The opening of every chapter, and the whole thing read aloud.

To my wife and kids,
my superpower and my everything.
For everyone who has ever sat in a parking lot,
unable to open the car door.
The system was broken. Not you.

I have been working on this book for two years. It began in 2024, as a set of notes scrawled during long walks, late nights, and the kind of conversations that leave you unable to sleep because the ideas will not quiet down. But the truth is, this book has been writing itself inside me for much longer than that.

For as long as I can remember, I have been deeply attuned to people. Not to what they say, necessarily, but to what they mean. To the currents running beneath a conversation. To the flicker of energy when someone talks about work that lights them up, and the almost imperceptible dimming when they describe work that does not. I am, by nature, a deeply sensitive human being. For years I thought this was a vulnerability. It took me a long time to understand it was a superpower.

My secret ability, if I can call it that, is this: from a brief conversation, I can see people. Not their resume, not their title, not the performance they put on in a job interview. The real person. What drives them, what frightens them, where their energy comes from, and what kind of team would let them do the best work of their life.

Two experiences shaped the vision in this book more than any others. The first was Better Place. In the late 2000s, this Israeli company set out to do something the world said was impossible: build an infrastructure for electric cars that would make gasoline obsolete. The company ultimately did not survive, but the experience left a permanent mark. I learned that the quality of the aspiration determines the quality of the people it attracts.

The second was Steve Jobs and Apple. Not as an employee, but as a student of what Jobs understood better than perhaps anyone in business history: that greatness comes from end-to-end integration. He understood that when you control the entire system, when every piece is designed to work with every other piece, you create something that is not just better but categorically different.

This book is a thesis about human nature: that people are dynamic spectrums of capability and potential, not static data points to be managed. It is a blueprint for a new kind of organization: small enough to be human, smart enough to operate at scale, and designed from the ground up for the way people actually work, think, create, and grow. And it is an invitation to build it. Not someday. Now.

The technology exists. The need is urgent. The only missing ingredient is conviction.

Matan Elmalam, March 2026

She sits in her car at 6:47 in the morning, engine off, coffee going cold in the cupholder. The parking lot is nearly empty. The building in front of her is glass and steel, tastefully landscaped, the kind of place that looks impressive in a recruiting brochure. Inside, her desk waits. Her inbox waits. The calendar waits, stacked with meetings she scheduled herself because that is what senior directors do, they fill calendars and call it leadership.

She is not lazy. She is not ungrateful. She has an MBA from a school whose name opens doors, a title that impresses people at dinner parties, and a compensation package that would make her twenty-five-year-old self weep with disbelief. She is, by every metric the professional world has invented, successful.

And she cannot bring herself to open the car door.

Not today. Not because of any particular crisis, no looming deadline, no difficult conversation, no restructuring announcement. Just the cumulative weight of a system that has slowly, methodically, almost politely drained every ounce of meaning from her work.

She used to love this. She remembers loving this. She remembers the version of herself who stayed late not because she had to, but because the problems were interesting and the team was alive and the work felt like it mattered. That version of her feels like a stranger now.

If you have made it this far, something in chapter one landed. Maybe it was the woman in the parking lot. Maybe it was the math: 21%. Maybe it was the line about the system being broken, not you.

I want to interrupt the book for a single page, because I do not want you to read the rest of it the way most business books get read. As an interesting argument. As a thought experiment. As something to nod at and then return to your inbox.

I am not writing this book as a detached observer. I have spent twenty years inside the architecture this book is about to take apart, and I have come to believe, with the kind of certainty that only failure and long observation produce, that the architecture must be replaced. Not improved. Not reformed. Replaced.

What follows is the case for what could replace it. It is a thesis about human nature, a structural argument about intelligence, a blueprint for a new kind of organization, and an honest reckoning with the economics. It is the most rigorous answer I have been able to construct to a question that has consumed me: how should human beings be organized to work together, given everything we now know about who we are and what we can build.

In the spring of 1911, Frederick Winslow Taylor published a slim volume called The Principles of Scientific Management. It was, in its way, a revolutionary document. Taylor argued that work could be studied, measured, and optimized with the same rigor that engineers applied to machines. Every task could be broken into component motions. Every motion could be timed with a stopwatch. Every worker could be trained to perform those motions in the one best way, eliminating waste, variation, and individual judgment.

Taylor was not a monster. He genuinely believed he was improving the lives of workers by replacing arbitrary management with scientific precision. And in the specific context of his time — factories producing physical goods, where the work was repetitive and the primary constraint was coordination of manual labor — his ideas had a certain logic. Production rates soared. Efficiency improved. The industrial economy accelerated.

But buried in Taylor's framework was an assumption so deep it became invisible: that workers were, fundamentally, machines with legs. Their job was to execute, not to think. Their value was in their output, not in their judgment. And the role of management was to design the system, enforce compliance, and extract maximum efficiency from the human components.

Henry Ford embraced this philosophy and scaled it to its logical extreme.

Try a thought experiment. Think of the person you know best in the world, your closest friend, your partner, a sibling you have grown up with. Someone whose quirks you know by heart, whose silences you can read, whose strengths and fears and contradictions you understand in the way that only comes from years of shared experience.

Now describe that person using only their job title, years of experience, and a list of skills. Software Engineer. Eight years of experience. Proficient in Python, JavaScript, and SQL. Experience with agile methodologies and cross-functional collaboration.

How much of who they are did that capture? Not just their work self, their full self. The way they light up when they are solving a problem that genuinely interests them. The patience they show when teaching someone something new. The way they withdraw when they feel undervalued, not because they are being difficult but because recognition is oxygen to them and they are suffocating without it. The creative leaps they make when they are trusted with ambiguity. The quiet leadership they demonstrate in a crisis, when everyone else is spinning and they become still, focused, the calm center that holds the team together.

None of that made it onto the resume. None of it ever could.

You have been measured your entire professional life. Grade point averages in school. Test scores for admission. Performance ratings at work. Engagement survey scores. Competency assessments during training. Skill evaluations during interviews. A lifetime of measurement, each instrument claiming to capture something essential about who you are and what you can do.

And yet, have you ever been truly seen? Not evaluated. Not scored. Not placed on a bell curve or fitted into a competency matrix. Seen. In the full richness of how you think, what moves you, how you show up with other people, what you are becoming.

Most people, when they are honest with themselves, answer no. They have been measured many times. They have been seen rarely, if ever, by the professional systems that govern their careers.

This chapter introduces a different way of looking at human beings. Not through the narrow lens of skills and experience that the resume demands, and not through the blunt instrument of personality types that reduces people to four letters. Through a framework that attempts to capture the full dimensionality of a person, the seven dimensions that together compose what we call the Spectrum.

The Spectrum is not a test you take. It is not a label you receive. The first dimension is Cognitive Style, how a person thinks.

The argument of this chapter is uncomfortable, so let me state it before I make the case. The right size of a working group, the right pattern of connection, the right relationship between autonomy and coordination, are not matters of management philosophy or corporate culture. They are matters of physics and information theory.

And when you take the question seriously, two completely independent disciplines, biology and computer science, converge on the same answer.

They both point at a number uncomfortably close to twenty-five. This is not a metaphor. It is a structural finding.

How I came to it is shorter than the case for it. In late 2024 I stood at a whiteboard for the better part of three days, trying to design an organization. I could not get the structure to feel right. Every version had the same problem. Either the central node had too much control, in which case the design was a holding company with extra steps, or the small circles were too disconnected, in which case it was a collection of independent companies pretending to be a network. I knew the answer was somewhere in the middle, but the middle felt arbitrary. Why this many people per circle? Why this many connections? Why this much autonomy?

The phrasing varies. It might be dressed up as a performance review, a talent calibration, a skills assessment, or a promotion committee. But the underlying question is always the same: does this person meet the standard? Are they performing at the expected level? Do they deserve to stay, to advance, to receive more?

This question feels natural because we have been asking it for so long. But it is the wrong question. It is backwards. It measures the gap between where a person is and where the organization wants them to be, which makes the organization's needs the reference point and the person the variable to be adjusted.

This is not a semantic distinction. It is a fundamental reorientation of the relationship between individuals and the systems they work within. One question treats people as resources to be evaluated. The other treats them as dynamic beings to be understood and developed. One question is about judgment. The other is about trajectory.

And the argument of this chapter is that the organizations that optimize for the second question, the ones that make human growth their primary metric, their organizing principle, their operating system, will systematically outperform the ones that optimize for evaluation. Not because growth-focused organizations are nicer or more idealistic. Because they are smarter.

There is a conversation about artificial intelligence that has consumed the public imagination for the better part of a decade, and it asks the wrong question.

This is understandable. When a technology promises to automate cognitive tasks that previously required human beings, the first instinct is to count the casualties. And the answers have ranged from the reassuring (“only routine tasks will be automated”) to the apocalyptic (“most white-collar jobs will be obsolete within a decade”), generating enormous anxiety and very little clarity.

This chapter asks a different question. Instead of seeing AI as a force that acts on existing organizations, displacing workers, automating processes, optimizing operations, it asks us to see AI as a force that enables entirely new kinds of organizations. Organizations that could not have existed before. Organizations that are better for the people inside them and more effective in the markets they serve.

The key insight is this: large organizations exist primarily because coordination is hard. Not because large is inherently better. Not because 10,000 employees are more creative than twenty-five. Not because hierarchy produces superior decisions.

There is a moment, somewhere around the fifteen-minute mark of a particular kind of conversation, when the person you are speaking with stops performing and starts speaking. You can feel it. Their shoulders drop a fraction of an inch. Their voice loses the slight glaze of preparation. They stop reciting the version of themselves they have been telling at dinner parties for ten years, and start telling the truer version, the one with more specifics, more contradictions, more things that do not fit the resume.

If you have ever conducted hundreds of interviews in a single year, as I have, you know that nearly all of the value of an interview lives on the other side of that moment. The first fifteen minutes are theater. The last forty-five are the actual person.

The problem is that most professional interactions never reach the second half. A typical job interview ends before the candidate stops performing. A typical performance review never gets there because the structure of the conversation prevents it. We have built an entire workplace culture in which the first fifteen minutes are the only fifteen minutes anyone gets, and then we are surprised that we do not actually know the people who work for us.

There is a tendency, when people first encounter the 25 Network, to read it as a workplace innovation. A better way to organize professionals. A nicer way to be at a job. An updated answer to the questions of the modern office.

That reading is too small.

The 25 Network is one specific instance of a much larger phenomenon, and you cannot understand it without understanding the phenomenon. The phenomenon is this. The same AI inflection that makes intelligence-shaped organizations buildable at human scale is also dissolving most of the institutions that twentieth-century life was organized around. Some of those institutions deserve to be dissolved. Others are about to be dissolved whether they deserve to be or not.

The work of the next twenty years is not to slow this process. It cannot be slowed. The work is to design what replaces the institutions that are going. To design intentionally, in service of human flourishing, while there is still room to design.

Enough theory. Enough diagnosis. Enough argument about what is broken and why. Let us build something.

This chapter walks you inside a single 25 org. One node in the network. One team of twenty-five human beings. The structure of a 25 org is what falls out of the architecture from chapter five once you ask it to take the form of a working team. There is no element of it that is preference. Every element is forced.

A 25 org has one chief and twenty-four team members. That is the entirety of the management structure. There is no VP. There is no director. There are no team leads with dotted-line reporting relationships to a matrix. There is a chief who leads, and there are twenty-four people who do the work.

This structure is forced by the sparse connectivity constraint. A neural network layer cannot afford a connection from every node to every other node, and a human team cannot afford a relationship from every person to every other person beyond the cognitive limit that evolution set.

The chief develops people.

A single 25 org is a remarkable thing. A team small enough for every person to be known, autonomous enough for meaningful ownership, intelligent enough to operate without bureaucratic overhead. But a single team, no matter how excellent, is still just a team. It can do one thing brilliantly. The world needs many things done brilliantly, simultaneously, across domains and geographies and timescales.

This chapter is about what happens when many 25 orgs connect. Not into a hierarchy. Not into a conglomerate. Not into a franchise. Into a network shaped like the brain we described in chapter five, where each node is autonomous, each connection carries value, and the whole is dramatically more capable than the sum of its parts.

Start with what flows through the network, because flow is the right way to think about it. The network is not a static structure of boxes and lines. It is a system of moving things. Four kinds of things move through the network.

Information moves. A breakthrough discovered in one node propagates to other nodes that can use it, through the community layer and the cross-team collaboration channels.

You have been on a team that clicked. You know the feeling. The meeting where ideas build on each other so fast that you lose track of who said what. The problem-solving session where someone's offhand comment unlocks a solution that no one would have found alone. The project that ran ahead of schedule not because people worked harder but because the collaboration was so fluid that effort translated into progress with almost no friction.

You have also been on a team that ground. Where every meeting felt like pushing through mud. Where good ideas died in the space between having them and getting them heard. Where the same interpersonal tensions surfaced in every conversation, consuming energy that should have gone to the work. Where people were individually talented but collectively mediocre.

The difference between these two experiences is not luck. It is chemistry. And chemistry is not random. It has a structure, a logic, and, for the first time in history, a technology that can design for it.

This chapter is about that technology, and about the question the technology answers, which is the question that every leader of every team in every era has had to answer in the dark: who should I put together with whom?

Every architecture needs a substrate. The thing underneath that makes the architecture run. A neural network running on the wrong hardware does not run. The architecture is correct; the substrate is incompatible. A brain in a body without circulation is a brain that fails inside three minutes. The architecture is correct; the substrate is starved.

A 25 Network without the technology that lets information, resources, talent, and signals flow across nodes is a 25 Network that does not function. The architecture is correct; the substrate is missing.

This chapter is about the substrate. The technology layer that runs underneath everything we have described in the previous chapters. The system that holds the architecture up.

The substrate is called 25OS. It is organized in three layers, each one corresponding to a different scale of the architecture.

Ideas that cannot survive contact with a spreadsheet are not ideas. They are wishes.

This chapter puts the 25 model through the financial gauntlet. Not because economics is the highest measure of a model's worth, we have spent thirteen chapters arguing that it is not, but because an organizational model that does not produce competitive economic returns cannot sustain itself, no matter how philosophically compelling it may be.

The good news: the economics of the 25 model are not just viable. They are superior. And the superiority does not come from working people harder or paying them less. It comes from eliminating the structural waste that traditional organizations have accepted as inevitable.

Start with the core metric: revenue per full-time employee. This number is the single most revealing indicator of an organization's productive efficiency. The average across all US businesses is approximately $200,000 per employee. A typical large consulting firm generates $300,000 to $400,000. A high-performing technology company generates $500,000 to $1 million. The 25 model targets $500,000 per full-time employee per year. This is ambitious.

He wakes up at 5:15. Not because he wants to. Because the flight to Chicago is at 7:40, and before the flight there are 17 emails that need responses, a presentation that needs one more revision, and a call with Singapore that could not be scheduled at any other time.

He showers, dresses, answers six emails while eating a piece of toast standing up, joins the Singapore call from his car (hands-free, he tells himself, as if that makes it safe), finishes three more emails in the airport lounge, revises the presentation during boarding, and spends the two-hour flight preparing for the meeting that begins forty minutes after landing.

The meeting runs long. It always runs long. The afternoon is three more meetings, back-to-back, no breaks, the calendar a Tetris board of colored blocks with no white space. He gets back to the hotel at 8 PM. Orders room service. Opens his laptop.

You have read the diagnosis. You have understood the human science. You have seen the role AI can play. You have examined the blueprint in detail, how a single team works, how the network connects, how teams are composed, how the platform operates. You have stress-tested the economics. Now the question is structural: what would it actually take to build the first one?

This chapter is not a summary. It is a reckoning with the threshold. The 25 Network, as this book describes it, is a thesis. The thesis says that an architecture shaped like an intelligence, made of people, would outperform the structures the twentieth century left us. To find out whether the thesis is right, the architecture has to be built.

And it begins, the way all networks begin, with a critical mass. The number that critical mass takes is approximately one hundred. Not because the number is magical. Because below it, the marketplace has no real depth, the mobility system has no real range, the community has no real critical mass, and the network cannot do the things that distinguish a network from a collection.

Dear reader, I am writing to you from 2035. Not the 2035 of science fiction, not gleaming towers and flying cars and a world unrecognizable from the one you know. A more grounded 2035. One that looks, in its physical details, much like today. The streets are the same. The coffee shops are the same. The sunset is the same.

But the way people live, the way they work, learn, connect, grow, and care for each other, that has changed profoundly. And I want to tell you about it. Not as a prediction. As an invitation to build it.

Let me tell you about a Tuesday. Noa wakes up at seven. Not to an alarm, she stopped setting one three years ago, when she realized that her body wakes naturally at seven when it is allowed eight hours of rest, and her work structure allows eight hours of rest because the day does not begin with a predawn email triage. She has breakfast with her two children. Not the hurried, phone-in-hand, toast-while-walking kind of breakfast. A real one. Conversation. Laughter. The ordinary magic of an unrushed morning.

This appendix collects the operational specifics that were referenced in the main chapters but deferred from the main narrative. It is organized by topic for reference, not for narrative flow. Readers who want to build, evaluate, or invest in 25 orgs will find it useful. Readers who want only the thesis can skip it without missing anything load-bearing.

1 The Gateway: Five Phases of the Conversation. The Gateway interview unfolds in five sequential phases over forty-five to sixty minutes. Each phase is designed to illuminate different dimensions of the spectrum while creating an experience that feels more like a profound conversation than an assessment.

Phase 1: The Story. The participant tells the story of their working life. Not the resume version. The actual story, including the moments that mattered, the decisions that shaped them, and the turning points that made them who they are.

Phase 2: The Challenge. The participant is presented with two novel, complex scenarios drawn from real-world problems.