# How large groups do things

As a group grows, we have a harder time [personally connecting](people-4_friends.md) with everyone in that group. Starting around 20 people, other members can just as often be acquaintances as close friends. We hit a hard limit on human interaction at about 150 people, both psychologically and [logistically](logistics.md).

A [social network](networks-social.md) user, a citizen of a [city](politics-city.md) or [country](politics-country.md), an employee of a corporation, and a member of a large club are all functionally in a large group. Broadly speaking, modern society consists *mostly* of large groups, segmented into [smaller groups](groups-small.md) according to who and what's most [convenient](purpose.md) or [familiar](symbols.md).

We tend to purpose ourselves toward larger groups for several reasons:

1. Larger groups give us the [impression](image.md) of [security](safety.md) against things we [fear](mind-feelings-fear.md).
2. Each person is capable of [*heavily* specializing](jobs-specialization.md) in things they want and giving others what *they* specialize in.
3. Everyone amasses more [shared knowledge](technology.md), meaning cutting down on [work](results.md).

In larger groups, we tend to associate with a few people in that group and maintain many acquaintances. Some people, like leaders, have many more connections, but there are too many people in a large group to stay in touch with all of them.

## Image

Large groups feel like a massive, monolithic entity, but they're made of the combined [souls](humanity.md) in that group, vaguely oriented toward the leaders' [communicated](people-conversation.md) [purpose](purpose.md). Unlike [smaller groups](groups-small.md), though, that purpose can be *very* far-removed from the leaders' guidance.

As groups become progressively larger, top-ranking people will become dramatically more [powerful](power.md) than low-ranking ones to create a [class divide](classes.md). To prevent too much [in-fighting](people-conflicts-war.md), the people at the top will naturally convey an [image](image.md) that dilutes how much power they legitimately have:

- Leaders will have others introduce them before they speak.
- [Knowledge](understanding.md) experts will use [language](language.md) to imply they're [uncertain](understanding-certainty.md).
- Multimillionaires will wear working-class clothing or avoid showing their wealth.
- [Religious leaders](religion.md) will feign humility (e.g., "I'm the worst of sinners").

To clearly understand *any* group, an observer needs to see 3 perspectives at once:

1. Legal entity: It's a separate existence which makes its own collective [decisions](decisions.md), though its [civil rights](people-boundaries.md) are generally treated as sub-human.
2. Holistic entity: It's made of people, most of them normal workers doing various [specialized roles](jobs-specialization.md), and some of those are only partly of that entity (e.g., contractors) and others with more power (e.g., shareholders, leadership).
3. [Power](power.md): It advances [an agenda](purpose.md), defined and curated by its leadership, which impacts its individual members and other people around it.

Unlike leaders in times past, their effort is to [identify](identity.md) with the rest of the people instead of demonstrating raw [force](power.md). It's [influence](power-influence.md) via [being relatable](people-4_friends.md) instead of [through fear](mind-feelings-fear.md).

Leaders must reflect [values](values.md) that align with the [purpose](purpose.md) of the organization. If they ever break from it, the organization will change in different directions. At its worst, it becomes an [intergroup conflict](people-conflicts-war.md) or a [bad system](mgmt-badsystems.md). Leaders have so many observers that they *must* publicly [identify](identity.md) with those values, or they'll sow [dissent](mgmt-badsystems.md) among the members.

More than anything, the group's represented value will bind all the members together through a shared belief in [emotionally](mind-feelings.md) similar [consequences](results.md). [Fear](mind-feelings-fear.md) of adverse [effects](results.md) is the greatest motivator to get everyone together, which is rarely subtle.

## Scaling

Every large group was once a [small group](groups-small.md), but kept growing. Any large group that takes decades to grow will have its own distinct [culture](culture.md) that members [pressure](power-influence.md) new members to adopt. Members in rapidly grown groups tend to maintain their old cultures as they come in, meaning the group adopts a strange hybrid of everyone's background and is easily dispersed by comparison.

More [cultures](culture.md) can resonate with an organization that promotes very fewer [values](values.md), so an organization grows proportionally to how few values it conveys, assuming it has [mass exposure](marketing.md). At the same time, many values allow a clearer group [identity](identity.md). At the same time, stated values that aren't shown can cause the entire culture to [backfire](mgmt-badsystems.md) and [break apart](people-conflicts-war.md).

As a group grows, each individual's [influence](power-influence.md) in the group becomes more [specialized](jobs-specialization.md). Members' tasks become more specific, and many leaders are simply a "connecting-point" to other smaller sections of the group. Once this starts happening, most members will become more [apathetic](mind-feelings.md) to the group's [causes](purpose.md) and the group will feel more inhuman, irrespective of how [relatable](power-influence.md) or [noble](morality.md) its origins were.

Over time, members' [specializations](jobs-specialization.md) in various aspects will create special-purpose authorities beyond the central leadership roles (e.g., bureaus, coalitions). [Bureaucracy](bureaucracy.md) is when those bureaus have most of the [power](power.md), and they are [subgroups](groups-small.md) of their own inside the meta-group that are subject to [intergroup conflicts and misunderstandings](people-conflicts-war.md).

Since every person has a different reaction to [new information](education.md), [unexpected](imagination.md) [consequences](results.md) will come from any large [decision](decisions.md) proportionally to the size of the group. The only way to plan ahead is to consult [history](stories-storytellers.md). Otherwise, the group should make [plans](imagination.md) for a trial-and-error endeavor, along with the resources to [hedge the inevitable risks](safety.md) from those decisions.

As a group expands, leaders don't see [results](results.md) as immediately as they had in a [small group](groups-small.md). For that reason, their inherent aptitude creates the ceiling of the group's size, and it's harder for them to learn how to lead through simple trial-and-error.

All the [unpredictable](imagination.md) elements of a small group compound across many people in a large group. There are more [power](power.md) dynamics, more [purposes](purpose.md) that diverge from the group's, and usually a broader mix of [cultures](culture.md). Each leader's ability to [manage a small group](groups-small.md) will be challenged in entirely new directions, especially regarding human universal tendencies to be [lazy](results.md), [evil](morality-evil.md), and [engage in conflicts](people-conflicts.md).

When the scale becomes enormous (i.e., surpassing thousands), every large group leader must act quickly to keep the entire organization in line. Their role frequently becomes a never-ending string of resolving [disasters](people-conflicts.md).

Large groups are far more risk-averse than [small groups](groups-small.md), since they have so much more [power](power.md) to lose and so much more easily. They're also susceptible to [trust](trust.md) *other* large groups more, since they often depend so heavily on them for their large-scale resource needs.

To gather more [understanding](understanding.md) or [certainty](understanding-certainty.md), most large systems give [formalized tests](education.md). While these tests *can* clarify, they represent tremendous [power](power.md) to the members, who will immediately find ways to cheat on the test. Thus, while leaders often [trust](trust.md) the tests, they're one of the quickest ways to make a [dysfunctional organization](mgmt-badsystems.md) if everyone knows what they're used for.

## Hierarchy

In any formal hierarchy or organization structure, [power](power.md) is abstracted and redistributed proportional to its rigidity. This makes large groups more able to perform gargantuan tasks (which are often easy to [measure](math.md)), even while being completely [non-human](humanity.md) in how they do it. The highest form of this abstraction comes through [money](economics.md).

Unlike a [smaller group](groups-small.md) with an organization run by the [personality](personality.md) and [decisions](decisions.md) of a few leaders, larger groups consist of a *very* formalized hierarchy of leadership, often at least three layers deep (e.g., boss, his boss, and the CEO).

Top leaders are usually at least 2-3 degrees removed from any actual work done, but they're responsible for making all the decisions. If trends or technology change, they must stay aware of it, or the entire group will become [irrelevant](mgmt-badsystems.md). This is why most top leaders spend most of their time [communicating](people-conversation.md) and [consuming information](creations.md), and why they're not very [useful](purpose.md) in any practical sense.

Often, top leaders will let lower-ranking leaders of smaller subgroups have unique roles. This ranges by how much [trust](trust.md) the leaders give, and can range from the top leader merely existing as a [symbolic](symbols.md) "publishing" or "producing" role all the way to completely micromanaging the subgroup. Absolutely nobody *ever* likes being micromanaged, and it's the product of a [dysfunctional system](mgmt-badsystems.md) caused by bad leadership not respecting individual [decision](decisions.md).

Middle leaders (e.g., middle management) are effectively an extension of their top leadership to manage the bottom leadership. Middle leadership often makes very few [decisions](decisions.md) compared to the top and bottom leadership. However, their [willingness](purpose.md) to comply with the organization's [policies](rules.md) keeps the entire organization unified in [purpose](purpose.md) (or, at least, the [appearance](image.md) of it). In practice, most middle management is simply the above-stated "connecting-point" to other subgroups.

Leadership getting feedback about the members, and acting on it, is *critical* for the entire group's success, and the entire group can become [dysfunctional](mgmt-badsystems.md) surprisingly easily. Leadership failures usually come from incompetence more than malice, though their [public image](image.md) may imply something else.

## Decisions

Top leaders only have limited [influence](power-influence.md) over the entire group's [culture](culture.md). They only maintain it by making specific [decisions](decisions.md) that craft an [image](image.md) for everyone to interpret for themselves. That image promotes collective [values](values.md) (at least in [appearance](image.md)) that everyone holds to:

1. Messages to everyone in the group about the shared [purpose](purpose.md) they're all aspiring toward.
2. Appointing, promoting, disciplining, and dismissing sub-leaders who conform with the [culture](culture.md) they want to promote.
3. Encounters with the sub-leadership and *very* limited encounters with the lower-ranking members to rally everyone toward that stated [purpose](purpose.md) and sort out details.
4. Formally declared, often written, [policies](rules.md) and [procedures](habits.md), enforcing them, or changing them.

These decisions are *always* designed for the interests of the people in the group or the people the group favors. For that reason, a group won't generally affect others it doesn't favor or assimilate already unless it's facing [competition](people-conflicts-war.md) with another group that may serve those other people or [subduing](slavery.md) those people for another [purpose](purpose.md).

Members' awareness of each large-scale decision ripples outward like a [trend](trends.md) at the speed of [technology](technology.md). It has a formative time with sub-leaders, then transitions outward to the majority, with the laggards *never* aware of the decision. This trend sticks proportional to the [leadership skills](mgmt-1_why.md) of the top leader.

Low-ranking members have *much* less freedom than if they were in a [smaller group](groups-small.md), since the leadership redirects most of the lower ranks' [power](power.md) to other [purposes](purpose.md). Low-rank members won't usually [trust](trust.md) their large group much, since they're not given much freedom, but they [pretend to](image.md) well enough that most leaders don't realize this.

## Subcultures

The sub-leaders each create their own [subculture](culture.md) inside that group's broader culture. They'll appoint their own sub-leaders and dictate their own [rules](rules.md).

The differences between individuals *in* the large group are often greater than the proposed differences *between* the groups. However, because of [ego issues](image.md) coming from the top leadership, those groups will never converge until there's a [greater threat](people-conflicts-war.md) from the outside. This will almost certainly happen if they become [obsolete](trends.md) from other, different groups coming in.

One of the most prominent [conflicts](conflicts-inner.md) for all sub-leaders is whether to enforce a [rule](rules.md) that became obsolete from the situation changing. Often, this becomes a battle between [morality](morality.md) and [maintaining image](image.md).

Each subgroup ends up behaving precisely like its own small group, with a few notable exceptions:

1. There's a more complex distribution of [power](power.md) and redirection of resources toward [purposes](purpose.md) that many of the members don't know than a smaller group.
2. Everyone in a subgroup has many more resources available than an independent group would have.
3. The large-scale leader's [decisions](decisions.md) have a far-reaching, soft effect. Everyone [feels](mind-feelings.md) like they're part of the organization without as much evidence as a small group, even when they're in a well-connected subgroup.

Each subgroup usually has different [purposes](purpose.md) to accomplish, and they [redefine](identity.md) their [culture](culture.md) around those purposes. This can create dramatic differences among subgroups in the same large group.

Even in a complete autocracy, the presence of others' [decisions](decisions.md) in [carrying out](results.md) the dictator's authority dilutes the [power](power.md) dynamic across multiple people. This makes *all* systems subject to [tampering and misuse](mgmt-badsystems.md).

In open-ended leadership (i.e., the sub-leaders run everything) the sub-leaders have the final authority to mismanage as they [please](purpose.md). Thus, a central leader who can't or won't punish [evil](morality-evil.md) will empower it and the group will partly [devolve](mgmt-badsystems.md).

These subcultures affect each other to create the aggregate group culture, with each one being its own "unit" interacting with other "units". As various subgroups gain or lose [power](power.md), the organization travels through [eras](trends.md) of control. Those eras create the shared [culture](culture.md) and [identity](identity.md) of the group.

The [traditions](habits.md) that the subcultures make pile together over time. If those traditions keep accumulating but aren't removed as the members who found [meaning](meaning.md) through those traditions leave that group, the group will slowly become a [bad system](mgmt-badsystems.md). Unfortunately, the changes must be gradual, or the entire group will have a [complete meltdown](people-conflicts-war.md).

## Social Engineering

Instead of *stopping* people from doing what they want with [rules](rules.md), social engineering involves using rules to make people *[want](purpose.md)* to do something ("libertarian paternalism").

Every large organization will try to socially engineer, at least a little. Even if they don't have the [power](power.md) to do it, the draw from how much [power](power.md) they can gain is too tempting to pass up.

Social engineering is risky, proportionally to how much it's done. [Human nature](humanity-universals.md) hates to be "boxed in", but leaders' human nature won't let them easily reverse a bad [decision](decisions.md), so it often becomes a [bad system](mgmt-badsystems.md) in the process of trying to [craft an ideal](politics-perfectsociety.md).

Every [imposed rule](rules-methods.md) can hurt people and damage [purposes](purpose.md), and social engineering is trying to use the bludgeoning [force](power.md) of the law to coax, so it usually leads to [negative externalities](economics.md) (e.g., overlooking people who don't *precisely* quality for beneficial things, or shouldn't qualify for beneficial things but still do).

Taxing, especially, will broadly destroy [purposes](purpose.md):

- Taxing *anything* far enough will motivate people to subvert the [rules](rules.md) entirely.
- Taxing what people produce makes people want to [create](creations.md) less.
- Taxing what people consume makes them less likely to consume.
- Taxing [taboo](morality-taboo.md) things too much without completely [outlawing](rules.md) it often makes the organization depend on that thing for income.
- Taking enough from the [rich](classes.md) makes removes the [incentive](purpose.md) for the middle class to [take risks](socialrisk.md) to become rich, and is never enough in a [large system](groups-large.md) to balance the group's budget alone.
- Taking enough from the middle class often makes them become the lower class.
- Taking enough from the lower class is almost completely [evil](morality-evil.md), since it barely benefits the leadership and hurts those with the least [power](power.md).
- In fact, the only [fair](morality-justice.md) tax is a progressive tax up to a specific level of survival, then everything else from middle-class onward as a flat and non-negotiable percentage.

When a [bad system](mgmt-badsystems.md) starts socially engineering, they're often operating off [theories](imagination.md) mixed with trial-and-error. This can be utterly disastrous, and often leads to [social unrest](people-conflicts-war.md), people losing property, and people dying.

## Results

Every far-reaching decision in a large group requires a lead time between [communicating](people-conversation.md) it and everyone [doing it](results.md). That can range from minutes to weeks, depending on the [technology](technology.md) available for communicating and acting. If the group has any [dysfunction](mgmt-badsystems.md), or it's a particularly [unexpected](imagination.md) task by most of the group, it'll take *much* longer or may result in an [inner conflict](people-conflicts-war.md) where the task is never performed.

Every single [result](results.md) of an organization is [created](creations.md) by the members' [decisions](decisions.md) and actions. Thus, even when the leaders get *all* the [credit](image.md), or it had a debatable [morality](morality.md), every member was partly responsible for it, no matter how small.

Like individual [human purposes](purpose.md), large groups are always concerned with growing when they're not facing a [challenge](people-conflicts.md) to their existing [power](power.md). They tend to grow their power both outward and upward:

- Outward growth comes through absorbing other groups and people, often with [large-scale conflicts](people-conflicts-war.md). If uncontested by other groups, it becomes an [empire or monopoly](politics-monopolies.md).
- Upward growth comes through increasing the organization's [knowledge](understanding.md) and [technologies](technology.md). This often comes through [formalized training](education.md) and expanding individuals' roles into broader group responsibilities.

The best [analysis](logic.md) in the world can't calculate all the variables of a large group, since it's impossible to [grasp](understanding.md) everything that happens or [might happen](imagination.md) at any slice of time. Thus, the group will always create [statistical outliers](math.md) that don't match the group's [purpose](purpose.md), regardless of *what* [decision](decisions.md) they make.

While it's impossible to track how a group will respond, there are a few conditions to more easily track changes:

1. Make frequent, small [decisions](decisions.md) and poll the results over time. This permits the group to adapt to [trends](trends.md) before another trend hits.
2. Maintain an endless [image management](image-distortion.md) that stays up-to-date with every change.
3. Expect that history *never* repeats itself, but has a similar pattern. It's never the same because the [culture](culture.md) is different between both cases of deciding, sometimes simply because people *remember* the prior decision.
4. Always give approximate estimates, and avoid trusting something simply because it was [mathematically](math.md) communicated.

*Many* [intergroup conflicts](people-conflicts-war.md) can adapt and [change groups](trends.md). This change can be surprisingly fast or only change its [external image](image.md), depending on how the leadership handles it. Since there's so much power in a group, they tend to persist indefinitely in some form or another unless they're destroyed or merged into another group.

However, with very few exceptions, the members who engage in most intergroup conflicts float through a unique sequence:

1. They sincerely [believe](understanding-certainty.md) they can dramatically [resolve some issues](https://adequate.life/fix/) within the group.
2. They'll make bold actions (e.g., [communicating](people-conversation.md), [building](creations.md)) to transform that group for the better.
3. To survive in the group, they'll eventually need to compete with other members of the group, so they'll compromise their message.
4. If they don't become one of the leadership or give up, the group's leadership will eventually eject them as a [pariah/heretic](morality-taboo.md) or will outright kill them.

## Competition

The leadership of most large groups treat the entire group as a singular entity, then compare themselves to other groups. When other groups outperform that group (e.g., [technology](technology.md), size, [achievements](purpose.md)), [human nature](humanity.md) works at scale, and they try to outperform them.

Some of them will perform noble achievements in response to the challenge, but most of them become [bad systems](mgmt-badsystems.md) through a few avenues:

- Banding together and allying with other groups that do *not* share the same values, leaving the entire group as a morass of [mixed ideals](values.md).
- Stealing from the other group and trying to imitate their success, but without the strength the first group had developed from having to [work at it](success-5_persevering.md).
- [Attacking](hardship-war.md) the other group and either destroying them or taking their things.

However, if a large group *doesn't* have a purpose outside itself via competition, it'll start getting restless and destroy things.

In short, large groups must always be [acting](results.md) toward a fixed [purpose](purpose.md) they [believe](understanding-certainty.md) they can perform without outside help, or they will descend into [chaos](unknown.md) and destroy or disband.