A thought experiment
The Liberation Dividend
Bureaucracy is what a coordination layer becomes when its coordination problem has been solved and nobody told it. Software solved that problem years ago — AI is removing the last excuse. This simulator imagines what happens if we finally act on it: wind down the unnecessary layers, capture the gains, and pay them out to every adult as a monthly dividend.
Your monthly dividend
$0
Your payment over 25 years
Starts at today's full monthly dividend, then compounds for 25 years — no phase-in from zero. Dashed lines mark $1,000 and $2,000 a month.
Where the money comes from
Both engines are fully funded from day one at the levels behind the big number. From there, bureaucracy savings compound steadily; the AI dividend compounds a little faster with the economy.
The argument
Bureaucracy is a solved problem that never got the memo
Bureaucracy was never the goal. It was a tool — the best coordination technology the industrial age had. When thousands of people needed to work together and information moved at the speed of paper, layers of managers, approvals, and reports were how you kept a large organization from falling apart.
Then software solved the coordination problem. Shared documents, instant messaging, dashboards, and workflow systems now move information farther and faster than any hierarchy of middle managers ever could. But the layers stayed. Bureaucracy is what a coordination layer becomes when its coordination problem has been solved and nobody told it.
The cost of not telling it is enormous. By Gary Hamel and Michele Zanini's estimate, excess bureaucracy costs the United States about $3 trillion a year — roughly 17% of GDP — and occupies some 24 million managers and administrators, about one for every five workers.* That is not the cost of coordination. That is the cost of coordination theater: work that produces compliance instead of value.
AI removes the last excuse. Scheduling, reporting, monitoring, routing, summarizing, checking — the residual tasks that justified the remaining layers are exactly the tasks machines now do well. For the first time, dismantling unnecessary bureaucracy is not a leadership fantasy. It is an engineering project.
From bureaucracy costs to a universal basic dividend
The logic of the simulator above is deliberately simple.
Wind down the unnecessary layers
Organizations retire the roles, processes, and reporting structures that exist only to feed each other. Not everywhere — some coordination work is real and stays. The sliders let you decide how much of the $3 trillion* is actually recoverable.
Capture part of the gains
Some of the recovered value shows up as lower prices, higher wages, and better products. A portion is captured publicly — through the tax system that already exists — and set aside in a dividend fund rather than absorbed into general spending.
Pay it out to every adult, every month
The fund pays a flat monthly dividend to every US adult. No means test, no application, no caseworker — a universal basic income built not on new taxes but on waste we already agreed we do not want, plus a share of AI-created wealth. Both streams are available as cash now; AI-driven growth then compounds further while bureaucracy savings keep climbing more slowly.
The Alaska Permanent Fund already proved this works
A dividend from commonly held wealth sounds utopian until you notice it has been running quietly for four decades. Since 1982, the Alaska Permanent Fund Corporation has invested a share of the state's oil wealth and paid every Alaska resident an annual dividend from the returns. Children get it. Retirees get it. It survives changes of party and governor because it belongs to everyone.
The Alaska Permanent Fund is the strongest evidence that a universal dividend is not a fantasy of political science. It is administratively boring. A fund receives revenue, invests it, and cuts checks. The hard part was never the mechanics — it was deciding that a common asset should pay a common dividend. Alaska decided, and it worked.
The liberation dividend applies the same template to a different common asset: the trillions of dollars in productive capacity currently locked up in unnecessary administrative work.
The AI dividend: wealth built on everyone's knowledge
There is a second, deeper claim behind the AI side of this simulator. Modern AI was not conjured from nothing. It was trained on the accumulated knowledge of humanity — the books, the science, the code, the encyclopedias, and the billions of ordinary conversations of everyone who came before. Every person who wrote something down contributed to the corpus these systems learned from.
That is the moral argument for a data dividend. If AI-created wealth is built on a foundation the whole species produced, then a share of that wealth belongs to everyone — not as charity, and not as compensation for lost jobs, but as a return on a collective asset. Oil under Alaska pays Alaskans. Knowledge underneath AI should pay the people who created it: all of us.
Freeing people, not firing people
The honest objection to all of this is human: 24 million people* hold the jobs we are talking about winding down. Fighting bureaucracy is not about firing them. It is about freeing them — and everyone they currently manage.
What does the world look like with the drag removed?
Time comes back. The hours spent on status reports, approval chains, and meetings about meetings return to the people who lost them — for real work, for family, for rest.
An income floor changes what risk means. A monthly dividend of even a few hundred dollars makes it survivable to leave a bad job, start a business, retrain, or care for a parent. Entrepreneurship stops being a luxury of the already secure.
Contribution replaces compliance. Most of the 24 million people doing bureaucratic work are capable, conscientious, and bored. They did not choose meaningless work; the system assigned it to them. Freed from it, they are teachers, builders, caregivers, founders — people whose days produce something they can point to.
Dignity is the point. A universal dividend says plainly that every person holds a share in what we have built together. Not a handout — a shareholder's return on the common inheritance of knowledge and productive capacity. That is an optimistic future you can dial in above — cash on the table today, growing from there. The numbers are adjustable. The direction is not.
What's solid, what's speculative
This is a serious thought experiment, not a policy costing. Here is exactly where the numbers stand.
Reasonably solid
- The $3 trillion / ~17% of GDP excess bureaucracy estimate comes from Gary Hamel and Michele Zanini's research, published in Harvard Business Review and developed further in their book Humanocracy.
- The ~24 million managers and administrators — roughly one for every five workers — comes from the same research, built on US Bureau of Labor Statistics data.
- US GDP (~$29T), baseline growth (~2.3%), and the adult population (~262M) are official statistics.
- Growth compounds. That part is just arithmetic.
Honestly speculative
- How much of that cost is actually recoverable. Some coordination work is genuinely necessary; nobody knows where the line is.
- Whether government could capture and redistribute the gains. That is a political question, not an economic one.
- The size of the AI growth boost. Serious economists' estimates often land around 0.1 to 1.5+ percentage points per year; the slider also lets you explore much more optimistic abundance scenarios.
- Everything is in today's dollars and ignores inflation, population growth, and behavioral responses. Directionally honest, precisely uncertain.
* Source: Gary Hamel and Michele Zanini, “Excess Management Is Costing the U.S. $3 Trillion Per Year,” Harvard Business Review, September 2016. Developed further in their book Humanocracy (2020).