Core architecturev0.1 · 19 Aug 2026

Ownership & Portability

How a personally owned professional augmentation can learn across a career without carrying an employer’s protected information out the door.

Personal ownershipEmployer boundaryPortabilityAuditability
The problem

If the employer owns the +, the model collapses.

A worker may spend years teaching an augmentation through examples, corrections, judgment and experience. If that accumulated capability simply becomes employer-owned automation, the worker has rational reasons to withhold knowledge and resist the system.

Human Capital .+ reverses the incentive: the worker should be able to build something that remains meaningfully theirs, while respecting every legitimate right of the employer.

Three zones

Personal layer

Generalized professional methods, personal workflows, permitted training history, preference structures and the portable augmentation identity.

Employer workspace

Customer data, proprietary documents, trade secrets, designs, pricing, internal systems and company-specific credentials.

Platform layer

Identity, permissions, audit records, model orchestration, evaluation, portability controls and economic accounting.

While employed

Your + enters the employer workspace.

The + receives only the permissions required for the work. It can use employer information inside that controlled context, but access does not automatically convert that information into portable personal property.

Personal + ↔ permission boundary ↔ employer workspace

All meaningful accesses, changes in permissions and export events should be auditable.

When leaving

The workspace closes. Your + leaves with you.

Employer-specific access is revoked. Protected information stays behind. The worker retains the portable augmentation and any generalized capability that can lawfully and technically be separated from the employer's protected information.

Employer credentials revoked
Protected employer data remains in the employer-controlled environment
Portable personal augmentation remains accessible to the worker
Transfer/export event is recorded and inspectable
Open design issue

What counts as generalized learning?

This is one of the hardest questions in the entire model. A person's ordinary professional learning is naturally portable, but machine-generated representations can blur the line between generalized skill and retained confidential information.

Working rule: never claim portability of a learning artifact until the legal and technical mechanism for separating protected employer information has been validated.

This page therefore defines the intended principle, not a finished legal doctrine.

Keep the record connected.

This is one part of the living Human Capital .+ documentation. Read the other working documents to see how the ownership, economics, identity and validation pieces fit together.