"It seems to have been reserved to the people of this country … to decide the important question, whether societies of men are really capable or not of establishing good government from reflection and choice, or whether they are forever destined to depend for their political constitutions on accident and force." — Alexander Hamilton (Publius), The Federalist No. 1, 1787

When the course of human life moves online — when our friendships and our work, our memory, our money, and our public square all come to live on the internet — then the terms on which that internet is owned and governed become questions of human freedom. And when those terms have, over years, been set against the very people who give the internet its life and its worth, a decent respect for one another requires that we say plainly why we mean to change them.

We hold these truths to be self-evident:

That the data each person generates is the shadow of a real life — it would not exist if that person did not — and it belongs, first and last, to them.

That the value created by the people of a network is created by them, and ought to flow to them.

That platforms, like governments, derive their just authority from the consent and the participation of those who use them.

That when a platform becomes an instrument of extraction rather than of service — taking without returning, ruling without accounting — the people who sustain it have the right, and in time the duty, to remake it, or to build in its place an internet that answers to them.

That these rights belong to no single nation, creed, or class, but to every person who lives a part of their life online.

The history of the present internet is a history of quiet expropriation — not by malice in every case, but by a structure that has, link by link and click by click, turned the many into the product and the few into its owners. To prove this, let these be submitted to a candid world:

• They have gathered the digital shadows of billions, and asked no leave, and returned no share.
• They have built vast fortunes, and now vast artificial intelligences, upon that gathered data, and have shared neither the wealth nor the say with the people whose lives that data describes.
• They have written the rules of our common spaces in private, and changed them at will, and answered to no one but themselves.
• They have watched us without ceasing, and called the watching a service.
• They have made our attention their inventory, and our trust their leverage.
• They have gathered power over speech, commerce, and knowledge into a few unaccountable hands, beyond the reach of those that power governs.
• They have built for the next quarter what was meant to last for generations, and called the strip-mining of it growth.
• And they have taught us to call all of this normal — to click I agree as though we had a choice.

Set these down together and they do not describe a service. They describe a reduction: a people who should be We the Users have been made, click by click and link by link, into we the used.

No one chose this. It was built; it can be built differently.

We, therefore, the users of the internet — its readers and its writers, its makers and its menders, the people whose lives fill it — publish and declare our intent:

To build a new internet that belongs to the people who use it: where each person holds their own data; where the rewards of the network flow back to the network; and where no founder, no engineer, and no investor is set permanently above the rest.

To build it for the long horizon — to measure it not by the next quarter but by the next thousand years, on the conviction that only an internet owned by the people who use it is built to last that long.

To settle its governance not by accident, and not by force, but, in Hamilton's words, by reflection and choice: to write its constitution in the open, in full public view, with framers invited from among those who have given themselves to a freer and fairer digital world.

To open this work to everyone, and to make belonging a matter of affirmation and participation — not of birth, or border, or wealth.

We do not pretend to hold every answer. The hardest questions are written down plainly, and they are the work now before us. But we know enough to begin, and we mean to begin together.

In support of this Declaration, with a firm reliance on one another, we pledge our effort, our good faith, and our names.

A provisional draft — not yet declared, and not yet signed. Offered to everyone who believes there is a better way: to argue with, to improve, and one day to ratify together, community by community.

Most of us have made a quiet peace with the internet. We know, somewhere, that the feed is built to keep us scrolling; that the "free" service is paid for with our attention and our data; that the companies we depend on answer to their shareholders, not to us. We know it, and every day we click I agree anyway — not because we agree, but because there is no other button, and it all seems to be simply how things are.

But no one chose this — we only clicked agree. Every part of the internet we touch was built by someone, for a purpose, and the purpose was rarely us. The bottomless feed, the data taken without asking, the rules written in private and changed at will — none of it is a law of nature. It was chosen, because it paid a few people very well. And what was chosen can be chosen differently.

What follows is an argument that we should, and that, for the first time in a long while, we actually can.

The shadow you cast

Begin with what you make. Every day, you leave a trail: the things you search and buy and watch, the messages you send, the places you go, the people you know, the moods you let slip at one in the morning. Call it your digital shadow. It is not you, but it could not exist without you. It is cast by a single real life, yours, and no one else's.

That shadow turns out to be worth a great deal. It is what lets a company predict what you will click, what you will buy, what you can be persuaded of, what you can be sold. Gathered across billions of people, these shadows are the raw material of some of the most valuable enterprises in history. And lately they have become something more than an advertising tool. They are the fuel of artificial intelligence — systems trained on the collected output of humanity that are beginning to reshape how nearly everything is made and done. The wealth of that intelligence is flowing, overwhelmingly, to the few companies that happen to hold the data.¹ Not because those companies are wiser than the rest of us. Because they hold the shadows.

That is the quiet bargain at the heart of the modern internet: you produce the value, and someone else keeps it. We the Users have quietly become we the used — a bargain never struck, only assumed.

Nothing natural about it

We have been taught to treat this as a kind of natural law — as though it were always bound to be the case that a few enormous firms would own the public square, the marketplace, and the running record of our lives. It is worth recalling how new this is, and how thoroughly it was chosen.

The internet did not begin as a machine for watching people. It began, not so long ago, as something closer to a commons — a place its early builders hoped would be open, decentralized, and free.² The surveillance, the consolidation, the sense that five or six companies simply are the internet now: these arrived later, and they arrived because someone decided that the surest way to make money was to harvest attention and sell prediction. It was a reasonable decision, given the incentives. It was still a decision. The engineers who built the feed to hold you could have built it to serve you. The businesses that kept the value could have been arranged to share it. None of this was handed down by nature. All of it was authored, and what is authored can be rewritten.

We know rewriting is possible, because pieces of the alternative already exist. Much of the internet's own foundation runs on software built by communities and owned by no one — given away, maintained in the open, relied upon by everyone. Wikipedia, written by strangers for free, outlasted the encyclopedias that were sold. And beyond software, the longer human story is largely the story of people taking power that looked permanent — that called itself natural — and making it answer to them. The means to own things together, to share what a network earns among the people who actually make it, are more available now than they have ever been. What has been missing is not the tools. It is the decision to use them.

What is yours

So let us say plainly what a fairer internet would look like.

Your data would be yours — held by you, not extracted from you. You would decide who may read it, write to it, and share it, and you would be free to take it and leave. The value the network created would flow back to the people who created it — not as a kindness handed down, but as a matter of ownership, the way a return belongs to an owner. The people who built and ran the thing would be paid, and paid well, but no founder, no engineer, and no investor would be set permanently above everyone else, drawing rent on a public square forever. And the rules of that square would be written in the open, by the people who have to live under them.

This is not a dream of an internet without companies, or without builders rewarded for building. It is the more modest and more radical idea that the rewards should not all collect at the top and stay there — that the people we still call "users," as though they were merely consuming something, should be recognized as what they actually are: the owners of the value they create.

"But it cannot be changed"

The honest objection to all of this is not that it would be bad. It is that it cannot be done — that the incumbents are too large, their advantages too entrenched, and the rest of us too scattered to do anything but accept the terms we are given.

That objection is serious, and it deserves a straight answer instead of a slogan.

Here is the straight answer. It will be hard, and there is no shortcut. A new network is least useful at the very moment it is smallest, when there is almost no one there yet, and becomes worth joining only once many people already have. There is no way to wave that problem off, and anyone who tells you otherwise is selling something. So we will not pretend that a finished, better internet is sitting on a shelf, waiting for you to switch to it. It does not yet exist. It has to be built, and it has to be built by the people who want it.

That is, in fact, how every founding works. The American Patriots did not wait until they had an army and a treasury to declare what they believed; they declared it first, and then they built what the declaration demanded. We are going in that order on purpose. The better tools, the shared rewards, the intelligence that belongs to the people it was trained on — those come after people gather, not before. A document like this one is not the finished country. It is the gathering.

And what makes this a moment rather than merely a wish is timing. The terms of the next internet are being set right now, as artificial intelligence reorganizes the whole digital economy and, with it, the question of who owns what. The settlement that hardens over the next few years could hold for a generation. It can harden in favor of the few who already hold our data, or it can be built, deliberately and in the open, for the rest of us. That window is still open. It will not stay open for long.

What we can do now

On the second of July, 2026 — two hundred and fifty years after a smaller group of people decided they would no longer be governed without their consent — we will publish a Declaration of Internet Independence. It makes no pretense of having already built the new thing. It says, plainly, that the internet should belong to the people who use it, that we intend to build it, and that we are inviting framers to write its constitution out in the open, where everyone can see.

You do not have to be an engineer, or an expert, or anyone in particular to be part of this. You only have to decide that the way things are was never really chosen — that it was built, and can be built differently — and to put your name beside the others who mean to build it.

That is how every founding has ever begun: not with a finished nation, but with a single sentence that enough people were willing to sign.

No one chose this. Let us build something better.

The shadow you cast

You cast a shadow now that you cannot see.

Every day you leave a trail — the things you search and read and buy, the messages you send, the routes you walk, the people you know, the hours you keep, the moods you let slip at one in the morning. Each trace is small. Gathered and kept, they form something larger: a detailed, searchable, predictive record of a single life. Yours. It is not you, but it could not exist without you, and it belongs to no one else the way it belongs to you. Call it your digital shadow.

For most of history a person's shadow was faint and forgettable; what you did on Tuesday was gone by Friday. What is new — what has changed everything in a single generation — is that the shadow is now permanent, copyable, and computable. It can be stored forever, combined with millions of others, and read by machines that find in it the patterns a human eye would miss: what you will want, what you will click, what you can be persuaded of, what you are worth. A shadow that can be read this well is a kind of power, and right now that power belongs almost entirely to whoever holds the data¹ — not to the person who cast it.

This essay is about taking the shadow seriously: what it is made of, and what rights a free person ought to have over each part of it. Because we have never really decided. We carry a second self around with us, growing sharper by the year, and we have almost no settled say over who may look at it, add to it, change it, keep it, or sell it.

Why "it's mine" isn't enough

The natural first answer is: it's mine. My data, my shadow, my rights.

It is the right instinct, and it is too blunt to build on. Look closely and the shadow refuses to be a single thing you simply own. A private message has two authors, not one. The tie between two friends belongs to both of them. A great deal of the shadow you did not even make: it was inferred about you — guessed, scored, and filed by a company you never spoke to, from patterns you never saw. "Ownership" assumes a clean line between what is yours and what is not, and the shadow has no such line.

So it helps to stop asking who owns the data and start asking a sharper pair of questions: which rights, over which kind of data, held by whom?

Rights, because what we loosely call ownership is really a bundle of separate powers² — the way owning a house is really the bundled rights to enter it, change it, rent it, sell it, and keep others out. For data, the bundle includes the right to see it, to add to it, to correct it, to erase it, to hold it (to keep the only copy), to share it onward, to use it (to compute on it, including to train an artificial intelligence), to profit from it, and to carry it away and leave. These can be held separately, and by different people.

And held by whom, because more than one party can have a real claim: you, who cast the shadow; the others who appear in it alongside you; the network that records and infers; and sometimes the public, when a thing is genuinely shared in common.

The work, then, is to walk the regions of the shadow and ask, for each, how that bundle should be divided.

The regions of the shadow

What you make. The clearest region, and the one with the strongest claim. The photographs you take, the words you write, the messages you compose, the things you create and post: these you authored, deliberately, and the whole bundle should default to you — see, correct, erase, carry, and profit, with use by anyone else only on terms you grant. This is the part of the shadow that feels most obviously like property, and the law half-agrees. It is also, increasingly, the smallest part.

What you do. Far larger is the exhaust of simply living online: what you click, how long you watch, what you scroll past, what you search, what you buy. You generate it without meaning to, as a byproduct of using a service, and that service has some honest claim to see it, because it cannot function otherwise. But the contested powers here are use and profit, and today the network quietly takes both: your behavior is studied and sold, and you see neither the study nor a cent of the sale. A fair settlement leaves you the right to see this record, correct it, erase it, carry it, consent to how it is used, above all whether it trains advertising or AI, and to share in the value it creates.

What you are, and where you are. A region apart, because of its sensitivity: your location over time, your face and your voice, your fingerprints and your heartbeat, your health. Much of this you did not author so much as emit, and to gather it is to follow a body through the world. It deserves the highest protection — collection minimized and opt-in rather than assumed, custody held by you by default, erasure easy and real, and a right the other regions barely require: the right not to be measured at all.

What's inferred about you. Here is where the power hides. From everything above, the network computes things you never told it: your likely politics, your vulnerabilities, your "type," your predicted next move, a score that quietly decides what you are offered and at what price. You did not author this region, the network did, and yet it is entirely about you, and it is often the part that matters most to how you are treated. Today you usually cannot even see it. A free person should be able to see what has been inferred, correct what is wrong, erase it, and consent to whether such inferences may be drawn and used at all. The network made it; but you are its subject, and being its subject is itself a claim.

What's between you and another. Private messages, transactions, the mutual tie that says two people know each other: these have two rightful claimants, and neither owns them alone. You can delete your copy of a conversation, but whether you can erase it from the other person's record of it is a genuine question with no clean answer. This region cannot run on sole ownership; it needs joint stewardship — rules of consent worked out between the parties, the kind of thing a network's members would have to agree among themselves rather than have imposed on them.

What others say about you. Finally, the things other people author about you: the photo they tagged, the review they left, the mention in someone else's post. Two interests collide here — their freedom to speak, and your stake in your own reputation. You should be able to see what is said and to respond, but your right to erase another person's speech is rightly narrow, reserved for harm rather than mere discomfort. This is the hardest boundary on the whole map, and the one most easily abused from either side.

Set side by side, the pattern becomes a table, and the table is the argument:

Underneath every row sit two powers the table does not show, because they apply everywhere. Custody, who physically holds the data, should default to you, hardest of all for the region of body and place; the network should keep only what it needs, in plain view, and let you withdraw it. And portability, the right to take any of this and leave, is the backstop that makes the rest real, because a right you cannot walk away with is a favor, not a right.

There is one more region, and it belongs to no one alone. When millions of shadows are pooled, the pool takes on properties no single shadow has — the shape of a whole population, and the intelligence that can be trained on it. This is the shadow we cast together, and it is where the value of modern artificial intelligence actually comes from. No individual rights table governs it, because it is a commons. The principle that should govern it is this movement's whole reason for being: it ought to be stewarded in common, and its rewards ought to flow back to the people who cast it³ — not pool, as they do now, with the few who happened to hold the data.

The hard edges

None of this is tidy, and an honest map shows where it tears.

Relational data is the deepest problem⁴: when a thing belongs to two people, whose consent governs it? If you want a shared conversation erased and the other person wants it kept, someone's right must yield, and no rule satisfies everyone. Inferred data raises a stranger question still: can a model truly forget what it learned about you, once your shadow has dissolved into its weights? Deletion there may be less like tearing up a file than like asking someone to unlearn a face. The line between your control and other people's speech will be contested without end. And some of the shadow is genuinely public — said in the open, on the record — and cannot simply be reclaimed.

These are not reasons to abandon the framework. They are the framework's agenda: the questions a constitution for a user-owned internet would have to answer in the open, deliberately, rather than leave to whoever happens to write the terms of service. They are hard precisely because they matter.

What a shadow could do for you

It would be a thin victory to win all these rights and use the shadow only to defend yourself. The deeper promise is that a shadow you actually hold becomes something you can use — an asset, and an instrument of freedom.

Hold your own shadow and you can lend it on your terms: grant a company, or an AI you trust, the use of part of it for a season, and be paid for what is today simply taken. Carry it and you can leave — move to a better service without abandoning your history, your reputation, your connections, which is the lock that keeps people inside platforms they have long outgrown. Keep it close and you can train an intelligence on it that works for you — a model that knows your life and answers to you alone, rather than to the company that assembled the profile. And you can pool it on purpose: thousands of people choosing to combine their shadows into a common fund, a data union⁵, with the bargaining power, and the shared return, that no individual has alone.

A shadow you do not control is only a liability, a thing that can be used against you. A shadow you hold is a kind of capital, and a kind of liberty.

The first possession

A free people building a digital world of their own will have to decide many things. This is the first of them. Before the question of how the network is governed, or how its rewards are shared, or who is allowed to do what, comes the question of the self each person carries into it — the shadow they cast, and whether it is theirs.

The Declaration says that it is: that your data is the shadow of a real life, and belongs, first and last, to you. This is the argument beneath that line. The shadow has a shape; the shape has regions; each region carries rights that someone will hold whether we choose them or not. Choosing them — deliberately, in the open, for the person who casts the shadow rather than the company that captures it — is where a free internet begins.

"If men were angels, no government would be necessary. … Ambition must be made to counteract ambition." — James Madison (Publius), The Federalist No. 51, 1788

The thing we got wrong

The companies that own today's internet hold four kinds of power at once, and that is the whole problem.

They control the machinery — the servers, the code, the keys. That control gives them power — the ability to decide what you see, what you may say, who is let in and who is cast out. They have made that power pay, and the money has entrenched them past the reach of any rival. And they have granted themselves the authority to set all the rules, by the simple device of a terms-of-service agreement no one reads and no one negotiates. Control, power, money, authority — gathered into the same few hands, answerable to no one but their shareholders.

We have lived under this so long it looks like nature. It is not nature; it is a design, and a bad one. The deepest thing a new internet has to get right is not any single rule but the separation of those four powers¹, so that running the software does not also mean ruling the people who use it, and so that money cannot simply buy the rest. This is the oldest lesson in the book on government, and it is the one the internet skipped.

Four kinds of power

It helps to keep the four straight, because they are easy to confuse, and the confusion is exactly where domination hides.

Control is operational command: whose hands are on the machine. Authority is the recognized right to make a rule or settle a dispute — power that people accept as legitimate. Power is the raw capacity to make others comply, legitimate or not. And money is the value moving through the system, which, left unchecked, converts into all three.

A free internet needs all four to exist — someone must run the software, someone must settle disputes, the thing must be paid for — but it needs them held apart, in different hands, checking one another. The design that follows is one long attempt to do exactly that.

Three tiers

The structure has three tiers, and you have met them: Individuals at the base — people, each with a name; Communities in the middle — the places people gather, each with its own name and its own rules; and above them a single thin federation: the Commonwealth. The pattern is the one the American founders reached for — federalism: a citizen, a state, and a federal government, each holding powers the others do not, with the contested questions pushed down to where the people they affect can settle them.

The whole design can be read off one table — who holds what:

Read down the columns and the separation comes into focus. The Individual holds the self. The Community holds the rules of its own house. The Commonwealth holds only the thin shared floor, the protocol, and a capped fee — deliberately the least of the three, because it is the one place where power could re-concentrate. Most of what makes the old internet oppressive — the rule-setting, the moderation, the revenue model — lives at the Community tier, close to the people, where it can be argued with and walked away from, not dictated from a single center.

Friendly rivals

Communities compete for members. That is a feature, not a danger — if the competition is the friendly kind. The aim is a world where the way to win members is to be better to belong to, and where the moves that hurt people simply do not pay.

The master mechanism is cheap exit. Because the Covenant guarantees that your data, your identity, and your connections come with you, a Community cannot hold you hostage; the only way to keep you is to earn you — with better terms, better moderation, a better share of the rewards. Everything else in the floor exists to keep that exit real: mandatory, comparable statistics so you can see how Communities actually treat their members; duties to interconnect so no Community can wall itself off; limits on concentration so none can swell into the next incumbent.

That single mechanism turns the rivalry from destructive to constructive:

The principle underneath is plain: friendly competition is just what a market looks like when leaving is easy, and unfriendly competition is what it looks like when leaving is hard. Keep exit cheap, and the rivalry stays friendly on its own — not because Communities are virtuous, but because predation stops paying.

Who belongs

Two questions of membership have to be answered honestly: who belongs to a Community, and who belongs to the Commonwealth at all.

For an Individual, joining the Commonwealth is by affirmation, and it is nearly irrevocable — no one can be exiled from the network itself, and your shadow and your identity are always yours. Joining a Community is a different matter: each one sets its own standards and may remove people who break them, but only within the floor, which guarantees due process (notice, reasons, a chance to answer, an appeal) and forbids any removal that confiscates your data or your portable identity. This is what makes communities-with-standards tolerable. Being removed from one Community is not exile; you leave with your shadow and join or found another. Exclusion is never digital death, and that single guarantee is what lets Communities differ sharply from one another without anyone being trapped or erased.

For a Community, the door works differently. Forming one is open to anyone. But to join the Commonwealth — to interconnect with the rest and draw on the shared protocol and the shared rewards — a Community must enter the Covenant: it commits, in advance, to the floor, to portability, to transparency, to the duties of interconnection, and to the terms under which it may itself be severed. That last point is the quiet key. Because the rules of severance are part of the agreement every Community accepts on the way in, a Community later cut off for breaking them cannot cry foul — it consented to those terms when it joined. This is consent of the governed, applied to the Communities themselves.

It is also where this design can learn from the one the founders left unfinished. The U.S. Constitution never specified whether or how a state could leave, and the silence was answered, eventually, by a civil war. Writing the terms of severance calmly into the founding agreement — in advance, in plain view — is how the Commonwealth declines to inherit that wound. Severance, when it comes, is not a power seized by the center; it is the enforcement of a rule the Community already agreed to, reached through due process, and even then the Community's members keep their portability, so no one is trapped by the punishment of their Community.

The money

Money is the solvent that dissolves every other safeguard, so the design watches it closely. The Commonwealth funds itself with a fee, a percentage of the activity it carries, but the fee is capped, its books are open, and it is set by the Commonwealth's governing bodies rather than by whoever happens to run the software; the honest expectation is that it should fall as the network grows, not rise. Communities choose their own revenue models — advertising, subscription, something not yet invented — and compete on how much of the proceeds they hand back to their members. And at every tier, the returns to founders, engineers, and investors are capped: they may be paid, and paid well, but none draws permanent rent on a public square. The throughline is a single rule stated four ways — money must not be convertible into authority.

Ambition against ambition

Who, then, holds the authority the old companies kept for themselves? Not one body, on purpose. The Commonwealth's governance splits the work the way a constitution splits a government: those who set the value-laden rules — where the floor sits, what counts as the unfriendly kind of competition — held apart from those who interpret the Covenant and judge whether it has been broken, and both held apart from the operators who run the software and merely execute. No one of them legislates, judges, and runs the machine at once; each can check the others. Exactly what those bodies are, and how each is chosen — a panel drawn by lot from the membership, say, so ordinary people and not only insiders hold real authority²; a seasoned body to interpret the rules; rotation and fixed terms against entrenchment — is among the most important things the framers have to work out. The principle is fixed; the design is theirs to get right.

Beneath all of it sits the deepest check, the one no incumbent ever faced: the software is open, and the Commonwealth can be forked. If the Commonwealth itself betrays the Covenant — if the center tries to become the very thing we left — the Communities and their members can take the protocol and route around it. The operators serve a public that can leave them. That is the final reason control cannot harden into domination here: in the end, even the center governs only by consent.

This is what Madison meant, arguing for the Constitution, when he wrote that the great security against a concentration of power is to give each part the means to resist the others — that ambition must be made to counteract ambition. We are not building a system that depends on good people staying good. We are building one in which the structure itself keeps any part from ruling the rest.

The hard problems

None of this is settled, and the honest edges are sharp. Who decides where vigorous competition becomes predatory, and can that line be drawn clearly enough to enforce? How is a Community's freedom to set its own standards squared with an Individual's right not to be excluded unfairly? How binding can the floor be before "federalism" becomes dictation from the center, and how loose before the floor stops meaning anything at all? What keeps the Commonwealth's power to sever from becoming a tool of capture dressed as enforcement? How is the fee set so it neither starves the center nor lets it grow fat, and should it ever cross-subsidize weaker Communities, with all the dependence and gaming that transfers invite?

These are not afterthoughts. They are the agenda of the Covenant — the questions the framers must answer in the open, deliberately, rather than leave to whoever ends up writing the code.

A structure, not a savior

The genius of 1787 was not that it produced wise rulers. It was that it produced a machine in which power was set against power, so that the wisdom or folly of any one person mattered less. That is the bet here, carried onto the internet: not a benevolent company, not a charismatic founder, not a flawless set of rules, but a structure — Individuals who can leave, Communities that must earn them, a Commonwealth held thin and checked and forkable — in which no part can rule the rest, and the ambition of each is bent, by design, toward the people they are all meant to serve.

The objection worth taking seriously

Every proposal to build something in common runs into the same objection, and it deserves a real answer rather than a slogan. The objection is that competition is what makes an economy work — that the rivalry of firms chasing profit is the engine that lowered prices, raised wages, and filled the shelves, and that anything built on "collaboration" is soft-headed, a charity that will be out-produced by the hard discipline of the market. People who say this are not wrong about the past. Competition under capitalism really did raise the standard of living for enormous numbers of people, over a long time, and any honest case for a different internet has to begin by admitting it.

So this essay does not argue that competition is bad, or that we should abolish it, or that prices are a trick. It argues something more specific, and harder to wave away: that a collaborative system can do everything a competitive one can — including run a competition whenever competition is the better tool — and can do more besides, and can share the result more evenly. Not a softer economy. A larger one, more fairly held.

Two systems, and the one we're not going back to

Start with the choice we are usually offered. On one side, communism and socialism, where wealth is shared, but shared by force. The state takes and redistributes; the sharing is not chosen, it is enforced. On the other side, capitalism, whose real improvement was to remove that coercion: it left the getting and the giving to individuals, acting freely. Whatever else divides these systems — and people have argued it to exhaustion, more ably than I can here — that is the line that matters for us. We are not proposing to step back across it. Coerced sharing is not the model. Whatever we build has to be voluntary, or it is not worth building.

But removing coercion did not make capitalism perfect, and pretending otherwise is how the conversation usually goes wrong. There is a flaw, and naming it precisely is where the argument starts.

The flaw is not inequality. It is the leash.

People who worry about capitalism usually point to inequality — the gap between what the top earns and what the bottom earns. But a gap, by itself, is not the deepest problem. Two people can earn very different amounts and still stand as equals, each free, neither holding power over the other. Inequality of income is a gap. What should worry us more is inequality of a different kind: the steady drift toward a world in which a growing number of people are in debt to a shrinking number of people.

Debt is not just a number. A debt is a leash. It gives one party a standing claim on another's future work — your time, years out, already owed before you have lived it. When more and more people are on the leash and fewer and fewer hold the other end, something has gone wrong that no rise in the average standard of living repairs. The economists Atif Mian and Amir Sufi have traced¹ how much of modern economic pain runs through exactly this channel: it is the concentration of debt, not income alone, that decides who absorbs the losses when things go bad, and it is the debtors, every time. The power that one set of people comes to hold over another through debt is the thing a better system should set out to reverse.

Reverse it, that is, by consent — not by seizing and redistributing, but by building an economy whose ordinary workings spread ownership outward instead of gathering it in. Which brings us to how.

The bigger circle

Here is the central idea, and it is simpler than it sounds. A collaborative system can choose to hold a competition. A competition cannot choose to collaborate.

Think about what that means. If a group of people working toward a goal decide that the best way to get there is to compete — to race, to let rivals push each other, to reward whoever does best — they can simply do that. The competition runs inside the collaboration, as a tool the group picked up because it worked. But go the other way: take a pure competition, a set of rivals each out only for themselves, and there is no neutral ground on which they can decide to collaborate. The moment they coordinate, the competition is no longer pure. Collaboration is the bigger circle; it can draw competition inside it. Competition is the smaller one; it cannot contain the whole.

This is what dissolves the objection we started with. We do not have to give up what competition is good at. A collaborative internet keeps as much friendly competition, and as much honest pricing, as it takes to get at least the full benefit a competitive market would have produced. Where rivalry sharpens the work, we let people compete. Where prices carry information that no planner could gather — and they do; this is the one thing the defenders of markets are most right about — we keep the prices. We give up none of competition's efficiency. We simply refuse to stop there.

Because there is a second kind of problem, where stopping at competition leaves real value on the floor.

Two races

Picture a track, and a group of runners with a shared goal. Two different goals, actually, and the right way to organize the runners depends entirely on which goal it is.

Suppose the goal is to get every runner across the finish line in the shortest possible time, and each runner has to cover the distance on their own legs. Then a friendly race is the best thing you can do. Everyone starts together, and the rivalry pulls each of them faster than they would have run alone — pushing one another, cheering one another on, every personal best lifting the group's time. This is competition at its finest, and a collaborative system, faced with this goal, would choose exactly this. There is no tension here at all.

Now change the goal. Suppose what you want is the greatest total distance covered in a fixed amount of time, and runners are allowed to tag in and out. The best arrangement is no longer a race. It is a relay: your fastest runner runs flat out while the others rest, then hands off to the next, who runs while the first recovers — each doing what they do best, no one wasting strength on ground a fresher runner could cover faster. The total distance a relay covers will beat what you would get by making every runner grind out the whole course alone. And a purely competitive rule — everyone runs the entire distance by themselves — produces exactly that worse result. It forces duplication: ten runners each running the same miles, when the miles only needed running once.

This is one of the oldest findings in economics, dressed in running shoes. Adam Smith called it the division of labor; David Ricardo called it comparative advantage²; the point in both is that pooling effort and letting each person do what they are best at produces more than everyone doing everything separately. The relay is that idea in motion. And the lesson of the two races is the lesson of the bigger circle: a collaborative system can run whichever race the problem calls for, while a competitive system is stuck running the second one as if it were the first — making everyone cover the whole distance alone, and calling the wasted effort "competition."

So the real question becomes: which kind of race is the internet?

Sandwiches and songs

Almost everything valuable on the internet is the second kind of race, and to see why, it helps to sort the world's goods into two piles.

Some things get used up when they are shared. A sandwich is the clearest case: if I eat half, there is half as much for you. Land, fuel, a seat on a train — these are rival goods, and competition over them makes sense, because there genuinely is not enough to go around. Most of what the old economy fought over was this kind of thing.

But other things do not get used up when shared. A song does not become quieter because a million people hear it. A map does not wear out because everyone uses it. A piece of software, a fact, a dataset, a connected network — I can use these without leaving any less for you. Economists call them non-rival goods, and they have a strange and wonderful property: sharing them is not a sacrifice, it is the efficient thing to do. To wall off a non-rival good — to lock a song or a map or a dataset behind a fence so only a few can reach it — destroys value for no reason at all, because letting more people use it costs nothing and helps everyone.

Here is the thing the modern internet would rather you not notice: its core goods are almost all non-rival. Data, software, knowledge, the network itself — none of it gets used up. And yet the whole industry is built on fencing these things off and charging for access, one walled garden competing with the next. The economists Charles Jones and Christopher Tonetti studied data³ in particular and reached the conclusion the larger picture implies: because data is non-rival, the arrangement that creates the most value is for it to be widely used, and the real question is not how to ration it but who gets to control it. Hoarding it behind walls leaves enormous value uncreated.

We do not have to guess whether collaboration can produce non-rival goods at world scale, because it already has. Linux — the operating system running most of the internet's servers and, through Android, most of the world's phones — was built by people collaborating, not by a single firm competing. Wikipedia out-produced every commercial encyclopedia that came before it.⁴ These are not charities; they are among the most-used things humans have ever made, and they were made by the bigger circle. And the old fear that anything held in common must collapse into neglect, the "tragedy of the commons"⁵, was answered decades ago by the political economist Elinor Ostrom, who won a Nobel Prize⁶ for documenting, case after real-world case, communities that govern shared resources sustainably and by choice. A commons does not have to fall apart. It has to be designed well.

What we keep, and what we change

So the argument is not that collaboration always beats competition. Said that baldly, it would be false, and you would be right to distrust it. Competition does something real: it tests ideas, rewards the ones that work, and, through prices, gathers up scattered knowledge that no central authority could ever assemble. The thinker who saw this most clearly, Friedrich Hayek, made it the heart⁷ of his case against central planning, and he was right to. It is because he was right that a collaborative internet must be voluntary and decentralized — many communities, real prices, genuine rivalry — and never a single planner deciding from the center.

What changes, then, is not whether we compete. It is two things. First, we choose the right race for each problem — friendly competition where that produces the most, collaboration where that does — rather than forcing everything through competition out of habit. And second, and most important: the rewards are shared evenly. The value the network creates flows back to the people who make it up, instead of pooling in the hands of whoever owns the walls.

Put those together and you get the actual claim of this essay, which is more careful, and stronger, than "collaboration wins":

A collaborative internet produces at least as much real value as a competitive one — because it can run a competition wherever competition is best — and more wherever collaboration is best, and it shares that value far more evenly.

An economist would call that dominance: a choice at least as good as the alternative in every case, and better in some. We are not asking anyone to accept less so that things can be fairer. We are saying the fairer arrangement also makes more, and the only reason it does not exist already is that the people who own the walls have no reason to tear them down.

The longer race

There is a deeper reason to choose all of this, underneath the economics, and it comes down to which race you think you are running.

The companies that own the internet today are built to win the next quarter — the next earnings call, the next funding round, the next product cycle. And a quarter is a sprint: you win it by outrunning whoever happens to be beside you, right now, over a short stretch of track. None of that is wrong on its own terms. But it is the wrong race. The internet is not a quarter-long thing. It is where a civilization will think and talk and trade and remember — not for a few years but for centuries, the public square and the printing press and the marketplace of an entire age at once. The game worth optimizing for is the next thousand years, not the next quarter.

And you do not win a thousand-year race by sprinting against the people next to you. This is not a change of heart about competition — we have already seen that a friendly sprint is sometimes the fastest way to cover a particular stretch, and we keep it exactly where it earns its place. It is the refusal of one specific mistake: treating the whole internet as a short dash to be won by outrunning your neighbor, when it is really a thousand-year endurance run — the kind that is won the way a relay is won, by everyone willing to run taking their turn, handing off, covering for one another, and going far further together than the fastest sprinter ever goes alone.

That is the longer race. It is the one we are actually in, whether we admit it or not, and it is the one a collaborative internet is built to win — not by leaving competition behind, but by putting it in the service of a horizon long enough to be worth the running.

The internet we'd actually run together

Strip away the theory and here is what it means for an ordinary person. On today's internet you generate enormous value — your data, your attention, your contributions, the network of people you bring with you — and almost none of it flows back to you. It collects instead on the far side of a wall, in the accounts of a handful of companies, and the gap between what you make and what you keep is the modern version of the leash. A collaborative internet is simply one that stops pretending the walls have to be there. It keeps the competition that makes things good. It keeps the prices that keep things honest. It adds collaboration wherever collaboration makes more. And then it does the one thing the walls were built to prevent: it lets the value land on the people who created it.

That is not a smaller economy than the one we have. It is a bigger one — at least as much made, and far more evenly held. The only thing it asks us to give up is the habit of running every race alone.

A little over fifteen years ago, an idea arrived with a real claim on the future: that a network could belong to no one and to everyone at once — that you could have the coordination of an institution without an owner sitting at its center, taking a cut and writing the rules. Bitcoin proved the core of it. A ledger that no company kept and no government could quietly edit, maintained by thousands of strangers who did not need to trust one another. That was new, and it was real, and it mattered. The revolution decentralized the record.

But a record is not a reward, and that is where the revolution stopped short. Follow the money rather than the technology, and the pattern is hard to miss: the ledger was decentralized; the spoils were not. Ethereum, the platform that defined the era, created most of its coins before the public could mine a single one — roughly a sixth of the founding supply set aside for the team and its foundation, the largest single share going to one founder.¹ That much was disclosed, and it paid for a decade of genuine building. But it became the template, and the copies grew less principled with each cycle. The wave of initial coin offerings that followed sold tokens to the public while quietly reserving the largest tranches for founders and venture backers;² and at the far end of the same lineage sit the memecoins — tokens with no product and no promise, whose only real function is to move money from the many who buy them to the few who issue them. The blockchain beneath them was a marvel of decentralization. The economics built on top were a machine for concentration.

Honesty requires three corrections here, and each makes the charge sharper rather than softer. The first is that Bitcoin is the exception that proves the rule. It had no premine at all; its anonymous creator wrote a newspaper headline into the very first block as proof that no one had been given a head start, published the software for anyone to run, and never moved the roughly one million coins he mined.³ Bitcoin's inequality is the ordinary kind — early risk, rewarded by time — not the kind that is handed to insiders in advance. The second correction is that the concentration was never confined to the rewards: the machinery itself drifted back toward the center, until a mere two mining pools came to produce around half of all Bitcoin blocks,⁴ and a single service grew to hold nearly a third of all the ether staked to secure Ethereum.⁵ The third is that the wealth figures must be stated with care — by one recent estimate Bitcoin's on-chain inequality sits a touch below that of the United States,³ and such figures are blunt instruments, since one person may hold many addresses and an exchange may hold millions of people's coins in a few. So the accurate charge is not that crypto made everything worse. It is narrower and more damning: crypto decentralized the ledger and left the wealth and the power roughly as concentrated as it found them — and in its most extractive corners, more — all while announcing that it had abolished exactly that.

This is not the complaint of a skeptic shouting from outside the field. It is, in substance, what the field's most serious builder has been saying for years. Vitalik Buterin's own well-known account of decentralization names three axes — how many machines run the system, how many people control them, and whether its logic is unified — and not one of the three asks who ends up holding the wealth.⁶ He has written plainly that one-token-one-vote governance is plutocracy, and that plutocracy is bad.⁷ He has co-authored a paper whose premise is that web3, as actually built, revolves around transferable, financialized assets and carries a long-standing, unsolved problem of wealth concentration that it must find a way past.⁸ When the person who built the second great network in this space tells you it decentralized the tokens but not the power, the diagnosis is no longer in serious dispute. What has been missing is not the insight. What has been missing is a design that does something about it.

The diagnosis matters more now than it did even a few years ago, because a second and far more powerful concentrating force has arrived. Artificial intelligence is, among other things, the most efficient engine ever built for turning the data of the many into value for the few. It learns from the digital shadows of billions of people and returns the gains to whoever owns the data and the models. Set the two technologies side by side and each turns out to carry one great virtue and one characteristic failure. Crypto's virtue is the idea that a network can own itself, with the rewards flowing back to the network rather than upward to a founder; its failure, so far, is that the rewards did not actually stay where they were promised. AI's virtue is that it can create extraordinary value from data at scale; its failure is that it delivers that value to the incumbents who already hold the most of it. The flaw of each is the cure for the other.

That is the wager of this project. We the Users is the deliberate merger of the two: AI's engine for turning data into value, mounted inside crypto's structure for letting a network own itself — built, this time, so that the rewards cannot quietly recentralize. Where the projects before it decentralized the ledger and let the upside pool at the top, this one fixes the upside in advance. Returns to founders, engineers, and investors are capped, so that no one is set permanently above the people who use the network. Each person's data stays with the person who casts it. And if those who run the network ever break the bargain, the Covenant gives the network the right to fork — to leave with itself intact and leave the operators behind. AI is what gives a user-owned network something worth owning; the caps, the fork, and the open Covenant are what keep it owned by its users once it is worth something — which is precisely the test the last fifteen years failed. Buterin has pointed at the same pairing from the other side, noting that crypto's methods are among the few credible ways to build AI that resists the very centralization the technology otherwise invites.⁹

There is one more thing the crypto era got right, and for a project built on collaboration it may matter most of all. Beneath the speculation, what these systems really demonstrated is that trust can be manufactured by amplifying the ability to verify — that trust me can be replaced, at scale, by check for yourself. Nick Szabo gave the idea its name: social scalability, the capacity of a system to take on more, and less familiar, participants without breaking down.¹⁰ Personal trust does not reach far past the people we know; to cooperate beyond that, every society has had to build expensive machinery — courts, auditors, regulators, escrow agents — to vouch for strangers, and that machinery becomes its own center of cost and capture. What a public, verifiable ledger adds is the ability to lock in the integrity of what matters and let anyone confirm it directly, so that confidence in an opaque and changeable intermediary can be exchanged for confidence in a transparent and fixed rule. The newest cryptography sharpens this further still: a zero-knowledge proof can establish that a claim is true while revealing nothing beyond it — that a person is old enough, that a treasury is solvent, that a system followed its own rules — so that verifying need not mean exposing, and recent advances have made such proofs cheap enough to check on an ordinary computer. That is no small thing for what we intend, because cooperation has always been bounded by exactly this: reciprocal cooperation survives only where cheating can be seen, and a shared resource is governed well only where it can be monitored.¹¹ Verification is cheap, continuous monitoring built into the system rather than bolted on by an authority, so a bargain that can be verified can hold together among millions of near-strangers, where one that had to be taken on trust, or policed by a central referee, would have collapsed or been captured long before. A network like ours, then, need not ask its users to believe its central promises on faith — that the rewards flow back, that the returns are capped, that the data stays with the person who casts it. It can make them checkable. Verification is how a betrayal would be seen; the right to fork is what answers it. Don't trust us; verify us, and leave if the verifying ever fails.

Honesty demands the same care here as everywhere else in this argument. Trustless is a marketing word; these systems do not abolish trust so much as shrink it and move it¹² — you no longer trust a bank to keep the ledger, but you still trust the mathematics, the people who audit the code, the assumption that enough participants stay honest, and — hardest of all — the truth of whatever is reported to the system from the outside world, which no proof can vouch for. That last gap, between a faithful record and a true one, is where nearly every spectacular failure of the past decade actually happened. Buterin is plainer about this than the industry's salesmen: no system is ever fully trustless, and the real question is never whether to trust, but whom, how much, and what happens when that trust is betrayed.¹³ The capability is real; what it is aimed at is a choice. Crypto mostly aimed it at speculation. We mean to aim it at collaboration.

None of this is guaranteed by good intentions; the graveyard of decentralization is crowded with them. The hard part was never decentralizing the record — Bitcoin did that in 2009. The hard part is keeping the rewards decentralized after the money arrives, which is the exact moment every earlier attempt quietly gave way. That is why the caps and the forkability and the public Covenant are not ornaments on this project but its load-bearing walls. The revolution that began with the ledger was right about the adversary and wrong about how much ground it had taken. We mean to finish it — on the side of the ledger that actually pays.

Twice now, a cost that quietly governed everything has fallen toward nothing. The first time, we did not understand what was happening until the new owners had already moved in. The second time is happening now, while we are awake — and this essay is about what it makes possible, what it does not, and where it leads if we choose.

The first zero

For nearly all of human history, the expensive thing was sharing. An idea could be had for free, but moving it — copying it, carrying it, putting it in front of another mind — cost a monk's year, a printer's press, a broadcaster's tower. Whoever owned the means of sharing owned much of what got shared.

Then the internet drove the cost of sharing to zero. Any thought, any song, any picture could reach anyone on earth for nothing. We called it freedom, and for the word itself, it was.

But watch what happened to the places. The feed, the index, the marketplace, the inbox — the presses of the new age — were built by a handful of companies, owned by a handful of companies, and everything we said and made and did inside them became their raw material.¹ The first zero freed the sharing of things while quietly cornering the ownership of the places where sharing happens. It cost nothing to speak, and the value of a billion people speaking flowed to the few who owned the rooms.

That was the first zero's lesson, learned late: when a great cost collapses, the question that matters is not what becomes free. It is who owns what gets built on the freedom.

The second zero

Now a second cost is collapsing: the cost of building itself.

Software that once took a company and a year now takes a small team and a season; increasingly it takes one person and a conversation. The machines can write code, test it, refactor it, document it — not perfectly, but relentlessly, and cheaper every quarter. The craft is not disappearing; the scarcity of it is. Building an internet platform is on its way to costing roughly what it costs to describe one clearly.

Be honest about the default, because the default is not on our side. Cheap building reaches the incumbents first. They hold the models, the compute, the data, and the billion-user front door; a tide that lowers the cost of construction lifts the biggest builders most. Left alone, the second zero concentrates — it deals the fastest hand to the players already holding the best cards. Nothing about this technology chooses us.

What the second zero does is narrower, and more important: for the first time since the internet began, it makes the platforms themselves rebuildable. The feed, the search, the market, the messenger — these were never miracles. They are software plus data plus trust. The software is becoming cheap. The data is ours by right — it could not exist without us.¹ The trust is the one thing that cannot be bought, only kept.

The first zero made it free to share what we make. The second makes it possible to own where we make it. Possible — not automatic. Ownership is not what the second zero does; it is what the second zero permits. It has to be chosen, and built, against the grain of what the technology would otherwise do.

The same rooms, a different deed

So picture the destination plainly, because a movement that cannot describe where it is going should not ask anyone to walk.

It looks, at first, unremarkable: the same kinds of rooms. A feed where your friends are. A search box that answers. A marketplace, a mailbox, a map. Built to the same standard of polish, because the tools that build them are the same tools the incumbents use. The difference is not on the screen. It is on the deed.

Each room is run by a community that sets the rules of its own house. The communities share a thin commonwealth — deliberately the weakest tier — holding only what must be common: identity, the Covenant, the Code.² Nothing is run in secret; the software that enforces the rules is public, and if the center ever betrays the people, the people can take the protocol and leave. That architecture is argued elsewhere in these essays. What has been missing, until now, is the plain answer to the plainest question: where does the money go?

Start with scope, honestly. This argument is strongest where the product is mostly information — social, search, media, messaging, coordination — and weakest where it is mostly atoms. Rebuilding the feed is software; rebuilding the warehouse network is not. The commonwealth begins where code and data are the whole game — which happens to be exactly where the shadows are harvested most.

How the money works

A user-owned internet will still earn money. Rooms cost something to run; services will be worth paying for; value will flow through the new network as surely as it flows through the old one. The question every co-op, commune, and republic has had to answer is the one we answer here: who gets what?

The answer has two parts, and the physical world already taught us both.

First, the revenue funds the government. Every functioning society takes part of what it produces and pays for the common work: the roads, the courts, the clerks, the people who keep the water clean. The commonwealth is no different. Out of what the network earns, it funds the people who do its necessary work — the builders who write and maintain the Code, the moderators who keep the rooms livable, the keepers of identity and security and the unglamorous machinery underneath. They are the civil service of the new internet, and they are paid — paid well, and paid openly, on a budget everyone can read — because a commons that expects its hardest work for free will watch that work quietly stop.³ These are wages. They are earned, they differ with the work, and they end when the work ends.

Second, the profit is divided equally — one person, one share. After the common work is funded, what remains belongs to the citizens, all of them, identically. Not weighted by usage, not multiplied for founders, not tiered by talent. The founder's dividend and the newcomer's dividend are the same number.

Understand exactly what is and is not being claimed. This is not the claim that everyone contributes equally — no one believes that, and the wage line above exists precisely because work differs. It is a claim about ownership: the commonwealth's profit is the return on things no one made alone — the pooled shadows of every member, the network that is valuable only because everyone is in it, the trust that belongs to the whole body or to no one. Ownership of the commons follows personhood, not output. Wages are for work. The dividend is for citizenship.

This is not a utopian invention. Alaska has run the shape of it for two generations: oil revenue funds the state's work, and every resident — every one, identically — receives the same yearly dividend from what the commons earns.⁴ No Alaskan pretends every resident contributed equally to the oil. The dividend is not a wage and was never meant to be one. It is what it looks like when a people owns something together and refuses to pretend otherwise.

And notice what the flat share quietly abolishes: the position of owner-above-the-rest. In the old internet, the people who run the platform also own the overwhelming share of it, and that double role — referee and richest player — is the root of nearly everything these essays indict. Here the roles are severed. The people who run the commonwealth draw salaries the citizens can see. Nobody — not the first believer, not the best engineer, not the largest community — draws rent on the public square. Returns on work are earned. Returns on position do not exist.

One person, one share

Strict equality has a price, and honesty requires the sticker.

The flatter and more unconditional the dividend, the richer the prize for a counterfeit citizen. If every member receives an equal share, the most profitable thing a bad actor can do is be many members — and eight billion equal shares are worth faking. Every scheme of common ownership eventually meets this problem; a digital one meets it on day one, at machine speed.

So the entire economy rests on a single gate: membership grounded in real, verifiable personhood. One human being, one membership, one share — verified strongly enough that manufacturing citizens is harder than the dividend is worth. This is a founding requirement of the constitution, not an implementation detail, and it is genuinely hard: proving unique personhood at global scale, without building the very surveillance apparatus this movement exists to end, is one of the open problems of the age. We do not pretend it is solved. We state it as the condition: the commonwealth's economics are only as sound as its answer to who counts as one person — which is why the Covenant must treat personhood the way constitutions treat citizenship: as the first question, answered in the open, amended in the open, and never delegated to a private party with something to sell.

If that gate cannot be built, the flat dividend cannot honestly be paid, and it would be better to say so plainly than to pay a dividend to whoever owns the most machines. We believe it can be built. It is the kind of problem the second zero was made for.

Leaving

No one will be marched into this. There is no expropriation in the plan, no law compelling anyone off the old platforms — the entire moral claim of the movement depends on the exit being chosen. That is also, frankly, the hard part. Network effects are why people stay in rooms they hate: everyone they know is already there, and a better room is worth little empty.

Which is why the way out is not individual but communal. A person alone cannot leave the feed; their people can. A congregation, a guild, a fandom, a town — a community that already knows where its members are can move together and arrive together, and its room is full on day one. This is why the call asks, before anything else, where would you gather? The map of communities willing to move is the map of the exit.

The first zero was used on us before we understood it. The second is arriving while we are awake, and this time the question is on the table before the rooms are built: who will own them?

No one chose the first internet. It was built — and now building is nearly free. The first zero freed the word. The second frees the work. Whether it frees us is not a law of nature. It is a draft, and the pen is on the table.