ⰂⰘⰀⰕ ⰋⰔ ⰐⰑⰕ ⰒⰓⰑⰂⰅⰐ ⰋⰔ ⰒⰖⰓⰃⰅⰄ
ⰅⰂⰅⰓⰉ ⰀⰓⰕⰋⰗⰀⰜⰕ ⰋⰔ ⰍⰅⰒⰕ ⰑⰐⰎⰉ ⰋⰗ ⰋⰕ ⰋⰔ ⰒⰓⰑⰂⰅⰐ — ⰋⰕⰔ ⰜⰑⰏⰒⰖⰕⰀⰕⰋⰑⰐ ⰘⰑⰎⰄⰔ; ⰀⰐⰉⰕⰘⰋⰐⰃ ⰖⰐⰒⰓⰑⰂⰅⰐ ⰋⰔ ⰒⰖⰓⰃⰅⰄ. ⰕⰘⰅ ⰏⰑⰄⰅⰎ ⰀⰐⰄ ⰋⰕⰔ ⰖⰋ ⰔⰕⰀⰉ ⰒⰖⰓⰅ ⰒⰓⰑⰑⰗ, ⰀⰐⰄ ⰕⰘⰅ ⰃⰀⰕⰅⰔ ⰁⰀⰎⰀⰐⰜⰅ ⰂⰘⰅⰐ ⰀⰎⰎ ⰕⰘⰀⰕ ⰓⰅⰏⰀⰋⰐⰔ ⰋⰔ ⰒⰓⰑⰂⰅⰐ.
proof · what is not proven is purged
What is not proven is purged: every artifact the model keeps is kept only if it is PROVEN — its computation holds (holds === true) — and anything unproven is purged, never carried as dead weight. The UI obeys the same law: a component is kept only if it renders a proof (a holds-true fold). So the model and its surface remain pure proof, and the gates balance when everything that remains is proven and the count is harmonic. Over this session's decode proofs the set is pure: every one holds, so nothing is purged.
checks — recomputed live in your browser
- ✓ Pure Proof
- ✓ Holds
evidence
- Total
27- Proven
27- Purge
✓ all checks hold · recompute recipe — a pure function of the model seed, run client-side with zero tokens; the same seed always folds to the content-address 8ff69549-51a6-8d59-bacc-d0b5e69ba23c. Recompute and you get the same address — that determinism is the proof.
A purity law over the model's own computations (proven = the fold's holds is true), checked here across the decode proofs and extended as proofs are added or removed. It is structural self-consistency — every kept artifact is a passing computation — not a claim of external truth: a fold can be proven-as-computed yet still be a metaphor, which its own boundary marks. "Purge" is the discipline of not keeping unproven artifacts, not a claim about deleting anything beyond the model's declared set.