Compute
Inference, retrieval, training, simulation, and local validation consume budget.
Internal budget
Resource Economy and R(t) Viability
A teleodynamic learner pays for representational structure using an internal viability budget that is replenished by success and depleted by work.
Resource closure
A structure is useful only if the system can afford to maintain it. When the budget drops near the viability floor, expensive structural actions are blocked.
Memory
Storage, indexing, dependency graphs, and trace retention create maintenance burden.
Review
Human review, audit interpretation, policy checks, and citation verification all cost time and attention.
Uncertainty
Ambiguity, fallback pressure, conflicting evidence, and unresolved meaning increase local cost.
Local objective function
The slow loop is myopic and local. It is not a global architecture search. It selects the candidate with the lowest expected local cost only if R(t) can pay the declared action cost.
Decision gatechoose argmin(J(candidate)) only when R(t) ≥ actionCost + viabilityFloor
If no candidate passes the affordability test, the correct decision is no-op.
Forced structural halt is a feature
Blocked action telemetry is evidence that resource closure is working. A system that cannot afford a change should refuse growth rather than silently accumulate complexity.
- Predictive success can replenish budget.
- Decay, compute, memory, review, uncertainty, and maintenance deplete budget.
- Near the viability floor, expensive edits are blocked.
- No-op preserves future viability and keeps audit traces readable.
Deep route polish
R(t) resource-accounting narrative
The visual gauge should be read as a bounded budget: maintenance burden, action cost, and viability floor decide whether growth is allowed.
Written narrative
Resource economy is the site’s guardrail against endless structure growth. A proposed distinction must repay its maintenance burden. If the budget is too low or the value too uncertain, no-op is the disciplined action.
Concrete example
Adding a specialized glyph operator may improve one trace, but if it raises maintenance cost across many routes, the resource economy blocks the addition until stronger evidence appears.
| Focus | What to inspect |
|---|---|
| Action cost | One-time cost of making the structural edit. |
| Maintenance burden | Ongoing cost of keeping the new structure useful. |
| Viability floor | Minimum reserve below which growth is blocked. |
Evidence note
R(t) panels on this site are explanatory simulations. They make review variables visible without claiming a deployed research scheduler.
Internal reading path
/architecture/
Teleodynamic AI Architecture for Self-Maintaining Systems
Teleodynamic AI architecture for resource-bounded learning, semantic glyph interpretation, two-loop control, and self-maintaining AI systems.
/work-constraint-cycle/
Work-Constraint Cycle for Self-Maintaining AI Systems
A practical explanation of the work-constraint cycle behind Teleodynamic AI and adaptive structure under viability pressure.
/claim-status-ledger/
Claim Status Ledger for Teleodynamic AI
A public claim-status ledger showing raw, reviewed, bounded, promoted, restricted, and rejected research claim dispositions with reusable boundary badges.