I read Joseph Earp’s essay on optimisation and classical music this morning and then went looking for the swans.
The swans belonged to Jean Sibelius. On 21 April 1915, he watched sixteen of them circle above him before disappearing into the haze. He called it one of the great experiences of his life. The sight, the sound, the lift of those bodies into the sky found their way into the final movement of his Fifth Symphony, where the horns rise above the rushing strings in what has become known as the swan theme.
I listened to it again after reading Earp. There’s no shortcut into that passage. You have to let the music carry you there. The earlier movements have done something to your sense of time. The repetitions have prepared the ground. Tension has gathered without explaining itself. Then the horns arrive, and for a moment you’re standing with Sibelius beneath the birds.
Nothing has been improved. No problem has been solved. The listener hasn’t become 12 per cent more effective.
Something has happened all the same.
This may be one reason art has become such a useful place to hide. So much of contemporary life asks to be assessed by what it produces, how quickly it produces it, and whether the process can be refined so that tomorrow it produces more. Art can still waste our time properly. A symphony can take forty minutes to make a movement inside you that no dashboard would know how to record. A novel can keep you wandering for weeks. A weblog can follow a thought down three side streets, gathering memories, links and unexpected encounters, and return with something lived and particular, even if it never reaches a tidy conclusion.
Earp writes about retreating into Mahler, Morton Feldman, Alfred Schnittke and Sibelius as the language of efficiency spreads from the workplace into sleep, exercise, entertainment and rest. The absurdity is now familiar. A poor sleep score can spoil the morning before the body has had a chance to say how it feels. The walk becomes step production. The books beside the bed begin to look less like invitations and more like a backlog. Even leisure is expected to return us to work sharpened, restored and ready to go.
We don’t simply rest. We optimise recovery.
And because the whole arrangement presents itself as self-improvement, it can be difficult to name the violence in it. Nobody appears to be forcing us. The foreman has disappeared. The factory whistle has gone quiet. The pressure comes from the watch on the wrist, the app on the phone and the small manager we have installed inside our own heads.
Byung-Chul Han calls this figure the achievement-subject: the person who experiences the demand to perform as freedom and therefore becomes both worker and overseer. Jenny Odell makes a similar connection in Saving Time. After studying Taylorism and the old industrial effort to make workers assemble things faster, she noticed that productivity culture asks us to apply the same logic to ourselves. The factory manager has been internalised. We reward and punish ourselves according to the numbers. We carry the clipboard home.
This is a particularly effective form of control because it recruits desire. I don’t need a boss to tell me to become better. I already want to become better. I want to write more clearly, move with more energy, use my days well, and make the most of whatever time remains. There’s nothing false about those desires. The trap opens when every part of life is fed into the same machinery.
The tool that improved one task becomes, almost without announcement, a philosophy of existence.
When the measure becomes the meaning
Optimisation works best inside a bounded problem. If I’m planning a delivery route, compressing an image, or trying to reduce the time spent copying information between systems, then faster and more efficient may genuinely mean better. Some friction is merely dead weight. I don’t become wiser by forgetting a password, filling the same form twice, or waiting for a badly designed website to load.
The trouble begins when the measure quietly replaces the meaning of the activity.
A walk can be made more efficient if the purpose is to reach a destination. But sometimes walking slowly is how I notice what I didn’t know I was thinking, and the detour is the method. A conversation can be shortened when its purpose is to exchange information. Its value may also lie in the pause before someone says what they really mean. A draft can be produced in seconds while the writer still needs three hours of not knowing to discover what the piece is about.
The same friction can be waste in one context and formation in another.
This distinction matters more as AI becomes woven into ordinary creative and intellectual work. The promise is irresistible: fewer blank pages, quicker summaries, cleaner prose, instant competence. I use these tools. I’m not writing from a cabin beyond the reach of electricity, warning the villagers about the machine. AI can remove drudgery, open doors, retrieve forgotten material, test an idea and help a solitary writer work with a range that once required a room full of specialists.
But frictionless competence has its own danger. If the tool always finds the first frame, writes the first sentence and resolves the uncertainty, then the finished work may improve while the person making it gets less practice at meeting the blank page alone. I can’t prove that the capacity underneath is thinning. I can feel the question becoming more urgent each time help arrives before I have formed a view of my own.
There’s already a hint of a related problem in the research, though the evidence is narrower than the argument I’m making here. In a 2024 experiment, Anil Doshi and Oliver Hauser studied 293 people writing eight-sentence stories. Some participants could request one AI-generated story idea and others could request up to five. Access to those ideas produced stories judged more creative, better written and more enjoyable, with the largest gains among people who scored lower on a divergent-association test. That is the good news, and it is genuinely good. The floor rises.
The same study found that the AI-assisted stories became more similar to one another. Individual performance improved while the collective field narrowed. Each person did better according to the measure, yet the room as a whole produced less variation. The study did not examine professional writers, long-form work, sustained collaboration or the gradual development of craft, and its authors warn against carrying the results too far. Still, the tension it reveals is worth holding.
This is the convergence trap in its most seductive form. The work doesn’t become bad. It arrives polished, grammatical and ready for publication. The minimum standard rises while everyone is quietly tempted towards the same centre. The danger isn’t a flood of obvious rubbish. We already know how to ignore rubbish. The danger is an endless supply of work that is good enough to pass, smooth enough not to offend, and familiar enough to require nothing from us.
The machine is particularly good at giving us the shape of an answer before we have fully inhabited the question.
That may save time. It may also remove the time in which a mind becomes capable of an answer worth giving.
Music that refuses to hurry
Morton Feldman’s later compositions can last for hours. His Second String Quartet is famously long enough to make an evening disappear. The duration isn’t an obstacle placed between the listener and the work. Duration is part of the work. The ear tires of waiting for development in the usual sense. Attention loosens, wanders, returns. Tiny changes begin to feel enormous because the music has altered the scale by which change is perceived.
You can listen impatiently, of course. You can skim Proust, play Feldman while answering email, or put Mahler on at double speed. Technology will permit almost any insult. But the work will withhold itself.
Some things are not slow because nobody has optimised them yet. They are slow because slowness is how they become available.
That thought opens a door onto an older argument. In his 1932 essay In Praise of Idleness, Bertrand Russell complained that modern people think everything must be done for the sake of something else. Leisure had become morally suspicious unless it could justify itself through future work. He was writing in the age of industrial machinery, but the sentence has aged well. We now ask not only what an activity is for, but how it can be stacked with another activity, tracked, improved, monetised, shared or turned into evidence of progress.
Read while walking. Learn while commuting. Network while exercising. Build a personal brand while having a life.
The hour is not allowed to remain an hour. It must report for duty.
Josef Pieper, writing sixteen years after Russell, made an important distinction between leisure and mere inactivity. Russell’s argument is political: if machinery reduces necessary labour, the reward should be time shared more justly rather than unemployment for some and longer hours for others. Pieper looks inward. Leisure, for him, wasn’t collapse after work or entertainment consumed to anaesthetise ourselves before Monday. It was a receptive condition, a willingness to let the world be more than raw material for our plans. It made contemplation, wonder and culture possible because it suspended the demand that everything prove its usefulness.
That sort of leisure is difficult now because usefulness has escaped the office. It follows us into the garden. We take a photograph before we have properly seen the thing. We turn a hobby into a side project, a side project into a revenue stream, a private pleasure into a public identity. Soon the activity has an audience, a schedule and a content strategy. The little patch of ground where nothing had to perform has acquired quarterly targets.
I understand the impulse. Making something shareable can deepen attention. A photograph can become a way of looking. Writing about a walk can reveal what the walk contained. The blog itself is a practice of bringing experience into language. The problem isn’t that we make forms from life. The problem is forgetting that life doesn’t owe us a form.
A three-hour walk may produce an essay. It may also produce sore feet, a photograph that stays on the phone and one half-remembered sentence. The walk hasn’t failed.
A rambling weblog may slowly become an archive of a mind in motion. It may also contain abandoned threads, odd enthusiasms, posts read by twelve people and questions that never arrive at an answer. The waste is part of its aliveness. Remove every tangent and the blog becomes a brochure. Remove every silence and the music becomes information.
The experiences that answer back
The sociologist Hartmut Rosa offers the word resonance for a relationship with the world in which something reaches us and we answer, with both sides changed by the encounter. Music is an obvious example, though resonance can happen with a person, a place, a piece of work, a religious practice, an animal, a book or a landscape.
The crucial thing is that resonance can’t be guaranteed. We can create conditions in which it becomes more likely, but we can’t order it to appear. The concert may leave us cold. The walk may be dull. The book everyone loves may refuse to open. Another day, a line of brass beneath imagined wings can undo us.
Optimisation dislikes this because optimisation depends upon control. It wants a reliable input, a repeatable process and a measurable result. Resonance keeps a wild element. The world gets a vote.
This is why attempts to optimise happiness, creativity or love often become strangely lifeless. The very effort to secure the experience changes the relationship. Attention turns towards self-monitoring. Am I relaxed yet? Is this meaningful? Have I reached flow? The hand keeps digging up the seed to check whether it is growing.
We are tired from the work, certainly, and from insecurity, noise, screens, bills, care, bad sleep and the ordinary weight of being a person among other people. Some portion of the exhaustion may also come from being unable to leave ourselves alone. Every mood is treated as a problem to regulate. Every interest must be developed. Every hour must choose a side between investment and waste.
Optimisation promises control over life. What it often produces is a second life running alongside the first: the life that measures the life, judges the life, records the life and keeps asking whether the life could have been better used.
We have the experience while another part of us audits it.
The counter-move isn’t to romanticise inefficiency. Pointless bureaucracy doesn’t become sacred because it takes all afternoon. Poverty, overwork and lack of time are not spiritual exercises. The right to waste time has always been unevenly distributed, often purchased by somebody else’s labour. A person working two jobs doesn’t need a lecture about putting down the productivity planner. A parent or carer may know exactly what unmeasured time is worth while possessing almost none of it. The clean, frictionless service enjoyed by one person may be held together by warehouse workers, delivery drivers, moderators and poorly paid people training the machine in the background.
Russell understood leisure as a political question because time can’t be separated from power. Odell extends that argument by asking us to see the self as relational rather than sealed off: my freedom rests inside arrangements of work, care, land, history and other people’s time. Pieper’s contemplative leisure matters, but it can’t float above those arrangements as a private spiritual achievement. If I turn leisure into one more possession I have earned through superior life design, optimisation has simply changed clothes.
A serious defence of wasted time must therefore move in two directions at once. It protects the private hour from the inner foreman, and it asks what kind of shared world would allow more people to have such hours. Otherwise one person’s freedom to listen to Mahler remains another person’s night shift.
Nor do I want to pretend I can simply step outside the technological world. I like useful tools. I like the strange intelligence available through a screen. I want systems that handle what needn’t consume a human life. The question isn’t whether we should optimise. It is where optimisation belongs, and whether we are still capable of recognising a value it can’t measure.
Ivan Illich called tools convivial when they enlarged people’s capacity to act, create and live together without turning them into servants of the tool. That feels like the live question for AI, productivity systems and all the devices gathering around the self. Do they return us to the world with greater freedom and attention? Or do they make the world feel like one more badly managed project?
The best tools, at least some of the time, return us to the world and disappear from view.
A small protected territory
I keep returning to the phrase wasted time because it still carries a little shame. It names the hours that can’t defend themselves in the language of production. The walk without a destination. The novel too large to summarise. The piece of music that refuses to become background. The afternoon spent following links through somebody’s old weblog. The conversation that wanders beyond its stated purpose and becomes the reason the meeting mattered.
None of this is automatically profound. Wasted time can be boring, lonely, anxious or genuinely wasted. That is part of the bargain. If we demand that it always deliver insight, restoration or transcendence, we have simply given it another performance target.
The point is to preserve some territory in which the result is not known in advance.
That territory may be one of the places where a self can still surprise itself. It is where taste develops without an analytics panel. Where thought takes the long route. Where an obscure enthusiasm gets to live before it is asked to become expertise. Where we encounter something that does not resemble us and stay long enough for it to alter the rhythm of our attention.
It is also where we practise being more than a bundle of capacities waiting for an upgrade.
There is an older meaning of cultivation that I prefer to optimisation. It allows for discipline, technology and ambition, but also for limits, seasons and response. You can prepare soil and remove what chokes the roots, but you can’t bully a garden into ripening on your schedule. Cultivation asks for participation rather than mastery. Odell’s attention to trees, flowering and seasonal time belongs here: the world is active, but it doesn’t run on interchangeable minutes.
Sibelius wasn’t looking at the sky to gather material for his productivity system. Sixteen swans passed overhead. He noticed them. The experience entered him and, after years of revision, entered the music. More than a century later, a listener hears the horns rise and feels something lift inside his own chest.
The swans crossed the sky. Someone was available to be interrupted.
Perhaps wasted time is one of the last territories optimisation hasn’t completely colonised. If so, I don’t want a five-step plan for defending it. I want to leave a gate open, walk through it occasionally, and resist the urge to count how long I was gone.
Notes and further reading
- Joseph Earp, “Exhausted by optimisation, I’m seeking solace in the past”, The Guardian, 21 September 2026.
- Andrew Mellor, “Sibelius’ 5th Symphony: A Song for Swans”, NPR, 2011.
- Bertrand Russell, “In Praise of Idleness”, 1932.
- Byung-Chul Han, The Burnout Society, Stanford University Press.
- Josef Pieper, Leisure: The Basis of Culture, St. Augustine’s Press.
- Jenny Odell, interview on Saving Time, Esquire, 2023.
- Hartmut Rosa, “Beyond the Echo-chamber: An Interview on Resonance and Alienation”, Krisis, 2019.
- Anil R. Doshi and Oliver P. Hauser, “Generative AI enhances individual creativity but reduces the collective diversity of novel content”, Science Advances, 2024.
- Ivan Illich, Tools for Conviviality, 1973.
- Morton Feldman, Speaking of Music interview, 1986.
