OOMNISENS AI docs · v0.1
Introduction

Substrate Engineering

The input-side discipline: strive for deterministic interpretation.

A substrate is a constraint specification S that restricts the interpretation space to IS(D) ⊆ I(D). Deterministic interpretation obtains exactly when |IS(D)| = 1: a single admissible reading remains.

When the space is a singleton and the canonical serialization is injective, independent models emit the byte-identical output, verifiable by SHA-256. Cross-model convergence is the empirical signature that the space collapsed.

example
# |I_S(D)| = 1, C injective
sha256( C(model_A(D)) ) ==
sha256( C(model_B(D)) ) ==   # byte-identical
sha256( C(model_C(D)) )      # = C(i*)