There is a large family of physical and mathematical principles that map unusually well onto AI systems, ACP-style governance, and institutions. What’s striking is that the best analogies are not about intelligence or computation at all, but about fields, constraints, accumulation, and path-dependence—the same things ACP keeps surfacing.

Below is a structured set of principles, each with a brief explanation of what it illuminates institutionally. We are not claiming these are perfect models; they are thinking tools that degrade more gracefully than “AI as brain” or “AI as employee.”

1. Gravity & Mass Concentration

(San Francisco, Washington DC, Wall Street)

Physics idea
Mass warps space; objects fall toward concentrations of mass even without intent.

Institutional mapping

  • Capital, talent, attention, and legitimacy accumulate in hubs.
  • Actors “fall” toward these hubs not because of ideology, but because movement is easier there.
  • Once mass concentrates, it becomes self-reinforcing.

ACP insight

Governance failures often come from ignoring gravity—pretending actors have equal freedom of motion when the field is already warped.
AI systems plugged into these hubs amplify gravity, not neutrality.

2. Electromagnetic Fields (Invisible Influence)

Physics idea
Fields shape motion without contact; charged particles respond even when forces aren’t visible.

Institutional mapping

  • Incentives, norms, prestige, and fear operate as fields.
  • Individuals believe they are “choosing,” but trajectories are shaped upstream.
  • Power is exerted without explicit commands.

ACP insight

Most AI risk is field-level, not action-level.
Focusing on outputs misses the invisible structures shaping behavior before decisions occur.

3. Potential Wells & Local Minima

Physics / optimization idea
Systems settle into low-energy states that are stable but not optimal.

Institutional mapping

  • Organizations get stuck in “good enough” processes.
  • AI systems reinforce existing workflows because deviation is costly.
  • Escape requires energy injection (political capital, crisis, leadership).

ACP insight

“Why don’t they just change?” misunderstands energy landscapes.
ACP’s emphasis on pre-authorization and constraints is about reshaping the landscape, not pushing harder uphill.

4. Vector Fields (Direction Matters More Than Speed)

Physics idea
Motion is defined by direction and magnitude; knowing only speed is insufficient.

Institutional mapping

  • Acceleration (faster decisions, faster models) is less important than direction.
  • Institutions often optimize speed while drifting strategically.
  • AI increases velocity without correcting vector alignment.

ACP insight

Most failures come from misaligned vectors, not insufficient capability.
ACP is about direction-setting, not acceleration.

5. Phase Transitions

Physics idea
Gradual change produces sudden qualitative shifts (ice to water, water to steam).

Institutional mapping

  • Trust collapses suddenly after long erosion.
  • Automation feels benign—until it isn’t.
  • Norms persist until a threshold is crossed.

ACP insight

Risk is nonlinear.
Governance that relies on incremental monitoring fails to detect approaching phase changes.

6. Conservation Laws (Nothing Disappears)

Physics idea
Energy and momentum are conserved; they only change form.

Institutional mapping

  • Responsibility displaced by AI doesn’t vanish—it reappears as blame, litigation, or moral injury.
  • Time saved becomes pressure for more output.
  • Authority abdicated re-emerges as informal power.

ACP insight

“Offloading” is usually transformation, not elimination.
ACP insists on tracking where things go when they’re displaced.

7. Entropy & Information Decay

Physics idea
Systems tend toward disorder without active maintenance.

Institutional mapping

  • Documentation rots.
  • Context evaporates.
  • AI-generated summaries accelerate entropy if not governed.

ACP insight

Order is a continuous cost, not a default state.
Ungoverned AI increases entropy faster than institutions can counteract.

8. Hysteresis (Path Dependence)

Physics idea
The current state depends on history, not just present conditions.

Institutional mapping

  • Past crises shape present behavior long after conditions change.
  • AI systems trained on historical data encode old constraints.
  • Reversibility is an illusion.

ACP insight

You cannot “reset” governance by policy alone.
Design must account for memory and irreversibility.

9. Knot Theory (Irreducible Entanglement)

Mathematics idea
Some knots cannot be untied without cutting; local adjustments fail.

Institutional mapping

  • Legal, technical, and social systems are intertwined.
  • Fixing one layer tightens another.
  • Simple reforms increase complexity elsewhere.

ACP insight

This explains why point solutions fail.
ACP treats systems as entangled, not modular.

10. Boundary Conditions (More Important Than Equations)

Physics idea
Solutions depend more on boundary conditions than governing equations.

Institutional mapping

  • The same AI behaves differently in courts, hospitals, or classrooms.
  • Interface design and permissions matter more than model architecture.
  • Context defines risk.

ACP insight

This is one of ACP’s core claims:
Governance lives at the boundaries, not in the core model.

11. Resonance & Feedback Loops

Physics idea
Small periodic forces can cause catastrophic oscillations at resonant frequencies.

Institutional mapping

  • Repeated nudges (KPIs, metrics, prompts) destabilize systems.
  • AI feedback loops amplify behaviors unintentionally.
  • Harm arises from rhythm, not intent.

ACP insight

Danger often lies in repetition, not magnitude.
ACP watches for resonance, not just spikes.

12. Fractals & Scale Invariance

Mathematics idea
Patterns repeat across scales.

Institutional mapping

  • The same governance failures appear in teams, firms, governments, and AI systems.
  • Micro-level incentives reproduce macro-level outcomes.
  • Fixes must be primitive-level, not scale-specific.

ACP insight

This explains why ACP applies equally to AI, offices, parenting, and states.
It is addressing primitives, not domains.

Why These Analogies Matter

Most AI discourse uses metaphors that collapse under scale: brains, tools, assistants, or oracles.
Physics-based and systems-based metaphors survive longer because they:

  • tolerate uncertainty,
  • foreground constraints,
  • respect nonlinearity,
  • and assume interaction effects.

ACP’s strength is not that it has the right metaphor, but that it refuses to settle on one—and instead asks which metaphors degrade least under failure.