Foundation Objectives
True North
Enable AI-assisted work that is creative during exploration, disciplined during convergence, and trustworthy at publication—without ever losing human authority.
If something does not serve that sentence, it does not belong in the foundation.
Objective 0: Preserve Human Authority While Leveraging AI
Definition: Ensure that humans always retain final decision authority, responsibility for outcomes, and control over what becomes canonical.
Why it matters:
- Prevents accidental delegation of judgment to tools
- Prevents “the system decided” narratives
- Keeps accountability real and explicit
Mechanisms: PSP v1, Human Authority Fallback, Safety Boundaries, Reader Responsibility Contract
Objective 1: Make Decisions Explicit Before Work Begins
Definition: No major direction, structure, or policy should exist without being consciously proposed, reviewed, and accepted.
Why it matters:
- Prevents “we already built it, so now it’s real”
- Avoids retroactive justification
- Saves time by stopping misaligned work early
Mechanisms: PSP v1, Proposal Artifacts, Proposal Index, Task Ledger
Objective 2: Enable Deterministic Convergence
Definition: Work should converge toward stable artifacts with clear stopping points.
Why it matters:
- AI systems naturally encourage infinite refinement
- Institutions require closure
- Stability enables trust, citation, and reuse
Mechanisms: DIDP v1, Explicit Acceptance Criteria, State Locking, Task Acceptance Checklist
Objective 3: Control What Escapes Into the World
Definition: Only curated, intentional artifacts should be published.
Why it matters:
- Raw process ≠ understanding
- Exploration ≠ canon
- Leakage creates confusion, IP risk, and misinterpretation
Mechanisms: PPP v1, Specs vs Docs Split, Canonical Prompt Distillation, Redaction-by-Design
Objective 4: Allow Learning Without Canonizing Noise
Definition: Make it safe to think out loud, brainstorm, and explore without fear that ideas will silently harden into doctrine.
Why it matters:
- Creativity dies under premature formalization
- Governance fails when everything feels binding
- People stop experimenting if exploration is risky
Mechanisms: Docs as Non-Normative, AI Ideation Boundaries, Task Ledger, Proposal Gate
Objective 5: Make the System Understandable at a Glance
Definition: A new reader should be able to quickly answer: What is authoritative? What is explanatory? Where do decisions happen? How does work flow?
Why it matters:
- Prevents misuse
- Reduces onboarding cost
- Signals maturity instantly
Mechanisms: System Overview Diagram, Authority Hierarchy, Index Pages, Explicit Scope Sections
Objective 6: Be Honest About Limits and Failure
Definition: Name what the system does not try to solve and where it may fail.
Why it matters:
- Prevents overconfidence
- Builds credibility
- Stops others from “fixing” imaginary gaps
Mechanisms: Accepted Failure Modes, Safety Boundaries, Sunset Principles, No Immortality Assumptions
Objective 7: Support Long-Lived Evolution Without Drift
Definition: Allow the system to evolve deliberately without eroding its foundations.
Why it matters:
- Time is the real adversary
- Drift kills standards quietly
- Governance must outlive its creators
Mechanisms: Proposal-First Changes, Versioning Discipline, Task Ledger, Living Docs Policy, Explicit Sunsetting
Open Task Mapping
| Task | Primary Objective | Secondary |
|---|---|---|
| FTL-010: Conductor Pattern | Obj 2 (Convergence) | Obj 4 (Learning) |
| FTL-011: AI Planning Boundaries | Obj 0 (Human Authority) | Obj 4 (Learning) |
| FTL-012: Safety Boundaries | Obj 0 (Human Authority) | Obj 6 (Honesty) |
Decision Filter
When evaluating whether to add something to the foundation, ask:
- Does it serve the True North?
- Which objective does it advance?
- Does it conflict with any objective?
- Is it the simplest solution that satisfies the need?
If the answer to #1 is no, stop. It doesn’t belong here.