The Score

A musical score is the strangest artefact in the cultural canon. A single document, composed once, in solitude, that becomes a living instruction set for hundreds of performers across centuries — each of whom must open the page and immediately understand not only what notes to play but what the composer meant by writing them in that order. The score is not the music. The music happens when human beings interpret the marks. The score is the medium through which intent survives translation, the document whose entire job is to be read by someone other than the person who wrote it, the artefact whose value is measured almost entirely by how cleanly its meaning travels from one mind into another.

Software has spent forty years pretending it is something else. We built our codebases as factories — measured in throughput, output, lines shipped, features delivered — and we rewarded the engineers who produced the most of it, fastest, regardless of whether anyone downstream could read what they made. The codebase as factory floor produced a generation of systems whose internal logic was knowable only to the original authors, only briefly, and only on a good day. The industry then built an elaborate apparatus of tribal knowledge, internal docs, onboarding programmes, and Slack archaeology to compensate for the fact that the source itself had stopped being legible somewhere around the second refactor.

That apparatus was always wasteful. But it was survivable, because the bottleneck was always the writing. A team of eight could produce only so much code in a quarter, and the limits of human typing speed acted as an unintentional governor on complexity. The codebase grew at roughly the rate the humans inside it could still hold in their heads.

That governor has now been removed.

The New Scarcity

Competent code generation has dissolved, in under two years, the constraint that shaped every assumption about how software gets built. A pair of engineers with the right tooling can now produce in an afternoon what a six-person team used to produce in a sprint — not because the engineers got faster, but because typing is no longer the bottleneck. Anything is no longer the bottleneck, except the thing that always quietly was: the human capacity to understand what has been written. Production cost has collapsed. Comprehension cost has not.

The first-order consequence is obvious. The codebases of the next decade will be larger, denser, more abundant than anything before them, and most will collapse under their own weight inside eighteen months, because no human will be able to walk in and know what is going on. The second-order consequence is the interesting one. The codebases that survive — that compound rather than collapse — will not be the ones with the best AI assistance or the cleverest abstractions or the most aggressive automation. They will be the ones whose source is scored. Composed for readers. Written so that a product manager, a developer hired this morning, a cinema operator with a quarter of an engineering background, and an autonomous agent acting for any of them can open the same file and read the intent before they read the implementation.

The industry keeps mistaking this for a quality-of-life concern. Readable code used to be what senior engineers asked for when they wanted juniors to use better variable names. That framing is now dangerous. When a single engineer can summon ten thousand lines of plausible code before lunch, the cost of writing has dropped below the cost of reading, and the gap widens monthly. Writing code that explains itself to other humans — across roles, across seniority, across the line between people who write software and people who only operate it — is no longer a courtesy to colleagues. It is the substrate any codebase that intends to stay economically viable in the agentic era has to be built on. The scarce resource is not lines. It is legibility.

What the Score Encodes

A score does not encode the sound. Anyone who has watched a violinist read Bach for the first time knows the marks are an instruction for how to think about a passage, not a recording of it. Dynamics, tempo, phrasing — interpretive cues, signals from the composer about what matters, what carries weight, what to listen for in your own hands. The score tells the reader what the writer was paying attention to.

Code can do this. Most of it doesn't.

A typed client method that reads client.sessions.list({ siteId, date }) is not merely clean the way an aesthetic preference is clean. It is scored. The shape of the method is a sentence — a verb acting on a domain object, parameterised by exactly the two things that matter, returning a value whose type is the same shape across every other endpoint in the system. A developer reading that line for the first time does not consult documentation, ask a colleague, run the code, or open a debugger. They have read the intent in the signature. The cost of comprehension is zero, because the writer paid it for them, in advance, by composing for a reader.

An error hierarchy that produces a RateLimitError with a typed retryAfter field is scored the same way. The error is not a string parsed from a response body, not a status code interpreted in three places by three engineers each making slightly different assumptions. It is an object whose presence in a catch block tells the next reader, in one line, both what went wrong and what to do about it. The recovery logic writes itself — not because the code is clever, but because the writer named the situation precisely enough that the name carries the meaning. Someone walking in six months later does not need to discover that the API uses HTTP 429 with a Retry-After header in seconds-or-HTTP-date format. They need to know rate limits exist and there is an idiom for handling them. The score already told them.

This is the discipline. Not commenting more. Not writing prose explanations beside the implementation. Composing the implementation itself so the names, the types, the structure, the order of operations all collaborate to make the next reader's interpretation effortless. Every method signature is a line of the score. Every type definition is a marking. Every directory structure is the architecture of the page, telling the reader where to look first and what carries the melody.

The Cross-Role Test

Cinema technology is the proving ground for this discipline, because cinema technology has no choice. The domain knowledge required to build a cinema platform is distributed across roles in a way almost no other industry replicates. The person who programmes films into a session holds a different mental model from the person who configures loyalty-tier rules, who holds a different model from the engineer building the seat-map component, who holds a different model from the accountant reconciling tax across territories, who holds a different model from the operator who opens the building in the morning and runs it until midnight. A booking flow is simultaneously a UI problem, an inventory problem, a tax problem, a payments problem, a concessions problem, a loyalty problem, and a real-estate problem — and the codebase that supports it has to be readable by anyone holding any one of those models, and increasingly by agents acting for any of them.

The horizontal-SaaS playbook handles this by hiding the domain behind a generic interface and asking every role to bend toward the abstraction. Cinema technology cannot afford that. The domain is too specific, the operations too varied, the cost of misreading too immediate — a misunderstood inventory state means an empty seat in a sold-out screening, or a double-sold seat in front of two paying customers, and there is no graceful failure mode for either. The codebase has to encode the domain precisely and legibly, in language the operator would recognise as their own. Sessions, not events. Held, not pending. Companion seat, not adjacent reserved. The names matter because the names are what let the engineer and the operator point at the same line of code and agree on what is happening.

A codebase composed this way reads, across roles, as a single document with a single voice. The product manager opens a file and recognises the vocabulary from the last operator interview. The engineer opens the same file and recognises the type contracts from the API docs. The operator, walked through it by either of them, recognises the model of their own building. The score is the same. The performers differ. The music — the operating cinema — happens when each of them plays their part against a document they can all read.

This is the territory the work I am building under the name Theatrical is being shaped to demonstrate: that the cinema-technology codebase composed for legibility outperforms the one composed for cleverness. The thesis is not that cinema needs better tools — that argument lives elsewhere. The thesis is that the tools cinema deploys over the next decade will be judged by a property the industry has not yet learned to measure: how cleanly their source reads to the humans, and now the agents, who have to live inside it.

The Industrial Argument

None of this is an aesthetic preference. The score is not a stylistic choice on the composer's part; it is the mechanism by which an orchestra is possible at all. Strip it away and you do not get a freer ensemble — you get silence, or noise, depending on the temperament of the players. The same is true of a codebase. The legible one is not the prettier or the better-documented or the more pleasant codebase to work in. It is the only kind that survives an environment where production has become free and comprehension is the only thing standing between a company and irrelevance.

The codebases that compound over the next decade — that absorb new engineers, new product managers, new operators, new autonomous agents, without each onboarding extracting a tax on velocity — will be the ones whose authors understood early that writing software had quietly stopped being a writing problem and become a reading problem. The differential will not be subtle. From the outside it will look like one company shipping confidently into adjacent territory every quarter while another spends every quarter rediscovering what its own systems do. The first company will look lucky. It will not be lucky. It will have composed for readers from the beginning, paid the small upfront cost of scoring every line, and collected the compounding return on every change after.

The companies that get this right will dominate the industries where domain knowledge is distributed across roles, because those are the industries where the cost of illegibility is highest and a shared score is worth most. Cinema is one of them. Cinema is, in fact, the cleanest test case any of us are likely to meet — an operation where a programmer, an operator, an accountant, and a moviegoer all touch the same system inside a single transaction, and the codebase either holds together as a coherent document they can all read or it doesn't.

Coda

A score is not finished when the composer puts down the pen. It is finished, and re-finished, every time a musician opens it and finds the meaning inside the marks intact. The composer's craft is measured, over the long arc, by how many performers across how many decades can pick up the same page and play. The engineering craft of the next era will be measured the same way. The codebases worth inheriting are the ones a stranger can open, on a Tuesday morning, and read like a sentence written for them — and only then will the music start.

Read this essay on brunohart.dev →

Something quiet has happened to the economy of making things, and it has not yet been priced in.

For most of the software era, the difficulty of production was the moat. Building was hard, shipping was hard, distribution was hard, and companies organised themselves around the friction of getting anything into the world — engineering teams, design systems, content pipelines, release trains, the whole apparatus of throughput. Advantage accrued to whoever could push more, faster: more features, more posts, more landing pages, more decks, more ads, more SKUs. The assumption beneath thirty years of software strategy was that production itself was scarce, and therefore valuable, and therefore the axis on which companies would compete.

That assumption has collapsed in about thirty months.

The cost of producing software is approaching zero. The cost of producing an image, a video, a piece of marketing copy, a logo, a song, a wireframe, a working prototype — approaching zero. The cost of producing an opinion, formatted for distribution, with a header image and three bullet points and a confident closing line — approaching zero. The machines have eaten production. They are still chewing, but the meal is essentially over. What used to take a team a month now takes an agent an afternoon, and the afternoon keeps getting shorter.

Companies reading this moment as a chance to produce more are about to learn what abundance does to economies of attention. The same thing it has always done to economies of anything: it destroys the value of the abundant good and relocates value to whatever just became scarce.

The newly scarce thing is craft.

The Death of Throughput as an Axis of Competition

Throughput won the last era because production was the bottleneck. The companies that learned to ship faster — Amazon's two-pizza teams, Facebook's move-fast-and-break-things, the whole DevOps movement — built durable advantages out of velocity, and the customer felt the difference. A product that shipped weekly beat one that shipped quarterly, because the weekly product compounded learning faster, fixed friction faster, evolved faster. Speed of production was a legitimate moat because speed of production was hard.

Speed of production is no longer hard.

A founder with three weeks of patience and an AI subscription can ship a working product. A marketing team of one can produce the output of a team of twenty. A designer with the right prompt stack generates variants faster than a stakeholder can review them. The bottleneck has moved. It is not in the making. It is in deciding what is worth making. In the taste applied to the output. In the judgment that separates the thing that should exist from the thing that merely can.

This is what craft is. Not handwork. Not artisanal indulgence. Not nostalgia. Craft is the disciplined application of taste, intent, and attention to the question of what gets made and how it is shaped — the layer of human judgment that survives when production itself has been automated away. In the old era, craft was a luxury good: nice to have, but margins flowed to whoever shipped fastest. In the new era, craft is the whole game, because everything else has been commoditised.

The companies that understand this will define the next generation of cultural and economic value. The ones that do not are about to spend a decade producing infinite mediocrity at increasing scale, wondering with genuine bewilderment why none of it lands.

The Failure Mode Is Already Visible

You can watch the failure mode forming in real time. LinkedIn feeds full of AI-generated thought leadership written in the same three rhythms by the same three prompt templates. Thumbnails that all look like the same thumbnail, because the same model trained on the same data made them. Marketing copy that reads like marketing copy because no human ever decided what it was for. Content factories — already at industrial scale — pumping out articles and summaries and listicles and SEO pages that exist for no reader, satisfy no curiosity, serve no purpose beyond the fiction of activity. The infinite scroll of generic. Taste flattened to a single beige average.

This is what abundance without craft looks like. Not a future risk — the present condition. And it will get worse before the market corrects, because the cost curve still favours producing more of it. For another year, maybe two, companies will pour resources into volume — generate more, post more, ship more, occupy more channels — and be rewarded with declining engagement, declining conversion, declining brand equity, and a slowly creeping suspicion that something is wrong with a strategy they cannot yet name.

What is wrong is that they are competing on the axis the machines now own. They are showing up to a race where the machines run for free, and bringing more machines. They are mistaking abundance for advantage. The advantage has moved.

The companies that will define the next decade treat production as solved and aim the entirety of their human capacity at the questions production cannot answer. What is worth making? For whom? Why now? What does this need to feel like? What is the texture, the weight, the integrity of the thing? What would make someone, somewhere, feel that this was made for them, by someone who cared?

These are craft questions. They have always been craft questions. For thirty years software was allowed to ignore them because production was hard enough to be the whole problem. Production is no longer the whole problem. Production is no longer any of the problem. The problem is everything else.

Cinema Is Positioned for This Moment

!A plate from Banister Fletcher's A History of Architecture on the Comparative Method — a building rendered in elevation and section side by side, the facade revealing the public face and the cut-through exposing the load-bearing structure that holds it up, columns, vaults, and arches drawn at a scale that lets you read both the proportion and the construction logic.%20344.jpg?width=800)

A plate from Fletcher's comparative history — facade and section at the same scale, because the discipline insists that what shows and what bears the load are the same question answered twice. Architecture is the longest-running ledger we have of which craft choices compound. The buildings still standing are not the ones built fastest; they are the ones whose makers refused to confuse production with worth, and whose proportions and materials and structural logic survived because they were chosen with the centuries in mind. Cinema sits inside the same lineage — film as the artefact made to outlast the moment of its making, the theatre as the room built for the purpose. The technology serving both has no business operating at a lower standard. — Source: Banister Fletcher, A History of Architecture on the Comparative Method, on Wikimedia Commons_344.jpg)

Of all the industries facing this transition, cinema is positioned with a precision almost no other can match.

Film is, by definition, a cultural artefact built on human craft and taste. Every frame is a series of choices — what to point the camera at, how to light the face, where to cut, what to score, which line reading to keep, what colour to grade the sky. Hundreds of people making thousands of choices that the audience experiences as a single coherent intent. The product is the craft. The craft is the product. There is no other layer to the thing.

And cinema is consumed in a shared, intentional space — strangers arriving at a building made for the purpose, agreeing to sit in the dark and attend to something specific. Not scrolling. Not skimming. Not multitasking. Nothing left running in the background. Theatrical exhibition is the only major content channel left that requires intentional human attention as the price of admission. Everywhere else, attention is fractional, ambient, recoverable. In a cinema it is whole.

Both halves of the equation privilege craft. The artefact is craft-dense. The mode of consumption is craft-rewarding. The industry is, structurally, the most craft-aligned content business at scale — not by accident of culture but as a property of the medium, and that property is about to become extraordinarily valuable as every other channel collapses into algorithmic sludge.

Cinema's product has always been craft and service — film as a tool for social good, a tool for communication, a way of serving your fellow human. The technology serving cinema has to inherit the standard. There is no other coherent move. Software supporting an industry whose entire business is craft cannot itself be careless. Tooling underwriting a medium of taste cannot itself be tasteless. The operational systems moving millions of seats and tens of millions of dollars through theatrical exhibition every week are not infrastructure in the engineering sense. They are infrastructure in the cultural sense — the underpinning plumbing of a craft economy — and they have to be built with the seriousness that implies.

What Craft Looks Like in Software

It would be possible to read all of the above as a sentimental argument about humans and machines, and it would be a misread. The argument is operational. Craft in software is a set of specific, observable disciplines that separate work that compounds value from work that does not.

In code, craft is the type system treated as a design surface rather than a constraint — schemas precise enough that the wrong call simply cannot compile, error messages that read like a colleague explaining the problem, signatures that document intent before mechanism. In design, it is component systems that hold under stress, where the third-level state of a fifth-tier element still feels like it belongs to the same family as the front door. In documentation, it is the manifesto disguised as a reference — the kind that leaves a reader oriented, respected, invited to do good work. In operations, it is dashboards that surface signal rather than telemetry-for-its-own-sake, alerts that fire only when something is genuinely wrong, runbooks written by someone who remembered the future person who would need them. And in the way a team is run, it is the discipline to say no to the seventh feature so the first six can be excellent, the willingness to delay a launch a month because the thing is not yet itself, the refusal to ship the version that almost-works.

None of this is slow. That is the misunderstanding worth pre-empting. Craft does not mean measured pace. It means the rate of production is no longer the variable being optimised. A team that has internalised craft as its operating system can move astonishingly fast, because every decision is made on the right axis. Time gets spent on the questions that matter. Time gets reclaimed from the questions the machines now answer.

The work I have been doing on a developer toolkit for theatrical exhibition has been an extended exercise in exactly this — a deliberate refusal to treat the technical surface as merely functional. Type-system rigour as a courtesy to the next developer. Documentation written as if it were the first impression of an entire industry. Architectural choices made with the attention a cinematographer brings to a lens. Not because the engineering benefits from being romantic about itself, but because the engineering is in service of an industry whose product is craft, and the technology has no business being sloppy when the medium it serves is not.

This is what the post-abundance economy looks like from the inside. The work is more careful, not less. The decisions take longer, the output is smaller, and the things that get made are worth more. The economics flip.

Coda

For a very long time, the question every software company eventually had to answer was can you build it? That question is closed. Anyone can build anything. The machines have eaten the bottom rung of difficulty and they are climbing the ladder.

The question that has replaced it — the one the next generation of companies will compete on — is older than software and older than industry. Every craftsperson, every filmmaker, every architect, every novelist has answered it for thousands of years.

Was this worth making?

In a world where everything can be made, the only durable advantage is making things someone, somewhere, was waiting to see made well. Made with intent. Made with taste. Made with the kind of attention that signals, without announcing itself, that a human cared about the person on the receiving end. Craft is not aesthetic indulgence. It is service. It is the form respect takes when it gets built into a product, into a film, into a line of code, into a room full of strangers who agreed to look in the same direction.

The film industry's whole business has always been this. The companies that build alongside it have a choice about whether to inherit the standard. The ones that do are going to define the next era. The ones that do not are going to spend it producing infinite mediocrity, wondering why nothing lands, and discovering — too late to course-correct — that abundance was never the advantage they thought it was.

Read this essay on brunohart.dev →

The world's largest companies are pouring their resources into something that does not yet have a settled name. On the surface it looks like building communities — that exhausted corporate word — deliberately assembled mixes of personalities, worldviews, and aligned intelligence, varied enough to throw off their own outliers and multiply thought into more thought. Underneath, it is more strategic than that. It is a contest over whose way of working becomes the default. Whose language wraps the core experience. Whose frame the work gets done inside. The companies winning it have noticed that the unit of competition quietly moved. Not features. Not pricing. Not distribution. Whose worldview the next decade of work is thought through.

Figma makes that case for design — Config keynotes that frame design not as a department but as a way of seeing, a driving force across strategy, climate, and code. Notion makes it for the networked knowledge worker, software modular enough that a company can feel in control of its own values and accumulated knowhow even as it scales at speed. Cursor makes it for programming itself — an agentic-first IDE whose real mission is to keep more people thinking about code rather than fewer, because the close craft of code still underpins every safe, secure, scalable system, including the ones now multiplying without a human at the keyboard. And beneath the named players runs a quieter, broader contest: the war for thought-partner context. Over who sits closest to the user's reasoning. Whose ontology the agent borrows when it reaches to be helpful. Whose vocabulary the next generation of professionals inherits without ever noticing they inherited it.

This is, at root, a domain-specific kind of knowledge, and it answers the question every software company eventually has to face: why do we need to exist? What these companies are building is a new layer of meaning laid over the highway of knowledge we already navigate by habit — and they are competing for your agent's context now as fiercely as they ever competed for your own. We are walking into a world where reasoning itself arrives mediated, through a thought partner whose sense of what matters was shaped by the architecture of the information it was handed. Everyone with something to protect should be competing to own a slice of thought in that layer, or watch the value their field holds quietly erode. The real question is how a company deepens its exchange with the thought partners now sitting between people and their own conclusions — and why certain domains, the ones built on craft and gathered human attention like film and theatre, begin the fight already holding an advantage.

The strategic axis

Industry to date has been built on human value and energy exchange. That will no longer be the inherited case. So we enter a new era — a fight for the relevance of a recognised shared reality, a recognised shared experience of what it means to be human.

The shift is to see the coming wave as a human-first world, not an agentic-first one. Humans are the possibility that gives thought hope, images meaning, words power. Agents are translators of intent — useful, powerful, increasingly indispensable — but translators. The seed, the idea caught like a fish in a swirling sea of maddening information, is inherently human. Value will accrue to whoever builds, creates, designs, and delights for the human-in-the-loop. The next great companies get to architect how that loop can better serve humanity. Not how humanity can better serve the loop.

This is the axis the next decade sorts on. Not vertical versus horizontal. Not on-prem versus cloud. Not even open versus closed. Human-first versus agentic-first. Two architectures that look alike from the outside and could not differ more in what they assume about where meaning begins.

The fish in the abyss

Ideas live in a deep place. They swim in the abyss of spacetime — elusive, independently spirited, indifferent to whether they are caught. Anyone who has tried to catch one knows the feeling: the half-formed thing flickering at the edge of attention, refusing to resolve until you stop chasing it, then arriving whole in the shower, on a walk, two minutes before sleep. Ideas are not produced. They are caught. And the catching is one of the most distinctly human things a human does.

A human-first company is a net for those fish. The whole organisation — its tools, its language, its rituals, its product surface — is a net cast into the abyss in service of the human standing on the deck. Every line of code, every workflow, every agent the company builds is judged by a single question: does this make the fish easier to catch, or does it scare them off?

An agentic-first company is a net cast in empty water. It assumes the agents are the fish. It optimises the loop, the context window, the reasoning chain, the throughput of tokens through a pipeline. A beautifully engineered apparatus with nobody standing on the deck. The catch, when it comes, is theatre — a recursive system making artefacts for other recursive systems to consume, an internal economy of motion mistaken for an external economy of meaning. The most sophisticated agentic-first products are technically remarkable and culturally hollow. They optimise the translator and forget the speaker. They mistake the medium for the message and the loop for the life.

The instinct to build everything around the agent will become almost irresistible over the next few years. As pools of context grow more reliable, more scalable, more maintainable, the temptation will be to let the agent become the centre of gravity — to design org charts, product surfaces, and whole industries around what the agent finds convenient. This is exactly the wrong instinct. It is the instinct of a generation that confused leverage for direction. The agent is leverage. The human is direction. Build a company that optimises the leverage and starves the direction, and you have built a lever with nothing on the far end of it.

What human-first actually means

Human-first is not a marketing posture. It is a bet about where life happens — and the bet is that life happens out there, in people busy living, not in here, inside the machine admiring its own motion. Bob Dylan put the whole philosophy in nine words: he not busy being born is busy dying. A human-first company is built for the being-born half. For the person mid-stride in their own life, who came to the tool to get back to that life faster and fuller — not to be kept inside it.

So it asks, before anything ships, a pair of questions most software never asks: whose intent does this serve, and whose attention does this protect? It builds the agent as a translator of human intent, never a generator of synthetic intent. It curates its context — the data, the ontology, the institutional memory — around a single conviction: the user is the author of the work and the system is the instrument. It refuses, on principle, to design surfaces that hijack attention to feed its own engagement metrics. And it shapes the loop to return the human to their own thinking with more clarity, not less — to hand them back to the morning, the room, the people they love, the work only they can do.

That is the celebration underneath the architecture. Shared breath. Open space. The agent coming to meet the human where the human already is, instead of dragging the human inside the machine to meet it. The robust version of this — the org charts, the API contracts, the context windows the agents will inherit — is the same conviction made load-bearing. Engineering in service of a life being lived, not a loop being fed.

The agentic-first temptation, by contrast, will be loud, well-funded, and seductively easy to demo. It will produce impressive videos. It will close enterprise deals on the promise of replacing rather than amplifying. And it will, in the long run, build products that are efficient, well-lit, and entirely empty of the one thing they were supposedly built to hold — a beautiful room with every light on and no reason for anyone to walk in.

Cinema as the canonical case

Cinema, and the whole world around the pictures, is fundamentally humanity-affirming. It holds open a relationship between intent and experience that thoughtless scrolling erodes — the algorithmic feed, the ambient agentic distraction now engineered into every surface a human looks at. Cinema preserves that relationship by its very form. A film is chosen, not surfaced. Arrived at, not delivered. Attended to in a room built for the purpose of attending. The lights go down, the phone goes away, and a person commits, for two hours, to a single act of focused attention among strangers doing the same.

This is intent and experience in their most concentrated form — the thing nearly every other surface in modern life is quietly degrading. Which hands the companies that serve cinema an asymmetric responsibility, and an asymmetric advantage, in the human-first era. The responsibility is to protect the form. The advantage is that the form is already aligned with where value will accrue. Every other domain has to be retrofitted to remember the human at its centre. Cinema never forgot. The whole industry is a load-bearing wall against the dissolution of intentional attention, and the technology underneath it should be built with that knowledge baked into every interface decision, every API contract, every context window the agents will eventually inherit.

The human-first cinema company is the inverse of the growth-at-all-costs playbook. It does not chase the human into the crowded channel. It builds the room the human chooses to walk into. It treats attention as a gift to be earned, not a resource to be strip-mined. It uses its AI not to manufacture more noise but to make the signal — the film, the showtime, the seat, the moment of arrival — more legible and more worth arriving for. The loop serves the evening. The evening serves the human. And the human, in turn, serves what humans have always served: each other, the stranger in the next seat, the people who will fill the room next week, the future the room is held open for. That is the only correct ordering. Leverage at the bottom. Life at the top.

Coda

Which is why craft matters now in a way it never quite has before. When agents can produce anything, what stays valuable is what is produced well — and what is produced in service of another human, with the care only another human can offer. Film has always been this. A film is craft assembled by hundreds of people on behalf of millions they will never meet, projected in rooms built to receive it. The whole product is craft. The whole service is care. In an era where the cost of producing artefacts has collapsed and the value of producing meaning has not, the industries whose entire output is craft and service stop being cultural curiosities. They become the leading edge of what the next economy is going to value. The human-in-the-loop is the seed. Craft is what the seed becomes when a human, with skill and intent, decides it is worth becoming something. That is the next argument, and the one this collection turns to now.

Read this essay on brunohart.dev →

The Ritual

A schedule, posted in advance, pulls strangers into the same hour. They arrive at a building made for one purpose. They choose a seat from a map of a room they are about to inhabit, surrendering a small piece of personal geography to the geometry of the space. They pay for something hot and salty that nobody is hungry for — popcorn as ceremony, sacrament adapted for secular use. They walk a corridor lined with posters advertising the next two months of attention, push through a door, and the light drops. Conversation thins to a hush. The room agrees, without anyone speaking the agreement, to look the same way for two hours.

This is a protocol. Not in the technical sense, though that follows. In the older sense — a procedure of assembly, choreographed by convention, rehearsed by repetition, held up by mutual willingness. Cinema is one of the last protocols of public assembly that human beings still physically inhabit at scale: in person, on a schedule, with strangers, no belief required, no artist required, no team required, nothing required except the screen, the seat, and the willingness to look one way for the length of a story.

Most of the institutions that used to do this work are in retreat. The architecture remains. The function has thinned.

What Used to Do This Work

The category cinema belongs to is older than cinema. The category is broadcast ritual — the periodic, scheduled, embodied gathering of strangers around a common focal point. Parliaments did it: the civic body in one room, oriented toward one speaker, bound by procedure to share attention for the length of a session. Houses of worship did it, weekly, toward an altar or a pulpit. Theatres did it. Opera houses, concert halls did it. The town square did it whenever a crier or a preacher or a politician on a wooden box pulled the population into a single oriented mass. Broadcast television did it for forty years — less physically, no less collectively — the whole country watching the moon landing together, the same news at six, grieving the same death on the same evening because no other channel was on.

These institutions did not survive abundance equally.

Parliaments persist as architecture and weaken as ritual; the chamber still stands while the gathering around it fragments into clipped posts and partisan feeds, watched alone. Worship attendance across most of the secular world has been in a forty-year decline no demographic has reversed. Theatres survive but have priced themselves into a tourist economy. Broadcast television, the great twentieth-century engine of synchronous national attention, has been disassembled by the streaming logic that swapped everyone watching the same thing at the same time for everyone watching a different thing, alone. One by one, the places that assembled strangers into shared focused attention have disappeared, gentrified, atomised, or moved indoors into the private feed.

The attention economy — the actual term of art for what happens now — runs on solitary devices. The phone in the pocket. The laptop on the desk. The headphones over the ears. The feed that reshapes itself for an audience of one. Almost every screen a human now looks at knows their name and addresses them alone. The category of public, oriented, synchronous, embodied attention has been hollowed out from every direction.

Cinema is the surviving exemplar.

Not because it is the largest. Not because it is the most economically significant. Because it is the one remaining instance of the category that still works in its native form — strangers, screens, schedules, silence, two hours of synchronised focus in a dark room with people they will never meet again. Concerts come close, but need a specific artist, alive and touring. Sports come close, but need a team and a tribe. Worship comes close, but needs belief. Theatre comes close, but runs at a fraction of the scale and a price that shuts most of the public out of regular attendance. Cinema asks for nothing but attention, and gives back a ritual that costs the price of a meal and runs three times a day on a Tuesday in every city on earth.

The Function the Market Cannot Price

What gets lost in the ticket-price discussion is that broadcast ritual is a function, and functions have consequences that never appear on the income statement of the building performing them.

Sociologists have a name for what happens when strangers share focused attention in a physical room: collective effervescence, Durkheim's term for the heightened emotional state that synchronised group experience produces. It is why a crowd at a concert feels nothing like the same concert on YouTube. Why a packed church on a holy day feels nothing like a podcast of the sermon. Why laughter in a cinema is louder, longer, more contagious than the same comedy streamed alone. The human animal is wired to register the presence of other humans attending to the same thing, and that registration lays down social bonds — weak ones, mostly, between people who will never speak — that aggregate into something durable at the scale of a society. Trust. Tolerance of difference. The baseline assumption that the stranger in the next seat experiences a story the way you do.

This is the social-cohesion argument, made in the academic literature for a century. It does not appear in the financial statements of cinema chains because nobody has figured out how to price the externality of public assembly. The cinema sells popcorn and tickets. What it actually produces, as a byproduct of selling popcorn and tickets, is a small, repeatable, scheduled instance of strangers practising the muscle of shared attention. A society that practises that muscle is measurably different from one that doesn't. Trust networks form differently. Public discourse calibrates differently. The fabric that lets a democracy function at all — the assumption that the person you disagree with is still in the same room — is the fabric that broadcast ritual weaves, screening by screening, schedule by schedule.

Strip the broadcast rituals out of a society and the society does not collapse. It thins. It gets harder to assemble for the things that need assembly. It gets easier to assume the people on the other side of the screen are not quite people. The decline of every other institution in this category, over forty years, maps almost perfectly onto the decline of the civic conditions those institutions existed to sustain. Correlation is not causation, and the argument is not that cinema alone holds the line. The argument is narrower and harder to dismiss: cinema is one of the few surviving instances of the kind of thing that holds the line, and what survives of a thinned category is load-bearing in a way it never was when the category was crowded.

The cinema across the street from the apartment block is civic infrastructure performing a sociological function its owner is not paid to perform and may not know they are performing. The Tuesday matinee is public assembly the city did not have to fund. The midnight horror screening is a hundred strangers practising the shared startle reflex of the species. These are not entertainments. They are exercises in the kind of attention that builds the kind of society most people say they want to live in.

What This Makes the Work

On the surface, the technology underneath this protocol does not look different from any other operational software. Session schedules. Seat inventory. Order processing. Loyalty. The vocabulary of any consumer-facing logistics system — and the temptation, especially from a horizontal-platform view, is to treat it as exactly that: assume what cinema operators need is what every operator needs, with cosmetic tweaks for the vertical.

That is the misunderstanding. The buildings these systems serve are not retail outlets that happen to sell time-based attention. They are the surviving architecture of a category of human practice almost nothing else still performs. A scheduling engine deciding which films play when across a circuit is setting the rhythm at which a city assembles for shared experience. A seat-map renderer offering a customer where to sit is mediating a small, real act of physical commitment to a public room. And a ticketing system that fails on a Friday night is not logging a lost transaction — it is generating the friction that nudges a marginal viewer toward staying home and watching alone, and the viewer who stays home tonight is incrementally less likely to come back next month.

The operators running these buildings, the designers shaping the interfaces, the engineers writing the systems beneath — they are participants in the maintenance of civic infrastructure, whether or not their employment contract says so. The market cap does not capture this. The ticket price does not capture this. The B2B SaaS language often bolted onto the category, with its throughput and conversion and ARPU, does not capture this. None of those frames are wrong; all of them are partial. The fuller frame is that the work belongs to the same lineage as running parliaments, maintaining places of worship, programming public broadcasters, designing town squares — the lineage of building and sustaining the architecture through which a society practises being a society.

A category that has lost most of its members concentrates its civic weight on the survivors. Cinema, as the surviving exemplar of broadcast ritual at scale, is no longer one option among many for the function it performs. It is, increasingly, the option. That changes what is owed to it — by its operators, by its technologists, by the cultural and economic machinery that decides which work is taken seriously and which is treated as ambient.

Coda

The seat map on the screen is a map of a room a stranger has not yet walked into. The schedule on the wall is a list of times at which strangers will assemble. The dark, in five minutes, will fall over a hundred people who are about to share an experience the algorithm and the feed and the headphones cannot reproduce. None of this is nostalgia. The phone in the pocket did not kill it. The phone in the pocket made it singular. What was once one ritual among many is now the one that still runs on a Tuesday afternoon in every city on earth, performed by anyone who shows up, asking of them only their attention and their willingness to share it.

The infrastructure beneath this protocol is, by inheritance, civic. The people building it are, by participation, building something older and stranger than software. The stakes have nothing to do with throughput. They have to do with whether the surviving instance of a load-bearing human practice keeps functioning at the scale a society needs. That is the work. That is what it has always been the work of. The category has narrowed. The responsibility has not.

Read this essay on brunohart.dev →