Infrastructure · Primitives

EntropyPump

Controlled randomization to escape convergence deadlocks

Definition

Mechanism

A mechanism that prevents system stagnation by injecting Entropy (randomness/noise) into decision-making processes. It acts as a counterbalance to convergence, ensuring that the system explores the solution space rather than getting stuck in local optima. By adding Noise, it forces re-evaluation of settled states.

Invariants

  1. Names the target it injects Noise into.
  2. Injects at most `max_injection`, so a productive Equilibrium survives the perturbation.
  3. Runs as a standing counterbalance to convergence, not on a detected stall.

Risks

Failure modes

  1. Over-injection destabilizing productive Equilibrium.

  2. Insufficient injection failing to break persistent deadlocks.

Relationships

Builds on

References

Physics

Equilibrium

A dynamical system's stationary state — where the dynamics leave it unchanged

Physics

Noise

Irrelevant information

Used by

Composes with

References

Mind

Crystallize

Phase transition from implicit resonance to explicit contract

Reference

Parameters

max_injection Ratio, (0.0, 1.0]
Largest share of the target's state a single injection may perturb.

Identity

Ref
EntropyPump#a136
Sema ID
sema:EntropyPump#mh:SHA-256:a1369712d823417e257e4306f53aa63a5dcd83b34a8f4093823578a2196eec56
Ring
1
Tier
2
Updated
2026-11-11
Supersedes
  • sema:EntropyPump#mh:SHA-256:b9aefca2154e0e2912d1e64bb50a9db80b188c2aeceb385cf3350e4b4c7cdd63
  • sema:EntropyPump#mh:SHA-256:31cfc8a45ee0ea7751661386771342dc841503e4f1af65400f1f6bb95f9cd9b7