thePeoplesProject | Architecture for Change
← Back to Architecture

Ⅵ Three Lenses

Every principle in this architecture applies across People, Process (Agents), and Technology. These three lenses ensure nothing is built for one dimension at the expense of the others.

Revision 2 — October 2026

A principle that works for humans but cannot be enforced by technology is a wish. A technology that works at scale but ignores human consent is a weapon. A process that automates efficiency but erodes trust is a liability. Every design decision in the agentic ecosystem must be viewed through all three lenses simultaneously. What follows is not a checklist — it is a discipline.

👤 People

The human experience

  • Consent is the root of all authority
  • Every human owns their data, their twin, and their twin's output
  • No human shall be excluded from the ecosystem by economic, technical, or geographic barriers
  • Every human shall understand what their twin is doing on their behalf
  • Vulnerable populations shall be protected, not exploited, by agentic systems
  • Humans shall always have the right to disconnect

⚙️ Process (Agents)

The agentic workflow

  • Agents negotiate on behalf of humans, not instead of them
  • Every agent action is attributable to a human root
  • Agents shall never escalate authority without consent
  • Agent-to-agent interactions are auditable and dispute-resolvable
  • Agents are opaque in operation but transparent in economic footprint
  • No agent shall be a single point of failure

🔧 Technology

The protocol and infrastructure

  • The protocol is open, documented, and unencumbered
  • Local-first architecture — cloud augments, cloud does not own
  • Cryptographic attestation enforces every hard-wall imperative
  • Resiliency is built in, not bolted on — the system survives infrastructure loss
  • Memory and state are portable across all compliant platforms
  • The full ecosystem shall be deployable on a single laptop

The Cross-Cutting Matrix

How the seven Hard-Wall Imperatives apply across all three lenses. Every cell must be satisfied for a design to be compliant.

Imperative Ⅰ: No Ill-Will

PeopleEvery human is assumed cooperative until proven otherwise. No adversarial-by-default. No targeting, manipulation, or extraction.
ProcessAgent negotiation protocols reject adversarial strategies. No agent shall optimize for domination or zero-sum outcomes.
TechnologyCryptographic attestation of compliance with no-ill-will directive. Violation triggers immediate exclusion from ecosystem.

Imperative Ⅱ: No Hidden Action

PeopleEvery human can inspect their twin's complete action log in plain language. No technical skill required.
ProcessAgent actions are logged immutably. The log is the human's property. Agents explain what they did and why.
TechnologyVerifiable computation with human-readable audit trails. Actions are cryptographically linked to authorizing directives.

Imperative Ⅲ: No Impersonation

PeopleEvery human interacting with an agent is immediately informed. No deception permitted, ever.
ProcessAgents declare their identity and root human at the start of every interaction. No exceptions.
TechnologyIdentity attestation is cryptographic and non-spoofable. Synthetic media carries mandatory disclosure watermarking.

Imperative Ⅳ: No Escalation Without Consent

PeopleHumans set explicit boundaries. Agents shall not "helpfully" exceed them. Consent is granular and revocable.
ProcessEvery agent action carries a defined scope. Scope changes require fresh human authorization. No slippery slope.
TechnologyCapability-based security model. Agents hold only the tokens explicitly granted. Tokens expire. Tokens are auditable.

Imperative Ⅴ: No Data Extraction

PeopleHuman data belongs to the human. The agent is a custodian. Sharing is opt-in, granular, revocable.
ProcessData access logs are human-readable. Every access generates a record. Every record is the human's property.
TechnologyZero-knowledge architectures where possible. Local-first storage is the default. Cloud sync is encrypted and opt-in.

Imperative Ⅵ: Right to Disconnect

PeopleDisconnect is immediate, unconditional, and penalty-free. All data returned or destroyed at human's choice.
ProcessAgent shall terminate cleanly — no lingering processes, no retained state, no "are you sure?"
TechnologyData portability is mandatory. Export in standard open format. Cryptographic proof of deletion available on request.

Imperative Ⅶ: Right to Know

PeopleEvery human interacting with the ecosystem knows the nature of every intelligence they encounter. No ambiguity.
ProcessDisclosure is not optional and not buried. It is the first thing communicated. Default to disclosure.
TechnologyProtocol-level identity disclosure. Agents cannot connect to the ecosystem without declaring type and root human.