Projects, Planning, Branches, and Long-Horizon Alignment

1. A project is purpose extended through time

A project begins when a desired change exceeds the reach of one cognitive movement. It must preserve an outcome across many artifacts, decisions, sources, assignments, interruptions, and discoveries while allowing the route to change as the work becomes better understood.

AIOS treats the plan as the project's live theory of completion. The plan explains how the present situation can become the intended future condition, which assumptions make that route plausible, what must be produced or decided along the way, and what evidence will establish that the transformation actually occurred.

This gives planning an intellectual function. It does not merely arrange activity. It keeps purpose, strategy, architecture, commitments, and immediate attention mutually intelligible over time.

2. Planning operates at five altitudes

Outcome names the future condition, its beneficiary, its scope, and the evidence by which completion will be recognized.

Strategy states the governing route, its assumptions and risks, and the important alternatives deliberately left aside.

Architecture organizes the artifacts, workstreams, dependencies, decisions, and integration points needed to realize the strategy.

Commitments bind defined work to an owner or role, an authority grant, a checkpoint, an expected return, and conditions for stopping or replanning.

Foreground selects the next meaningful movement and explains why it deserves attention now.

These are planning altitudes: they express how purpose descends from a future condition into present work and how evidence rises to change the plan. They give project form to the Core Thesis's intent altitudes: outcome carries the project horizon; strategy, architecture, and commitments organize structural descent; and foreground carries immediate attention. Context and memory resolution describe the grain of ground composed at any one of these altitudes.

The downward line turns purpose into present work. The upward lines let actual results revise commitments, architecture, strategy, and even the outcome itself. Every immediate movement bears on a larger intention; every long-horizon intention reaches ground through work that can produce evidence.

3. The project has a reconstructable semantic situation

The durable project situation includes its purpose and desired outcome, the plan at each altitude, its field of artifacts and sources, accepted decisions and alternatives, active and parked branches, assignments and returns, dependencies, unresolved tensions, accepted risks, and completion evidence.

The Project Home gives this field a compact entry point. It tells a person or model what the project is trying to become, where work stands, what now governs, and where exact ground lives. The projection remains subordinate to the files, decisions, plans, and records from which it is composed.

This is how a project survives interruption. Continuity comes from reconstructable purpose and state rather than an unbroken conversation or persistent agent memory.

4. Dependencies state what later work needs

A semantic dependency exists when one movement requires the product, decision, evidence, acceptance, or condition established by another. The plan names that requirement directly.

A recommendation may require evidence that distinguishes two mechanisms. A public claim may require source and quotation review. A design may require an accepted definition of its audience. An integration may require destination authority. A workflow stage may require the distinction produced by the stage before it.

Sequence alone cannot express these relationships. “Complete A before B” gives an order. “B requires the evidence account produced by A” gives the project intelligence needed to judge whether the dependency has actually been satisfied.

5. Planning and production occupy different altitudes

Planning works on route and structure. It can establish a workstream, record a decision, change movement order, create or relate artifacts, name a dependency, commission writing, or authorize a workflow or delegated inquiry.

Writing and production realize those commissions in the artifacts themselves. Holistic review then tests whether the completed work still serves the accepted structure and project purpose. The updated ground returns to planning at the proper altitude.

Keeping these movements distinct protects both. The plan does not silently become prose, and the producer does not acquire implicit authority to redesign the project.

6. A branch gives one line of work an independent world

A branch becomes valuable when a question needs a different source field or method, an alternative needs protection from anchoring, an adversarial view deserves independence, the main line can continue without the answer, or a disjoint object can be developed safely in parallel.

The branch record preserves its question, parent purpose, reason for independence, granted and intentionally withheld ground, source scope, status, expected product, receiving object, stop condition, and re-entry condition.

This turns a branch into a bounded contribution. It has enough shared Why to remain relevant and enough local freedom to produce information the parent line could not obtain by merely repeating itself.

7. Parallel work consumes a branch budget

Every additional branch creates an integration obligation. The project weighs the independence gained against shared assumptions, expected disagreement, collision risk, reconciliation effort, latency saved, and the number of live lines the person can still understand.

Parallelism earns its cost when sources are genuinely independent, specialist methods matter, alternatives benefit from separation, adversarial comparison improves the result, mutation surfaces do not overlap, and the parent has a named judgment for comparing returns.

It loses its value when purpose is unstable, every branch awaits the same person decision, several branches reproduce the same work, artifacts collide, or no directing line has capacity to reintegrate the results. The branch budget is a discipline of coherence rather than a fixed numerical quota.

8. Commitments connect work to consequence

A commitment identifies the object of work, its relationship to the parent outcome, its owner or delegated role, the context and tools granted, its mutation ceiling, expected product, latest useful checkpoint, completion evidence, and conditions for stopping or replanning.

The receiving object is essential. A research return might change a named decision. An edit might land in one artifact movement. A prototype might test one strategy assumption. Work without a receiving consequence can consume resources while leaving the project's theory of completion untouched.

Commitments make authorization inspectable. They also let the project distinguish future possibility, accepted work, active execution, returned products, and integrated change.

9. Returns become project knowledge through reintegration

A branch return re-enters at the parent altitude, where project judgment determines what it confirms, qualifies, challenges, or leaves open. An accepted delta can change the question, plan, artifact, or decision. The same judgment can close, redirect, split, park, or continue the branch.

Chapter 17 develops the complete reintegration movement and its disagreement analysis. At project scale, the governing requirement is that returns change their named receiving objects through a judgment capable of reading their evidence, assumptions, limits, and relation to current ground.

Conflicting conclusions are not averaged. Their differences are explained and brought back to the decision that can use them.

10. Alignment is a living relationship

Long-horizon alignment exists when the foreground movement, current commitments, project architecture, strategy, and desired outcome continue to bear intelligibly on one another.

An alignment review asks:

• Does current work still advance the outcome?
• Has evidence weakened a strategic assumption?
• Is an artifact being polished after its role changed?
• Do commitments retain their authority and resources?
• Does a new product change the project architecture?
• Has a parked branch met its re-entry condition?
• Has the project reached an honest stopping point?

The review makes drift and changed conditions visible. Authorized planning judgment then decides what to reaffirm, alter, defer, or close.

11. Replanning is evidence of intelligence

A plan changes when purpose changes, a critical assumption fails, contradictory evidence shifts confidence, a dependency becomes impossible, an artifact reveals a different problem, a branch alters the decision field, resources or authority change, completion criteria prove inadequate, or residual risk exceeds what was accepted.

The prior plan remains in lineage together with the trigger and resulting decision. This preserves the reasoning history of the project and lets later participants distinguish disciplined adaptation from unexplained drift.

Rigid adherence to an obsolete plan is not alignment. The plan fulfills its role by changing when the evidence changes the theory of completion.

12. Completion requires proof at the project scale

Project closure states the intended outcome and criteria, the evidence for each criterion, the artifacts and decisions delivered, unresolved defects or uncertainty, deferred work and its reason, accepted residual risk, required person acceptance, and the conditions that would reopen the project.

Activity provides supporting history. Files created, tasks checked, model calls completed, and time elapsed never substitute for evidence that the future condition was achieved.

Completion also renews the larger situation. The project can produce durable artifacts, decisions, new questions, revised standards, reusable methods, and lessons about its strategy. Reintegration determines which of these products bear on future work.

13. Example: a service redesign program

A public-service team wants people to complete a complex application accurately without repeated support calls. The outcome names the changed experience and the evidence that would demonstrate it. The strategy combines process simplification, clearer language, and assisted review. The architecture separates policy interpretation, form redesign, user testing, technical implementation, and decision records.

One branch studies failure points in current submissions. Another tests alternative language without seeing the preferred design. A third examines the operational burden on staff. Their returns change different receiving objects and are compared before the team commits to an implementation route.

If testing shows that the underlying policy, rather than the interface, creates most failures, the project replans. The purpose persists while strategy and architecture change. That is long-horizon alignment in practice.

Boundary: planning governs movement, not discovery

The plan gives work bearing, authority, and a theory of completion. It remains answerable to evidence produced through the work. Planning becomes a cage when it treats current assumptions as permanent truth; formation becomes drift when it cannot relate discoveries back to the outcome and its governing authority.

14. Failure conditions and evaluation

The principal failure conditions are distinct:

• Task-list planning: activity replaces a theory of how the intended future will be achieved.
• Detached foreground work: attractive immediate work loses its relationship to the outcome.
• Branch proliferation: parallel lines exceed the project's capacity to compare and reintegrate them.
• Duplicated parallel work: several branches reproduce the same assumptions without adding independent information.
• Stale commitments: admitted work survives after its authority, dependencies, or value have changed.
• Undigested returns: completed branches accumulate without changing a receiving object.
• Closure by exhaustion: depleted time or attention is mistaken for evidence of completion.
• Hidden scope or authority change: the project moves under assumptions no accepted decision established.

Evaluation tests traceability from foreground to outcome, quality of dependency accounts, recovery after interruption, speed and fidelity of replanning, branch-budget judgment, reintegration of conflicting returns, reduction of abandoned or duplicate work, and the difference between completion evidence and activity metrics.

A living plan makes purpose executable across time. Its value lies in helping the project remain itself while becoming wiser about how it should proceed.