Published in Niin & Näin 4/2022. This is a machine-assisted translation of the Finnish original
Computers are universal calculators. Computational universality, formulated mathematically by Alonzo Church and Alan Turing in the 1930s, means that the set of possible computational tasks is the same regardless of the structure of the machine. Even the largest modern supercomputers or the most futuristic quantum computers cannot do anything that a fifty-year-old IBM mainframe or a small microcontroller hidden in a toaster could not do if programmed properly – given enough storage capacity, time and energy.
The development of computing capacity in respect to matter and energy has been successfully described to this day by Moore's law, discovered in 1965 by Gordon Moore[i]: every two years, the amount of electronic components (transistors) that can be crammed into the same space can be doubled. The fastest and most expensive supercomputer of the 1970s, the Cray 1, consumed 115 kilowatts of electricity, while today a microchip costing a few euros and consuming perhaps a watt of electricity may have a similar computing power.
In principle, the universality of computing could enable a computing culture in which no device becomes technically obsolete; instead, the capacities of even old devices could expand over time, as software technology develops. Moore's Law could mean that the environmental burden of computing would decrease even as its use increases. But the reality is quite the opposite: more and more energy is being consumed by communication networks and server rooms, and ever shorter hardware life cycles are creating more and more environmental problems.
This development can be explained as a Jevons' paradox. In 1865, the economist William Stanley Jevons observed that the increasing efficiency of new steam engines led to increase of coal use rather than reduction of it. Moore's Law, while exploding the resource efficiency of computing, creates an equally explosive Jevons paradox, when the computing business attempts to find as many uses as possible for ever cheaper and ever more abundant computing resources.
According to Wirth's Law, formulated by computer scientist Niklaus Wirth in 1995, software gets slower more rapidly than hardware becomes faster.[ii] Although there is some provocative exaggeration in this statement, the rampant bloating of programs is a fact. Word processors have always done virtually the same thing, but their sizes have ballooned from kilobytes to gigabytes. Of course, today's programs are in many ways better and more versatile than the old ones, but hardly a million times better. Rather, the culture of software development, with its practices and tools, favors a shameless fill-up of the entire available space regardless of its abundance.
The universality of computing tempts us to reduce the capabilities of a computer to two quantities: storage capacity and computing speed. Since Moore's law allows both to grow exponentially, it is all too easy to see the entire development history of digital technology in their light alone. Some people even go as far to think that any new developments absolutely require a quantitative increase. In this type of thinking, which I call maximalism, there is no real progress without higher computational requirements.[iii]
Of course, maximalism exists wherever the economic system values an endless increase of consumption and production, for example. In computing, however, it manifests itself in an exceptionally extreme form - as if raised to the second power. When the consumed things consist of bits, whose physical manifestations constantly shrink and multiply, the phenomenon maykj reach cosmic proportions. Thus, it can also foreshadow transhumanist future visions where humanity expands exponentially into outer space while maximizing the intelligence and lifespan of individuals. One planet simply cannot be enough!
A characteristic of digitalization is that the mindsets embedded in computing spread to other areas of life as well. For decades, the picture of the analog TV consisted of about five hundred horizontal lines, but in the age of digital TV, no image resolution can be more than a momentary phase. In the new mindset, moderation would mean the end of progress. Technological determinism and maximalism demand that we move from 4G networks to the more energy-hoggy 5G, even if the supposed progress only means doing the same old thing more wastefully, in 4K instead of HD.
Although maximalism is deeply rooted in today's computing, it is not an inseparable feature of computing. The universality of computing allows computing to be adapted to all kinds of purposes and ideologies: thus, digital technology developed in a growth-oriented society naturally gets a shape that serves growth-oriented ideals. Universality acts as the mirror of these ideals, Moore's law as their magnifying glass. In a world that values reasonableness, computing would also evolved differently.
Even in computing, values have changed over the decades, however. In the 1970s, the personal computer was largely a counter-cultural project to transform oppressive technology into something that empowers the individual. As companies such as Apple and Microsoft began to prosper, growth capitalism and transhumanism became part of this counter-culture. Barbrook and Cameron have called the resulting mindset "The Californian Ideology".[iv]
But the counterculture created its own illusions. Microcomputers have been presented as a green technology that processes "clean" and "immaterial" digital data, thus freeing it from the confines of dirty matter. This may be a reason why the environmental impact of computing has not been widely considered until rather late: Sustainable ICT as a field of research only emerged around 2007 and has failed to address the root causes of the problems.[v] A type of thinking that breaks away from growth-centricity and recognizes the limits of the planet is only just emerging at the margins, such as the Computing within Limits workshops held since 2015.[vi]
We live in a world where there is an urgent need to moderate material and energy use, and digitalization is often presented as part of the solution. However, this approach can easily backfire if digital maximalism is not recognized and alternatives are not offered. It is not enough that individuals rethink their consumption choices and governments limits the industry, if the ideology that drives industry can only pull the rope in the opposite direction.
[i] Thomas N. Theis & Philip H.-S. Wong, The end of Moore's law: a new beginning for information technology. Computing in Science & Engineering, No. 2, 2017, 41–50.
[ii] Niklaus Wirth, A plea for lean software. Computer, No. 2, 1995, 64–68.
[iii] Ville-Matias Heikkilä, Digitaaliestetiikka, ympäristömuutos ja tietokonetaiteen alakulttuurit. Teoksessa Ympäristömuutos ja estetiikka. Toim. Jukka Mikkonen, Sanna Lehtinen, Kaisa Kortekallio & Noora-Helena Korpelainen. Suomen Estetiikan Seura ry, Helsinki 2022, 268–313.
[iv] Richard Barbrook & Andy Cameron, The Californian Ideology. Science as culture, No. 1, 1996, 44–72.
[v] Per Fors, Problematizing sustainable ICT. Acta Universitatis Upsaliensis, Uppsala 2019.
[vi] On-line: computingwithinlimits.org.