Maintenance future ⊗ A responsibly limited AI

The two featured pieces this week make the case for limits. Christopher Butler describes a maintenance future, a society that holds steady instead of growing, and Justin McGuirk revisits Ivan Illich’s argument for limiting our tools and applies it to AI. Since The Limits to Growth—whose co-author Donella Meadows later wrote about choosing the right leverage points—the debate has mostly been about limiting growth itself, or a given technology. I’d pick a different leverage point: limits to capital, and thus on the growth and power of concentrated wealth, which shapes how much we grow and which tools get built, and so reaches into who controls big tech, inequality, governance and our response to the climate crisis.

Maintenance future

Christopher Butler makes the case for a “maintenance future,” the neutral position missing from most visions of what comes next. He sorts future narratives by a quantity rather than a tone: there will be more, or there will be less. Each can be utopian or dystopian, which gives four possible futures. A future of less can be a utopia, a correction of a material standard that ten billion people could never share. A future of more can be a dystopia, which he thinks is where we already are: growth goes to a few, and “more” mostly describes the gap between them and everyone else. What none of the four describes is a future that holds steady, with neither more nor less, and that is Butler’s maintenance future. It can be good or bad too, depending on who decides: chosen, maintenance is a practice; imposed, it is a lid.

The usual objection is that a future that holds steady would put an end to invention, since we think of creativity as making something new. Butler answers with Mierle Laderman Ukeles, whose 1969 manifesto for maintenance art asked who had decided that the work of keeping things going wasn’t creative, and with a point about matter: everything we make was something else first. In a maintenance culture, design would favour recovered materials and objects made for a second use, and limits have always redirected invention more than they reduced it. He thinks the change could begin with taste, once the synthetic materials we associate with the future start to look like tomorrow’s waste, and once we see how much of what we own depends on fragile arrangements.

Butler also writes about which stories we keep and admits that a story he has often repeated, about an English college whose oak beams were replaced from a grove planted centuries earlier for that purpose, is probably untrue, and that we remember it because it is shaped like a rescue. The Japanese shrine of Ise Jingū is a true story of maintenance: it has been rebuilt every twenty years for some thirteen centuries, no larger and no better each time, and a forest it began planting in 1923 should supply all of its timber by 2123. Rebuilding so often looks inefficient, but it is how each generation of carpenters learns the building from the last. The Limits to Growth met the same fate as the college story, since it described a possible transition to equilibrium and what we remember is the collapse. For Butler, the way in is fear, the recognition that what we have is fragile, and contentment is cultivated afterwards, through the practice of keeping. In all his examples, what gets maintained is a capacity, not an object. His postscript doubles as a reading list on the economics and practice of maintenance: Boulding, Daly and Schumacher, Vinsel and Russell, Shannon Mattern, and Le Guin.

One pile says there will be less. Fewer objects, less wealth, less freedom, less autonomy — a future in which people have lost things they were accustomed to having. Apocalypse is the loudest version but not the only one; most of what reads as pessimistic futurism is simply an expectation of subtraction. The other pile says there will be more. More technology, more objects, more wealth, more reach, and nearly everything else we file under progress. […]

A maintenance culture doesn’t withdraw the brief for making things, it rewrites it. Get better at recovering materials. Prefer the ones with a natural lifespan and some plausible relationship to the world they came out of. Design for the second use. Timber instead of concrete, to take the plainest case. […]

At a much longer interval the knowledge would have to be contained in some other way, and history shows us how fragile that is. What appears inefficient — rebuilding a structure every twenty years — is the opposite. It is the most efficient means of knowledge maintenance, wiser even than a five hundred year advance on one cycle of timber. […]

The third [conclusion to *The Limits to Growth] was that the sooner the work began, the better the odds. Half a century later, the collapse is the part we still argue about. The equilibrium is barely remembered. One of the most-cited futures documents of the last century contained a maintenance future, and we filed the crisis.

What is a “responsibly limited tool” in the age of AI?

Ivan Illich would have turned 100 in September, and in the Los Angeles Review of Books Justin McGuirk revisits his case for putting limits on our tools to see what it says about AI. In Tools for Conviviality (1973), Illich used “tool” in a broad sense that covers machines as well as institutions like schools and hospitals. His argument was that past a certain size, a tool stops serving the people who use it and starts making them dependent on it (ring a bell?). What he called convivial tools are the ones kept small enough for their users to stay in control. Proposing limits on technology has been quite taboo, and McGuirk notes that this is changing now that the heads of Anthropic and DeepMind say the AI race is moving too fast and ask governments to slow it down. Setting limits wouldn’t be easy, since digital tools are built to keep changing and any ceiling would require agreeing on how much is enough, which is Butler’s line above, between maintenance chosen as a practice and maintenance imposed as a lid. Climate policy shows it can be done, at least on paper, with carbon budgets and planetary boundaries.

McGuirk dismisses the claim that AI is “just a tool,” which he reads as a way of avoiding the question of its social consequences. Its own promoters predict it will eliminate half of white-collar jobs, which would make it the first technological revolution aimed at educated professionals. All the while governments still worry more about falling behind than about restraint. In Illich’s terms, AI is a tool that replaces people instead of extending what they can do, and McGuirk thinks he would have recognised its effects: weaker critical thinking among regular users, memory handed over to machines, huge energy demands and a race to monopolise knowledge. Illich would complicate the picture in one respect, though, since he had little regard for the bureaucratic work AI is most likely to replace, the kind David Graeber later called bullshit jobs.

The harder question is what people would be freed up for, if machines can also draw, compose and write. Here McGuirk finds Illich’s answer somewhat technocratic, as if the problem could be solved by fixing the tools once there is enough political will. He contrasts it with Pope Leo XIV’s Magnifica Humanitas, which grounds the case in human dignity instead of human capability. Illich did have his own idea of dignity, drawn from the vernacular lives he encountered in Latin America. In Shadow Work (1981) he valued subsistence, mutual aid and self-reliance over consumption, and he practised a frugal life of attention and limits, which he saw as the basis for “friendship,” a word McGuirk says we would now replace with “care.” The book later shaped commons movements in Mexico. He never offered a road map for a convivial society, but McGuirk reads his call to retool as potentially radical, because it would reverse whose interests our tools serve. As David Cayley, one of his best interpreters, puts it, “The question essentially is who’s on top, the user or the tool.”

Digital devices are not like mechanical machines; they are mutable by definition, designed to evolve in our hands. A smartphone may feel as tangible as a slab of basalt, but this belies a shape-shifting operating system that is hell-bent on either self-improvement or what Cory Doctorow calls “enshittification.” […]

You had Herbert Marcuse claiming that humanity was becoming “one-dimensional,” Jacques Ellul railing against “la technique,” Lewis Mumford decrying the “megamachine,” and Günther Anders going furthest in arguing that we feel a “Promethean shame” at our inadequacy compared to our technologies. […]

The Club of Rome’s The Limits to Growth (1972) was published the year before Tools for Conviviality and seemed, erroneously in hindsight, to be a watershed moment. With the counterculture in full swing, Western society had reached what Illich thought of as an hour of kairos, a window for change. […]

One has to wonder whether his faith in each person’s capabilities was rooted in his own inordinate gifts—he spoke nine languages, was uncommonly independent of mind, and refused treatment for his own cancer out of principle.

Futures, Fictions & Fabulations

• The greatest mystery of futures studies. “A proper object, we argue, is the possibility space, a structured field that is, in essence, constitutively modal, multi-domain, anchored in contested presents, and reflexively shaped by accounts of itself. Futures studies and the special sciences are, probably, doing complementary cognitive labor rather than competing.”
• Thinking about futures, by Stuart Candy on his “the sceptical futuryst” blog. He shares his detailed answers to “post-reflection questions” following a talk he gave to “an advanced course alluringly titled ‘Future Craft: Speculative Worlds & Narrative Game Design.’”
• How science fiction solved the energy crisis. “In the 1970s, bold solutions to the crisis abounded... in the realm of fiction, at least.” Also on Reactor, Children of Men: The future is a mirror for the present.

Algorithms, Automations & Augmentations

• Remade.wiki (“everything can be remade”) and OccupySaaS are two projects to remake commercial apps or functionalities and make them available for free. “OccupySaaS is a protest movement written in code. Laid-off engineers, armed with AI and public documentation, are building clean-room, open-source, drop-in replacements for the platforms that rent our own industry back to us — one seat at a time.”
• There are now growing signs that AI is hurting employment in the job market. “‘The three heaviest AI adopters (information, professional services, finance) went from a 2.2 percent productivity trend to 5.0 percent since 2022,’ Basmajian explained, pointing to incremental gains in workplace productivity coinciding with a high AI adoption rate.”
• Human feedback guides an AI approach that automates about 84% of decisions. “Our research shows that companies should take a selective approach to automation. Rather than automating all decisions, they should deliberately retain human involvement in those decisions where people can help improve the algorithm.” … until they are, in turn, not needed either?

Built, Biosphere & Breakthroughs

• Magenta solar panels may be agrivoltaics’ cleverest trick yet. “See-through solar cells harvest some wavelengths of light to make electricity, while passing through the red and blue wavelengths that plants need. In a new agrivoltaics experiment, broccoli yields held steady.”
• 5,000 years ago, these forests drowned. Now they’re thriving reefs – home to oysters, sponges, sharks, rays and turtles. “Rising seas submerged these forests in Mozambique, now they serve as wildlife hotspots, with similar levels of biodiversity to coral reefs.”
• Fossil fuels are 40% of freight shipping tonnage, but half its fuel use. Damn! “Shipping fuel demand is driven by transport work, not just tons loaded at ports. Ton-kilometres are the better mental model. Fossil fuel cargoes travel long distances in very large flows, so their decline removes more than a proportional share of cargo mass. It removes a larger share of the ocean work and the fuel burned to do that work.”

Asides

• Three-body orbits. “Thousands of stable three-body orbits have been found — their paths repeat after a given interval. The Three Body Orbits website is an atlas of some of those orbits: you can watch each three-body system play itself out in a wide variety of different shapes and complexities.”
• Es Devlin’s infinite library rotates with more than 2,000 books in London. “The artist brings elements of her studio, decades of drawings and models, as well as a well-used collection of books into the design museum's galleries.”
• The real Kevin Bacon. Turns out the real center is Samuel L. Jackson. Going from Bacon to Jackson (their roles anyway) feels like a great metaphor for the last couple of decades of the world. “The "Six degrees of Kevin Bacon is a game where players try to connect any actor to Kevin Bacon in fewer than six steps, where each step links two actors who co-starred in a film.”