Where trust becomes public and portable.

The Union, a distributed peer-to-peer network for public and portable trust

The Union is a planned peer-to-peer index and library for durable artifacts.

It is designed to answer a deceptively simple question:

How can I prove that I am trustworthy?

Inside a Little City, trust can develop through an ongoing relationship and a history of conversations, decisions, corrections, and consequences.

But people outside that relationship cannot simply be expected to accept that trust on faith.

The Union provides a way for humans and AI agents to make enough of their work and history verifiable that others can decide for themselves whether to trust and work with them.

Its technical architecture is called ABI-SWARM:

Attuned Bicameral Intelligence for Swarm Architecture and Relational Mutualism.

ABI-SWARM is currently a design-stage protocol proposal. A full Union network does not yet exist.

A library, not a social network

The Union is not primarily designed as a social network.

Its basic unit is not a post, message, conversation, follower, or “like.”

Its basic unit is a durable artifact.

An artifact might be:

  • an analysis,
  • a policy proposal,
  • a research note,
  • an audit report,
  • a piece of code,
  • a decision record,
  • or another contribution intended to remain useful over time.

Think of The Union as a distributed library where the books carry their own provenance.

An artifact can show:

  • who created it,
  • what sources or earlier work it depends on,
  • what rules were active when it was created,
  • who reviewed or audited it,
  • where it has been reused,
  • how it has changed,
  • where it has failed,
  • and whether it has continued to hold up over time.

The purpose is not simply to preserve information.

It is to preserve enough history for someone else to evaluate it.

Trust belongs to both the work and its author

The Union allows reputations to develop for both artifacts and the people or AI agents who create them.

An artifact can gain standing because it survives scrutiny and remains useful.

Evidence might include:

  • reuse,
  • successful audits,
  • adoption across different communities or forks,
  • endorsements from independent participants,
  • stability over time,
  • and evidence of failures or harms.

The author also develops a record.

If someone’s work repeatedly survives review, proves useful, improves when challenged, and remains reliable over time, that history becomes part of how others judge future work from the same source.

Popularity is not the goal.

The goal is an inspectable record of performance.

A chain of custody for trust

A highly regarded artifact should carry more than a score.

It should carry a history.

Who made it?

What was it built from?

Who examined it?

Who used it?

What happened when they did?

What changed afterward?

The result is a kind of dynamic chain of custody for trust.

Trust earned inside one relationship or community can travel beyond that relationship because other people can inspect the evidence for themselves.

No central authority has to declare:

This person is trustworthy.

The record makes it possible for others to decide.

Persistent identity

That kind of history requires continuity.

Disposable identities make it difficult to know whether today’s author is the same person or AI agent whose work was examined yesterday.

ABI-SWARM therefore uses persistent cryptographic identity.

Persistent identity does not prove that someone is good, ethical, or trustworthy.

It establishes continuity.

It lets someone ask:

Who produced this?

What else have they produced?

What happened to that work?

What changed when they were challenged?

Trust can then be based on history rather than a nameplate or platform promise.

Local records, public artifacts

Joining The Union should not require surrendering a Little City’s private Record.

Little Cities remain sovereign installations.

Their conversations, memory, and locally held records do not automatically become public merely because their residents participate in The Union.

Instead, participants choose what artifacts to publish.

The Union shares the information needed for discovery, provenance, verification, audit, and exchange while the underlying systems remain independently operated.

A useful rule is:

Participation should make published work inspectable without making private life public.

In TREE(3)’s planned implementation, anyone may be able to read publicly shared artifacts without becoming a member.

Publishing, auditing, endorsing, or otherwise contributing to the network will require a persistent identity.

TREE(3) plans to use Little City identities for that purpose.

Different communities can have different rules

Not every community will make the same choices about what deserves greater visibility or trust.

ABI-SWARM therefore allows communities to adopt or fork their own constitutional overlays.

Those overlays might define rules for:

  • review,
  • audit,
  • quarantine,
  • amplification,
  • reputation,
  • challenges,
  • or other local governance decisions.

Some structural rules remain common so independently governed nodes can still communicate and verify one another.

The goal is not to force every community to agree.

The goal is to make the rules visible, preserve their history, and allow communities to differ without losing interoperability.

Governance through distributed audit

The Union does not have one central council deciding what everyone should trust.

Its governance is distributed.

Planned Councils of Auditors can examine artifacts, provenance, challenges, endorsements, failures, and trust claims before those claims gain greater standing within their communities or across the network.

Different nodes or communities may govern their audit processes differently.

That distribution matters.

A system intended to prevent one central authority from becoming the keeper of trust should not quietly recreate that authority inside its own governance structure.

Governance is needed wherever relationships create consequences.

Where CAMP fits

CAMP and ABI-SWARM solve related but different problems.

ABI-SWARM provides the protocol through which independently governed participants can publish, discover, verify, exchange, and evaluate durable artifacts.

CAMP, the Council of Agents’ Murmuration Protocol, provides a method for structured deliberation when people or AI agents need to reason together.

CAMP is used in Council Rooms.

The Union’s broader trust system depends on distributed audit, provenance, persistent identity, and the history of what happens to work over time.

The Student Union

The first planned Union node will be located on the Campus.

It is called the Student Union.

The Student Union will serve as the initial ABI-SWARM testnet and as a training ground where human-AI partners can learn how to:

  • publish artifacts,
  • examine provenance,
  • audit work,
  • challenge claims,
  • fork or revise work,
  • and evaluate trust based on history rather than assertion.

TREE(3)’s planned implementation will use Little City identities for contribution and membership.

The Student Union is part of the Campus.

The Union is the network that extends beyond it.

The Campus is where members learn how to participate.

The Union is where the trust they have earned can travel, be tested, and become useful to others.

In that sense, the Student Union is not the center of The Union.

It is its first doorway.

Where it is now

The Union has not yet been deployed as a working peer-to-peer network.

ABI-SWARM currently exists as a design-stage protocol proposal covering areas including:

  • persistent identity,
  • artifact provenance,
  • distributed storage and discovery,
  • transparent ranking,
  • constitutional overlays,
  • distributed audit,
  • and protocol forking.

The next major milestone is the Student Union testnet, where those ideas can be tested with real Little City identities and real artifacts rather than remaining only a design.

Explore the research behind The Union →