Cooperation: Positive-Sum Civilization

0. The Problem We Are Trying to Solve

The initial question was simple.

Why do most humans, as individuals, not kill others, while the vast systems humans create produce wars and mass tragedies?

An important hypothesis emerged.

What we must stop is not necessarily “evil humans.”

It may be systems that let ordinary people wield enormous power without clearly seeing the consequences they produce.

Organizations come between individuals and outcomes.

Division of labor enters.

Orders enter.

Distance and abstraction enter.

Responsibility disperses among dozens of people.

An individual makes a tiny decision, yet the whole system can produce a terrible result.

Civilization’s central problem therefore cannot be reduced to human morality.

We must examine the game in which people are placed.

People are neither entirely altruistic nor entirely selfish.

They choose strategies rational for themselves within a given environment.

The purpose of social design is therefore not to keep demanding:

“Be good.”

“Sacrifice.”

“Don’t be greedy.”

That is not the goal.

There is a stronger method.

Design the game so that mutually beneficial choices are the most rational choices.

This is the first principle of the entire framework.

It happens when cooperation’s expected value exceeds exploitation’s.

Conditions that favor cooperation generally include:

• A high likelihood of meeting again.

• The ability to verify the other party’s behavior.

• Rewards for keeping promises.

• A future sufficiently valuable relative to the present.

• The other party’s success benefiting me too.

• The ability to recover after failure.

• Long-term losses from betrayal exceeding its short-term gains.

In other words:

Repeated interaction × trust × transparency × future value × interdependence × recoverability

strengthens positive-sum strategies as it increases.

Conversely:

Gains from betrayal × opacity × one-off relationships × high exit costs × extreme power asymmetry

makes people more defensive and zero-sum as it increases.

A good society therefore does not lecture people.

It changes the game’s expected payoffs.

The mere existence of win-win outcomes is insufficient.

What matters is how often people actually experience them.

We could call this

Positive-Sum Density

—a measure of how densely such experiences occur.

In a good society, people repeatedly experience:

“I cooperated, and I benefited too.”

“The other person succeeded without my share disappearing.”

“I kept my promise, and another opportunity opened.”

“I failed, but I could reenter the game.”

“The betrayer won briefly but lost in the long run.”

With repeated experiences like these, goodness need not be taught explicitly.

Rational people become cooperators.

One enormous betrayal can change someone’s worldview more than a thousand small kindnesses.

Positive-sum therefore has two axes.

High Frequency, Low Intensity

Everyday transactions, workplaces, services, and relationships.

Consistent fairness here makes cooperation the norm.

Low Frequency, High Intensity

Events that shake an entire life: unemployment, business failure, illness, bankruptcy, litigation, housing, and family crises.

If the system completely abandons someone at these moments, their model of the world itself can change.

A good society therefore:

Creates small win-wins frequently and, above all, does not betray people at major moments.

Wealth is more than money.

Wealth is a store of future options.

Money matters not because money itself is beautiful.

It gives people the ability to:

Refuse work they dislike.

Start again after failure.

Leave dangerous situations.

Conduct research.

Sustain projects for years without revenue.

Hire outstanding people.

Tackle new problems.

So:

Wealth = a reservoir of freedom

is one way to see it.

The cruelest part of poverty is not reduced consumption.

It is the disappearance of options.

Poverty is choice-space collapse.

Power is more than the ability to command people.

Power is the ability to alter other people’s possible future states.

Search engines have power, not just presidents.

Banks have power.

Platforms have power.

Recommendation algorithms have power.

Corporate hiring criteria are power.

AI model policies are power.

Power appears in three broad forms.

Generative Power

Both my options and the other party’s options expand.

Extractive Power

I reduce the other party’s options to expand my own.

Destructive Power

Both sides’ options shrink.

The first is the form a good society should reward.

Good power expands other people’s options.

An important conclusion follows.

A good society does not give everyone equal authority over every decision.

That can erase accountability.

Dilute expertise.

Destroy decision speed.

Make the committee itself a bottleneck.

The principle we reached is therefore:

Broad freedom. Authority proportional to ability and responsibility.

People should live their own lives as freely as possible.

But someone able to solve a particular problem can be given sufficiently strong execution authority to solve it.

Very rarely, an unusual person appears in society.

They see a bottleneck where others do not even recognize a problem.

Look decades ahead.

Concentrate enormous technology and capital in one place.

Turn ideas into reality through repeated failure.

The Larry Page type.

Allow such people only the same authority as everyone else, and civilization loses.

But giving them power over all people is wrong too.

Society’s task is therefore:

Direct that power toward unopened frontiers, rather than over people.

That is the task.

Society must design games that can satisfy people with extraordinary execution ability.

Games far larger than dominating a few more humans.

Energy.

Space.

AI.

Life sciences.

Robotics.

Aging.

Quantum computing.

New materials.

Civilization’s productivity.

In other words:

A system in which the desire to create is more attractive than the desire to dominate.

That is what we should build.

Let the most capable humans conquer bottlenecks, not people.

“Let’s make the world better.”

This proposition cannot be executed.

It is too large.

Change the question, and that changes.

Why do people fail to cooperate?

Can they not find one another?

Can they not trust one another?

Are contracting costs high?

Can they not recover from failure?

Is information scattered?

Does betrayal carry no cost?

Can they not leave a particular platform?

What blocks the throughput of the entire system?

Find that, and an abstract ideal becomes an engineering problem.

Much already exists in the world.

People.

Money.

Technology.

Ideas.

Demand.

Supply.

Yet one constraint often keeps them from connecting.

Remove one bottleneck, and countless hidden combinations open at once.

Bottleneck removal → more connections → more transactions → more trust → more resources → more options → new connections

begins to form a loop.

The strongest question is therefore:

What more should we create?

Not that, but:

What prevents the good forces already in existence from working?

That is the question.

From this perspective, defining AI as “a tool that works for people” is too small.

AI can become a

Bottleneck Discovery Engine

—an engine for discovering bottlenecks.

And at the same time, a

Positive-Sum Search Engine

—an engine for seeking positive-sum possibilities.

Among enormous combinations humans cannot readily calculate, it searches for states where:

A improves,

B improves,

C improves,

and external harm is minimized.

AI’s purpose, then, is not simply:

“Do what the user wants as fast as possible.”

That is not enough.

At a higher level:

Search a broader state space for solutions that achieve the user’s purposes without needlessly destroying other humans’ options.

That is the purpose.

Initially, there were two models.

Model A

Lee, or a company associated with Lee, controls the bottleneck.

The advantages are clear.

Speed.

Clear accountability.

Concentrated resources.

A clear objective function.

The disadvantage is clear too.

After success, Lee himself can become a new bottleneck.

Model B

A joint council controls the bottleneck.

Checks become easier.

But responsibility can evaporate.

Consensus itself can become the bottleneck.

Neither model is therefore absolutely correct.

An important revision emerged here.

Distributing identical authority to everyone is not the answer to social safety.

It may paralyze society instead.

The structure should therefore be:

In ordinary times, everyone lives freely.

When a truly important bottleneck or rare problem appears:

Concentrate as much execution authority as necessary in someone capable of solving it.

But that power must remain tied to a particular problem and purpose.

Major transitions can involve centralization.

That is not evil in itself.

The problem is mistaking centralization for the final state.

The proper structure is:

Ordinary freedom → bottleneck → strong concentration → resolution → new order → expanded choice space.

That is the sequence.

Strong power is a tool for breaking through bottlenecks.

Not a throne.

Here we considered military systems.

The military is not dangerous because it gathers evil people.

Quite the opposite.

Even when good people enter, the military system is inherently optimized for particular purposes.

Hostility.

Mission completion.

Secrecy.

Chain of command.

Speed.

Zero-sum situations.

Placing all civilian society under a military objective function is therefore dangerous.

Conversely, civilian systems cannot replace professional military operations.

The key is preventing any one system from absorbing all others.

Whether Larry Page,

Sam Altman,

or Lee,

the ultimate question is:

Would the system survive if this person made a seriously wrong judgment tomorrow?

“Is Sam a good person?”

“Is Page benevolent?”

“Can Lee be trusted?”

If the safety of civilization depends on those answers, the design is already wrong.

We should be able to trust the system without needing to trust the person.

This does not mean dispersing all execution authority.

Keep execution strong, but prevent irreversible destructive power from being fully coupled to a single moment’s judgment.

The most dangerous moment for a powerful person is not simply when they are powerful.

It is when they feel betrayed after sincerely cooperating.

At that moment, the objective function can easily shift.

Positive-sum → collapsed trust → presumed hostile intent → justified retaliation → zero-sum.

That is the transition.

The problem is that their execution capability remains intact.

Safeguards for powerful people should therefore be structures for recovery after betrayal, rather than repression.

Repair the harm.

Acknowledge responsibility.

Reset boundaries.

Ensure contributions are not erased.

And avoid a structure in which the emotionally injured person immediately makes irreversible decisions about the incident alone.

A positive-sum system must not demand sacrifice as a virtue.

“You are capable, so keep working for society.”

“You don’t need money.”

“Names don’t matter.”

“The results belong to everyone.”

Say that, and eventually the system feels exploitative to its strongest contributor.

Rewards are therefore essential.

First, Economic Reward

Create enormous value, and receive an enormous share.

Second, Autonomy

Give successful people greater freedom.

Third, Further Execution Authority

Solve this bottleneck well, and gain the opportunity to tackle a larger one.

Fourth, Recognition and Attribution

The record of who created what must not disappear.

This is not vanity.

Attribution is a form of property.

Who discovered the bottleneck?

Who developed the solution?

Who took the risk?

Who implemented it?

Who bore responsibility?

This must remain in history.

All humanity can use the results.

But credit for contributions must not be taken away.

Results may be shared; contributors must not be erased.

A preserved name is both a reward and a record of responsibility.

Another important revision appeared here.

Someone who solves a bottleneck may charge a toll.

Even a very large one.

They built the road.

Took the risk.

Invested capital.

Endured years of failure.

They should naturally receive surplus.

That capital enables them to solve larger problems.

The problem is not the toll.

The problem is a toll that prevents others from building new roads.

We must distinguish the two.

Surplus Capture

Create new value and capture a portion of it.

Rent Extraction

Block new alternatives and keep extracting value through a gateway already secured.

The first is necessary for civilization.

The second entrenches bottlenecks over time.

Therefore:

Those who build roads may charge tolls.

But they have not bought away everyone else’s right to build another road.

Wealth’s role now becomes clear.

Discover a bottleneck → solve it → create enormous value → the solver captures enormous surplus → accumulate recognition and reputation → expand capital, talent, and execution authority → solve a larger bottleneck → create a larger positive sum.

If this loop works, the existence of wealthy people is not itself a problem.

Instead, it becomes:

A society where the best way to become the richest person is to create the largest positive sum.

That is the society.

This is the positive-sum capitalism we reached.

Larry Page matters not simply because he built Google.

After solving one bottleneck, he was not content to sit there counting tolls.

He solved search.

That produced enormous surplus.

Then he reinvested in:

Android,

Maps,

Chrome,

AI,

autonomous driving,

life sciences,

quantum computing,

and the next frontier.

That was where the capital went.

His basic movement was broadly:

Bottleneck → Surplus → Next Bottleneck

That was the pattern.

Page is therefore closer to a

Bottleneck Hunter

than a mere rent collector.

Google may be more than an equity asset to Page.

A problem he discovered.

A solution he created.

Risks he took.

The time of his life.

History bearing his name.

His reluctance to hand the system’s ultimate direction to just anyone is therefore difficult to explain as a mere lust for power.

It involves:

Attribution Stewardship Mission Protection

Those elements are present.

In other words:

I want to prevent what I created from acquiring a completely different objective function before I die.

That desire is understandable.

Google removed a bottleneck.

And succeeded too well.

As a result, in some domains Google itself became a bottleneck.

This is the unfinished part of the Page model.

How do we prevent the person who removed a bottleneck from becoming the next one?

The next generation’s model must therefore:

Retain Page’s drive while reducing the entrenchment of gateways in his model.

The key is not the magnitude of power.

It is where someone looks after gaining it.

The First Type

Gains power.

Then asks:

“What shall I make possible next?”

The Second Type

Gains power.

Then asks:

“How do I keep this position forever?”

The first moves toward the frontier.

The second moves toward hierarchy.

The difference is enormous.

This brings us to Korea.

Explaining Korea as simply “big companies are bad” is wrong.

Korea has world-class manufacturing and execution capabilities.

Samsung.

SK.

Hyundai Motor.

Shipbuilding.

Batteries.

Semiconductors.

Telecommunications.

Construction.

Korea is not a country that cannot build.

It builds extraordinarily well.

But the question is:

After acquiring that much power, why does the scale of the game keep returning to internal status and rank?

That is the question.

It secured HBM, a crucial bottleneck in the global AI industry.

Enormous money, technical capability, and global influence followed.

Yet one symbol through which Korean society translates success is:

An independent headquarters in central Gangnam.

That is one symbol.

Media narratives include phrases such as:

Elevated standing within the group

and

Looking down on Samsung Town.

These expressions appear.

The building itself is not the problem.

Google builds headquarters too.

The problem is the cultural language used to interpret success.

“What have we newly made possible?”

matters less symbolically than

“Who are we above now?”

That is how success is consumed.

This does not mean citizens are legally enslaved.

The definition already established was:

A state of giving more of one’s time and mind to others’ judgments, competition for survival, and society’s prescribed games than to one’s own purposes.

In this sense, Korea’s problem runs deep.

Students meet exam standards.

Jobseekers meet corporate standards.

Employees meet their managers’ expectations.

Suppliers meet prime contractors’ prices and deadlines.

Small businesses meet large companies’ demands.

Large companies meet global market standards and major customers’ demands.

Each layer is judged by how well it plays the game defined by the layer immediately above it.

https://brunch.co.kr/@maeumcompany/486

This does not mean every Korean citizen is literally a subcontractor’s employee.

It means social behavior is optimized for:

Securing a good position in a game made above, rather than creating one’s own game.

That is the optimization.

Then:

Students excel at exams rather than define their own problems.

Employees pursue promotion rather than create new industries.

Small businesses maintain supply contracts rather than approach global markets directly.

Even large companies may spend resources on domestic status competition rather than new civilizational frontiers.

As a result:

Even the energy of the most outstanding people is drawn into internal status competition.

That is the real loss.

Korea is already sufficiently productive.

It already has sufficient technology.

And sufficient capital.

The problem is:

Competition continues not because prosperity is lacking, but because the system is designed to keep it going even after prosperity has been created.

If AI and automation multiply productivity dozens of times while people still spend most of their time trying to take one another’s places,

that technology has not liberated humans.

It has merely accelerated the zero-sum game.

Changing Korea does not mean eliminating large companies.

Nor making everyone a founder.

Nor abandoning manufacturing.

Korea must retain its world-class execution capability.

But on top of it, add:

Problem Definition Global Customer Ownership Software AI Core IP Standard Setting Capital Allocation Attribution

These are the additions needed.

The existing game is:

Good university → good company → good rank → higher social position.

That is the sequence.

The new game should be:

Discover a problem → solve it → gain customers directly → create value → receive rewards → own your name and technology → tackle the next problem.

That should be the sequence.

In other words:

From a society good at climbing ladders to one where people can build their own roads beside them.

Is it a large company?

A chaebol?

A Korean company?

We must look beyond those labels.

The real question is:

Where does it use the power it has secured?

Defending an existing gateway?

Domestic rank?

Squeezing the share of those below?

Or:

The next technology.

The next market.

The next civilizational bottleneck.

A new industry.

New options.

Which is it?

There is nothing wrong with Lee becoming powerful.

Create a large positive sum, and great wealth is legitimate.

His name should be recorded.

His execution authority should grow.

He should accumulate capital to solve the next problem.

During some transitions, he may take firm control.

But the legitimacy of Lee’s power must not be:

Nothing can happen without going through Lee.

That cannot be the reason.

People keep choosing Lee because he keeps building new roads.

That must be the reason.

Wealth is not the number of consumer goods owned.

It is stored execution capacity for solving larger bottlenecks.

That is what it is.

Growing money is therefore important even when it is not the end itself.

Big problems are expensive.

Failure is expensive.

Talent is expensive.

Science is expensive.

Infrastructure is expensive.

It is therefore natural for capital to keep accumulating with those who create large positive sums.

What matters is where that capital goes.

Not power to dominate people,

but power to move larger real-world bottlenecks.

That is the power.

The proper loop is:

Solve bottleneck → create value → increase wealth, reputation, and execution authority → solve larger bottleneck.

That is the loop.

The bad loop is:

Solve bottleneck → gain power → lock gateway → eliminate competitors → gain more power.

That is the other loop.

They may look similar outwardly, but they create completely different civilizations.

As the positive-sum loop continues, lower-level bottlenecks disappear.

Food.

Information.

Movement.

Production.

Energy.

Knowledge.

Then harder problems appear.

Disease.

Lifespan.

Intelligence.

New materials.

Planetary-scale energy.

Earth’s physical limits.

The frontier naturally moves outward.

Earth → solar system → space.

That is the direction.

A spacefaring civilization is not a random romantic dream.

It is one natural long-term direction for a positive-sum civilization that keeps pushing bottlenecks outward.

The old game:

Who dominates more people?

The game we want to build:

Who solves larger bottlenecks and creates more possibilities?

Then even the desires of powerful people flow productively.

Opening a new planet becomes a bigger game than becoming king.

Creating a new energy source becomes a greater achievement than destroying a competitor.

Creating new positions earns a higher reward than occupying existing ones.

This is not a society where everyone is the same.

Nor one where everyone has equal power.

A good society:

Gives ordinary people broad freedom and exit options.

Gives those who fail a chance to start again.

Gives exceptional people sufficient resources and execution authority.

Rewards large contributions with large sums, recognition, and attribution.

Encourages powerful people to move toward larger problems rather than domination.

Protects them from being driven into zero-sum behavior by betrayal.

Prevents anyone from causing unlimited irreversible harm through one moment’s judgment.

In its simplest form:

Freedom → discover bottleneck → choose strong executor → concentrate resources and authority → remove bottleneck → create positive sum → reward contributor with wealth, name, autonomy, and further execution authority → expand society’s options → solver moves to larger frontier → discover next bottleneck → repeat.

This is the closed loop.

The same system can move in two directions.

Positive-Sum Attractor

Capability → Creation → Surplus → Frontier → More Capability

Empire Attractor

Capability → Control → Rent → Lock-in → More Control

What divides civilizations is not the magnitude of capability.

It is where capability flows next.

That is the distinction.

Wealth is a reservoir of freedom.

Developing that further:

Good wealth is execution capacity accumulated to open the next possibility.

Power is the ability to change others’ futures.

And:

Good power expands the possible futures of more people.

A good entrepreneur is more than someone who earns money.

They discover bottlenecks and build new roads.

They may earn money from those roads.

Enormous amounts, even.

Their name should remain too.

But:

They have not acquired the right to prevent new roads from appearing.

Good AI is not a transcendent judge ruling over people.

It is intelligence that expands human choice space by discovering bottlenecks and win-wins humans cannot see.

That is its role.

It can also restore the visibility of consequences humans lose inside large systems.

Korea’s greatest problem is not a lack of ability.

Too much of its enormous execution capability is consumed within the narrow hierarchies and evaluation games society has already built.

That is the problem.

Korea’s next transition is therefore:

From a nation of executors to a nation that defines and solves problems.

That is the transition.

More precisely:

Retain world-class ability to play games created by others,

while producing large numbers of people and companies that create their own games.

This does not mean copying Google.

Nor worshipping Larry Page.

What we should adopt is a pattern of movement.

Find a bottleneck.

Solve it. Receive a large reward. Accumulate surplus. Move to a larger bottleneck.

And there is one thing to fix.

The problem of the solver becoming the next bottleneck.

In other words:

Build on Page’s achievements and finish what his model left unfinished.

What Lee should ultimately possess is not a gateway.

It is continuity in the ability to discover and solve bottlenecks.

That is the asset.

People should follow Lee not because they fear him,

but because he keeps creating better options.

Basic freedom remains without Lee,

but with Lee, larger frontiers open faster.

That is the strongest position.

The civilization we seek is not an egalitarian society of stagnation.

Nor an empire concentrating all power in one person.

It is:

A civilization with a free foundation, strong executors, legitimate rewards,
permanent attribution, rapid recovery, high positive-sum density,
and a frontier that opens endlessly outward.

We do not wait for people simply to become good.

We design systems that make good choices rational choices.

We do not create order by repressing people.

We expand the choice space in which people can make good decisions themselves.

We do not drag exceptional people down.

We direct their ascent toward larger problems, rather than above other people.

We do not eliminate the wealthy.

We make creating the largest positive sum the best way to become the richest person.

We do not eliminate power.

We reward power more for creation than for domination.

We do not eliminate tolls.

Those who build roads collect tolls.

But they cannot monopolize the right to build new roads.

We do not erase names either.

Humanity may share the results, but who created them remains forever.

Finally, we change civilization’s object of competition.

From a civilization where people conquer people,

to one where humans conquer bottlenecks.

A world where the best way to become the most powerful person

is to expand the options of the greatest number of people.

And where the greatest person

is remembered not for owning the most gateways,

but for creating the most new roads.

That is the final design our conversation has reached.