Order and chaos

There is a room in my house that I keep tidy. Everything has a place: cables coiled and labelled, books spine-out and sorted, tools in their drawers. Left alone for a few weeks, it drifts. A cable gets pulled and not returned. A book lands on the wrong shelf. A project gets started at the desk and its components spread. The room hasn’t been touched by disorder — it has simply been left alone. That is enough.

Chaos is the natural state of things. Not chaos in the dramatic sense — not catastrophe or collapse — but the quiet, persistent tendency of everything to spread, drift, and accumulate. Entropy. The desk accretes objects. The codebase grows. The inbox fills. The day, unmanaged, runs away from you.

Order is always chosen. It does not arise on its own. You have to decide that something deserves a structure, and then impose it, and then maintain it. That imposition costs effort. Which means the real question is never how do I eliminate chaos — it is which parts of chaos are worth the effort of framing.

I’ve been thinking about this for a long time without a clean answer. What follows is my working position. I’m not sure it is finished.

---

Chaos is not a feeling

When I say I feel chaos, I mean something specific. Not overwhelm or noise, though both can be symptoms. I mean the state of not being able to find the structure of something — of looking at a problem, or a room, or a cluster of thoughts, and not being able to see how the pieces relate.

A room in disarray is chaos in the external sense. The objects are there; nothing is missing. But the relationships between them — what belongs where, what is available when you need it, what can be trusted to be in its place — those have dissolved. The chaos is the absence of that architecture.

Internal chaos is the same thing, just harder to navigate because the objects aren’t physical. When my thoughts won’t settle — when I’m trying to work through something and the arguments scatter, or when I’m feeling something I can’t name — what I’m experiencing is the absence of a frame. Something is there. I can’t see how it’s organised. That absence is what I call chaos.

Both kinds are uncomfortable in the same way. It’s not the content that makes chaos unpleasant — it’s the illegibility. A difficult emotion I can understand is not chaotic. A simple problem I cannot see the structure of is. The discomfort is in the not-knowing-where-anything-is.

---

What chaos looks like from the outside

I came to Hindustani classical music in my twenties. This is late. Most students of the tradition begin as children; by the time they reach my age they have had fifteen years of daily practice. I arrived as an adult with formed habits of listening and no framework for what I was hearing.

The first recordings I sat with were dhrupad — the oldest continuous form in Hindustani music, slower and more austere than anything I had heard before. The opening section of a dhrupad performance is the alap: unmeasured, unaccompanied, without a rhythmic cycle, the musician exploring the raag’s personality alone. It begins very slowly. It can go for forty-five minutes. It goes nowhere and everywhere at the same time.

My first response, sitting with it, was that this was beautifully chaotic. Unconstrained. Organic. A musician wandering freely through a landscape. I found it moving precisely because I assumed it was improvised without rules.

It is not. The more I learned, the more structure I found. The alap has sections, each devoted to a different register of the instrument, each with a distinct purpose in how the raag’s personality is revealed. Within each section there are conventions: which notes are dwelt on and for how long, how the octave is approached, what constitutes a complete phrase versus an incomplete one. The wandering is real. The freedom is real. But the frame it moves inside is as precise as anything in any musical tradition I know.

What looked like chaos from the outside was, from the inside, one of the most structured creative practices I had encountered. The apparent freedom was the product of the structure, not the absence of it. You can only wander well when you know exactly where you are.

---

Chaos as source

None of what I’ve said is an argument for eliminating chaos. That would be the wrong conclusion.

Chaos is a source. Not a problem to be solved — a resource to be worked with.

When I encounter something I can’t structure yet, there is an energy to that. An unmapped problem has a quality that a solved one doesn’t. A system I haven’t yet built has more possibility than one that is running. A raag I haven’t yet learned has a kind of openness — it could be anything, go anywhere. The moment I understand it, the moment I can see the architecture, that openness narrows. I trade possibility for precision. That is a worthwhile trade. But the energy was in the not-knowing.

Chaos is raw material. It is high in energy and low in direction. Order is what you choose to do with some of it. You bring a frame to a piece of chaos, and inside that frame the energy becomes usable — it can do something, produce something, go somewhere specific. Outside the frame, the energy continues. It is not wasted. It is waiting.

The alap, again: the form is not a cage. It is a channel. The musician is not suppressing improvisation — they are directing it. The structure is what makes the energy of the music move somewhere instead of just dissipating. Without the raag’s constraints, you can play anything. With them, what you play means something.

When I’m playing tennis well, the experience is precisely this. All the chaos — the energy, the physical intensity, the opponent, the pressure — has been channelled into precise technique and moment-to-moment adaptation. The game is fast, adversarial, and constantly changing; there is no script. But when the shots are landing and the patterns are working, I’m not generating the feeling of order from nothing. I’m directing the chaos that is already there. The difference between a focused game and a scattered one is not the amount of energy — it’s whether the energy has somewhere to go.

---

Subtraction as method

Black and white photography is where I learned the most precise version of this principle.

Colour adds chaos. Not noise in the technical sense — chaos in the sense of too many competing layers demanding attention at once. A street scene in colour carries the warm orange of the chai stall, the blue of a shirt, the grey of the concrete, the green of something further back. Each colour carries an association, a temperature, a mood. Managing all of that in a frame is a complex act of ordering. Removing colour eliminates one layer of it entirely — and what remains is geometry, weight, light and shadow. A simpler problem. A more honest one, I think.

But removing colour is only the first step. The frame itself is still chaotic. A city at any given moment is an accumulation of hundreds of competing things: a person passing who catches the light, a composition that almost holds, a gesture that is almost right. Almost is not enough. So I wait.

Most of the photograph happens before the shutter opens. The hour spent walking the same street a second time. The ten minutes standing in one position, watching how the light moves. The conversation that takes the tension out of someone’s face. Waiting is not passive — it is the work of recognising when the chaos of the scene resolves into something legible. The shutter is not the act of creating order. It is the act of recording order at the moment it arrives.

This is different from the music, where I build the structure myself. In photography, the structure is out there, and my job is to be present when it briefly appears. The patience the form asks for is the patience of not forcing it — of trusting that the scene will organise itself if you wait long enough, and being ready when it does.

---

The adversarial form

Tennis brings something none of the other disciplines do: an opponent.

In music, in photography, in engineering, the chaos comes from the problem or the material. In tennis it comes from another person who has their own structure, their own strategy, and who is actively trying to find the gaps in yours. The chaos is adversarial. It arrives faster than you can plan for it, and it changes every time you think you have understood it.

When a match slips — when a set I was winning starts coming apart — what I’ve learned to do first is distinguish between the two possible causes. Either I didn’t perform well: my shots tightened, my footwork got sloppy, the structure I rely on began to fail. Or my opponent was genuinely better: they found something in my game and exploited it with greater skill than I could answer. These two things require different responses, and confusing them is one of the more expensive mistakes you can make in a competitive context.

When I’ve done everything within my control and it still slips, I accept it. Acceptance is not defeat — it is accuracy. It is knowing where your control ends and the world begins, and not wasting effort on what is past. The match is over; the result is fixed; what remains is understanding.

After the match I look for patterns. Where did my game hold? Where did it break? Where was my opponent genuinely strong, and where were they simply finding the places I hadn’t yet shored up? That information goes into practice. The loss becomes data, and the data becomes structure, and the structure makes the next game more ordered. Chaos, again, as source — not something to be avoided, but something to be learned from.

What I find interesting about tennis is that the best state of play — when a match is going well, when everything is working — doesn’t feel like I’ve eliminated chaos. It feels like I’ve channelled it. The opponent’s pressure, the physical intensity, the need to make decisions in fractions of a second: all of that is still there. But it is directed. It is going somewhere. The order is not the absence of chaos. It is chaos pointed.

---

Complexity and simplicity, at the same time

I build things that are more complex than they need to be, if need were the only criterion.

I run my own servers instead of using hosted services. I write retrieval systems when I could use better models. I practise a musical tradition that would be several degrees simpler if I chose a different one. In every case, the path I chose involved building more structure, more complexity, more of the underlying machinery visible to me and managed by me.

This isn’t an aesthetic preference for difficulty. It is something closer to a conviction: that the complexity you build yourself is navigable in a way that complexity you inherit is not. When my self-hosted infrastructure breaks — and it does break, regularly — I know where to look. The chaos of the failure is legible to me because I built the order around it. When a cloud service fails, the chaos is illegible because the structure is hidden. I can restart a container. I cannot restart a data centre.

This is what high-performance computing is, at its largest scale. Thousands of threads running in parallel, memory hierarchies that behave differently depending on access patterns you cannot fully predict, emergent behaviour at a scale where no single person can hold the whole system in their head. The job is to impose enough structure that this enormous chaos can do something specific — something useful, something that would not be possible without it. What excites me about this work is exactly that: complex systems of chaos coming together through order to solve something grand. A well-structured piece of code, a module clean enough to be reused, a parallelism problem that finally yields — these carry the same satisfaction as a well-placed note in an alap, or the moment a photograph resolves. The form is different. The quality of the feeling is not.

But the other half of it: I want the thing that results to be simple. The user of a system I build — which is often myself — should not feel the complexity. The complexity is supposed to disappear behind an interface that is as clean as possible. I live between these two requirements: the system must be as complex as it needs to be to do the thing well, and it must be as simple as possible to use. I find both interesting. I’m not sure you can have the simplicity without doing the complexity work first.

---

Order is not neutral

The cost of order is underestimated.

Order is not just structure — it is maintained structure. A room is not tidy because you once tidied it. It is tidy because you keep tidying it. A practice is not consistent because you once showed up — it is consistent because you keep showing up. The cost is not one-time; it is ongoing. Every system you decide to maintain adds to the weight you are carrying.

The cost of order and discipline is very high. Not as a complaint. As fact. The 5am alarm. The long stretches of a raag that go nowhere, or go somewhere you did not want to go, and must be unpacked and tried again. The match you lose badly and have to sit with until you understand it. The system that breaks on a Saturday and takes three hours to fix before you can get back to the thing you were actually trying to do. None of these are romantic. They are the actual texture of maintaining order in a life.

I don’t try to impose order on everything. Where I lack the conviction that the structure will produce something meaningful — where I can’t see a clear purpose for the frame — I leave the chaos alone. Not because I am comfortable with it, but because building structure for its own sake is one of the more reliable ways to exhaust yourself. The structure has to earn its existence.

---

Joy and excellence

Why do any of it?

The cost is high, the maintenance is ongoing, and the chaos returns the moment you stop. So what is the reason to keep choosing order in these particular places?

Two things. Joy and excellence. Things that make me happy, and things that make me the best version of myself I can be.

The two aren’t separate for me. The things that produce genuine joy are the same things that demand growth. Music doesn’t let you stay comfortable — the raag you have played a hundred times still finds the gaps in your understanding. Tennis doesn’t let you stay comfortable — the opponent you cannot yet beat is the one who is teaching you the most. The photograph that doesn’t work is the one that shows you what you haven’t yet learned to see. The engineering problem that resists you is the one that, when it finally yields, you will understand completely.

I play music at five in the morning because I go to bed wanting to play again the next morning. I wait for the right light on a street for an hour because the moment the scene resolves is one of the more satisfying experiences I know how to reliably produce. I study the pattern in a lost set because the next time I meet a similar situation and handle it better, I feel genuinely glad. I build complex systems because the moment a parallelism problem finally yields — when the structure holds and the chaos is doing exactly what you needed it to do — it is a specific and irreplaceable pleasure.

Over time, I’ve discovered which things give me that dual feeling and which don’t. The ones that don’t have both — that give pleasure without growth, or demand growth without pleasure — I’ve stopped doing. Not every pursuit has survived this test. But music, tennis, photography, and engineering have stayed, and deepened, across years. They are the forms I return to even when they are hard, which is most of the time. The difficulty is part of why they stay. A form that cannot get harder cannot make you better, and a form that cannot make you better eventually stops bringing joy.

The joy is not in the output. It is in the act of channelling. The satisfaction is not in the finished photograph or the won set or the clean module — it is in the moment of directing chaos into something purposeful, and finding yourself slightly more capable than you were before. That experience is available every time you sit down, regardless of how the output turns out. It is, in the end, what the whole practice is for.

---

At five in the morning

There is a reason I practise at five in the morning, before the day begins.

It is partly about quiet. A surbahar resonates at a depth that carries through walls; the building is still at that hour. But it is also about the quality of what I arrive with. At five in the morning, before the day has accumulated anything, my mind has not yet been shaped by what has happened. Whatever internal structure I have is whatever I woke up with — before the news, before the messages, before the particular pressures of the particular day.

The raag takes that material and gives it somewhere to go. The form holds the morning’s unsettled quality and makes it productive. I am not trying to calm anything. I am trying to direct it.

This is what all of it is, finally: forms that take the chaos I arrive with — whatever state I am in, whatever the day carries — and give it direction. Some of those forms are ancient and transmitted across generations. Some are physical and adversarial. Some ask for subtraction. Some ask for an enormous amount of technical knowledge to impose at scale. But they are all, at root, the same thing: a chosen structure that makes the chaos useful.

I keep choosing them because of what happens when the channelling works. That feeling — of chaos finding its direction, of energy becoming precise, of yourself becoming slightly more than you were — is worth a great deal. Including the alarm at five in the morning.

Joy and excellence. That is the working position. I reserve the right to revise it.