that are fundamentally human: production, connection, creation, construction, writing, etc.

Yes, technology is often used for rationalist goals. Why? Because rationalist goals enable financial and organizational developments that help dominant economic systems function better. But just because digital tools have been used (perhaps primarily) for rationalist purposes does not mean their very nature is rationality. They are merely technologies. Rationalist goals long predate digital technology—just watch Charlie Chaplin’s Modern Times (1936).

The Example of the Cinématographe

To take a non-digital example, the Cinématographe—a technology invented in 1895 to record and mechanically reproduce photographic images, creating the illusion of motion—is just a technology. It was others, not its inventors, who used it to tell stories, mostly turning it into filmed theater. It can do countless other things: capture nature’s movements, preserve traces, enable live communication, surveil, etc. Not just storytelling. So, while the choice to use a technology backed by certain industrial powers is political, it’s crucial not to conflate technology with rationality. Many artists have long created with digital tools, for example, without any pursuit of rationality.

From Minitel to ChatGPT

From its inception, Minitel was widely used for the “Minitel rose” (adult chat services). The same goes for the internet… Scientists have long created games involving chance with the earliest computers. Language games, for instance, were abundant. And machines make countless errors that can be exploited creatively. Just look at how supposedly rational AI “hallucinates,” invents entirely false information, and struggles with verification systems—because they were anything but rational, even if that was the industrial goal. But they’ve had little success, and we’re far from reliable AI. Anyone chatting with ChatGPT notices its inaccuracies. Images generated by DeepMind a decade ago were utterly delirious—a “brain” that was nothing like a child’s, but absurd, incoherent, and strangely poetic.

Toward a Poetics of the Bug

Microprocessors, the heart of our machines, have bugs—they make random errors. In fact, their price correlates with bug frequency. The most powerful microprocessors are often the same as weaker ones but tested to minimize bugs, enabling faster calculations. Error-correction systems exist to rationalize them, but at their core, they are wildly imprecise machines. Signal digitization, for example, is incredibly imprecise, whether audio or visual—hence multiplying pixels to mask the inherent flaws of digital sampling. Claude Shannon’s 1948 information theory was groundbreaking precisely because it accounted for the immense “noise” (errors) in digital data transmission, devising strategies to compensate and work around them.

So, digital technology is fundamentally material—waves, electrical or optical signals, etc. It’s just physics. There’s nothing magical, nor is there anything in “the cloud.” Digital technologies are riddled with imprecision and irrationality. This is evident even with quantum computing prototypes, which could be millions of times more powerful than current computers. Some already operate in labs at -265°C, but they produce a staggering number of errors. They are inherently irrational—which is normal, because we’re taking the natural, material world and imposing our ideas onto it, and the world resists. Even Thomas Edison’s 1887 phonograph, recording on wax cylinders, had significant background noise degrading audio quality. It was an imprecise machine, and only by increasing rotation speed did the signal become slightly clearer.

The Stupidity of GPS

In short, digital technology is inherently irrational, like all machines—imprecise, error-prone. It’s through workarounds and tricks that we inch toward some semblance of rationality. Just using a GPS makes this obvious. It’s a seemingly refined technology, publicly available for 25 years, yet who hasn’t cursed at its absurd route suggestions—or occasionally been inspired by its brilliant detours? We experience this irrationality because the machine does its best within technical and financial limits. Of course, industries sell perfection. It’s a lie, and we’re sold many false promises.

Some might argue machines will grow more rational over time. Sure, future AI-enhanced GPS may work better, but we’ll just shift irrationality elsewhere. When we ask ChatGPT to write for us, for instance, what formatting, what irrationality of thought…

Humans to the Rescue: Debugging Absurd Machines

Programmers—those who write low-level machine code and are essential to keeping digital tools functional—know this well. They spend three-quarters of their time debugging, hunting down imprecisions, aberrations, and system crashes caused by hidden flaws. Remember the Y2K bug (which didn’t happen thanks to preparation)? Or old PCs freezing with a blue screen? We encounter bugs daily—apps crashing, phones glitching. We know: digital tools are not rational, and they’re still in their infancy. Don’t be fooled into thinking they’re anything more than fledgling technologies.

Of course, marketers sell rationality, but that’s not their essence. It’s the use they’re selling. If we learn to understand them better, to use them at a lower level, we’ll see them for what they are: simple machines.

The Matrix Revolution

For a cinematic reference, take The Matrix quadrilogy (Wachowski, 1999–2021), where machines enslave humans in a simulated world. In The Matrix Resurrections (2021), the free humans underground collaborate with machines to grow food and sustain their humanist rebellion. Here, machines are used virtuously, unlike the dominant ones that exploit humanity.

Technology is not inherently rational. It’s just technology—material objects in the world, acting and being acted upon, far less rational than us, than our near-divine metabolism. For more on this dialectic between computing and nature, I recommend Marc Alizart’s Informatique céleste (2017).