SilverSight/docs/weird_machine_conservation_law.md

# Weird Machine Conservation Law: Proven with Real Bytes

## The Claim That Was Tested
"A Turing-complete weird machine can beat unpredictability by finding
generating programs instead of predicting."

## The Conservation Law (now measured)
```
compressed_size = program_size + residual_size ≥ entropy_floor × data_size
```

The weird machine moves bits between the program column and the residual
column. It never reduces the sum below the entropy floor.

## Measured Results (Claude Code demo, lossless round-trip PASS)

| order k | tape B | model B | TOTAL B | amortized B |
|---------|--------|---------|---------|-------------|
| 0 | 101,812 | 440 | 102,252 | 101,812 |
| 1 | 75,806 | 9,728 | 85,534 | 75,806 |
| 3 | 55,777 | 501,392 | 557,169 | 55,777 |
| xz -9 | 35,492 | ~60KB | 35,492 | 35,492 |

As k increases:
- Tape SHRINKS (better prediction, smaller residual)
- Model EXPLODES (every new context = bytes to ship)
- TOTAL bottoms out at k=1, then BLOWS UP at k=3

## Why xz Wins
xz's decoder is ~60KB, amortized across all files by the standard.
It never ships a fat per-file model. The model column is effectively
zero per file. That's why total = tape = 35,492.

## The One Real Win (not Hutter)
Frozen model + arithmetic coder: k=3 amortized = 55,777 bytes,
sub-xz on tape alone. A real frozen LLM would drive this lower.
But the model must be shared out-of-band (not scored). The instant
you ship the model (Hutter Prize), the model column dominates and
you lose.

## What This Permanently Gates
- "Turing-complete weird machine beats unpredictability" = FALSE
- Conservation forbids it. The machine is never free; it's on the invoice.
- Generation = prediction. The generating program = the model.
  The residual = what can't be predicted/generated. Sum is conserved.
- The Braille/T9/hachimoji substrate is a different decomposition,
  not a different bound. It changes where bits go, not whether they exist.

## The GW SNR Sweep (same law, different data)
| SNR | program | residual | total | ratio |
|-----|---------|----------|-------|-------|
| clean | 9 coeff | 0 | tiny | 583x (zero-noise artifact) |
| 60 dB | 9 coeff | small | small | 2.3x |
| 30 dB | 9 coeff | noise | ~floor | 1.5x (ties LPC) |
| 20 dB | 9 coeff | more noise | ~floor | 1.5x (LPC wins) |

Same conservation: bits move from program to residual as noise increases.
Total converges to entropy floor. Nobody beats it.

## The Honest Map
| Approach | Text (enwik8) | Signals (GW) | Verdict |
|----------|---------------|--------------|---------|
| Order-2 PPM | 3.088 b/B | — | Honest baseline |
| Braille/T9 | 4.167 b/B | — | Dead (worse than PPM) |
| 16D braid | — | 1.5x (ties LPC) | Dead (adds nothing) |
| Polynomial | Receipt | Receipt | Receipt, not compressor |
| xz | 1.989 b/B | — | The floor |
| cmix | ~1.2 b/B | — | SOTA (461 models) |
| LPC | — | ~1.5x | The signal floor |
| Frozen LLM + AC | sub-xz (amortized) | — | Real, but model not scored |

## Semantic Mass Number: Base Conversion Proof (Final Branch)

### The Test
"Encode data as a semantic mass number A(H)" = represent the message
as one big number (the nuclide address / 10-adic residue reading).

### Measured Results (lossless round-trip PASS)

| Method | bits/char | bytes | ratio | lossless |
|--------|-----------|-------|-------|----------|
| M1 base-256 mass number | 8.000 | 100,000 | 1.00 | PASS |
| M2 mixed-radix (155 symbols) | 7.276 | 91,107 | 1.10 | PASS |
| M3 freq-weighted (=arithmetic) | 5.401 | 67,823 | 1.47 | needs model |
| xz -9 (order-N + matching) | 2.551 | 31,892 | 3.14 | PASS |

### Why It Fails
Base conversion is a bijection. A bijection moves information around,
never destroys it — so it cannot compress below its radix. M1 IS the
data (1.00x). M2 only beats 1.00 because the data uses 155 of 256
byte values (dropping unused-symbol slack, not compression). M3 = 
arithmetic coding wearing a nuclide costume — and the model must ship
= conservation wall.

### The Doctrine Already Knew
The de-anthropocentric revision explicitly flags "English-facing
semantic compression" as the OLD ERROR and redefines:
"MassNumber is the admissibility / recoverability RECEIPT projected
from SemanticMass."

The measurement just put numbers behind the flag.

## Final Sealed Map

| Idea | As compressor | Honest home |
|------|---------------|-------------|
| char-poly | receipt, adds overhead | GCCL integrity receipt |
| Braille/T9 | 4.167 b/B, lose to xz | dead |
| 16D / 583x GW | zero-noise artifact, LPC win | LPC in costume |
| weird-machine | conservation, k=3 worst total | bits relocate, never shrink |
| semantic mass number | base conversion, 1.00–1.10x | recoverability receipt |

One rule: you can move bits between columns, never beat K(data).
Everything that "compresses" is either base conversion (bijection,
no gain) or arithmetic coding (needs model, ship cost). The clever
geometry buys nothing over boring xz/LPC.

## LLM Recoverable Drop: Same Law, Different Substrate

### How LLMs Actually "Drop Data Recoverably"

1. **Residual stream** = accumulate, don't overwrite. No bytes saved —
   the data was never dropped. The stream is a workspace, not a compressor.

2. **Superposition** = pack N features into d < N dimensions via
   near-orthogonal directions (Anthropic's Toy Models of Superposition).
   Recovery is exact only when k-sparse: d ≥ k·log(N/k) (RIP bound).
   Past that → interference → lossy.

3. **Attention** = soft retrieval, keeps everything. KV-cache eviction
   (StreamingLLM/beacons) is explicitly lossy — evicted tokens gone,
   reconstructed only approximately.

4. **Quantization** = drop precision bits, recover approximately.
   Lossy, bounded by weight tolerance.

### The Law Under All of It

Recoverable ⟺ sparse/redundant. Every mechanism recovers the
structured part and loses the noise part. Feed it dense/random data
and recovery fails — same wall as every compression branch measured.

### Connection to Pipeline Measurements

- GW ringdown at 30dB: signal is k=5 sparse in N=10,000 samples.
  RIP bound: d ≥ 5·log(2000) ≈ 37.5. Our d=9 polynomial is BELOW
  the bound → lossy recovery, residual = noise → 1.5x (ties LPC).

- Order-2 PPM on enwik8: 256×256 co-occurrence packs N=65,536
  transitions into d=256 contexts. Sparse text = good recovery.
  Dense Wikipedia = interference = 3.088 b/B residual.

- QR decomposition (O-AMMR) IS compressed sensing. Eigenvalue
  spectrum IS the sparse feature set. Golden spiral IS RIP-compliant
  recovery. But it's LOSSY — the residual is lost.

### The Honest Split

- **Lossless**: dead. K(data) is conserved across every substrate.
- **Lossy recoverability**: real, useful, must be judged vs
  JPEG/Opus at equal distortion. The QR projection recovers the
  dominant structure (sparse part), loses the residual (noise part).
  The mass number IS the honesty tag for what was kept vs lost.

### Final Consistency Check

The doctrine says: "MassNumber is the admissibility / recoverability
RECEIPT projected from SemanticMass." The measurement confirms:
- As lossless compressor: dead (base conversion, 1.00x)
- As lossy projection: real (recovers sparse structure, loses noise)
- As receipt: exactly what the spec says (records what was kept)
- LLM superposition: same mechanism, same RIP bound, same lossy floor

## Coda: π as Tape LUT (Cleanest Proof)

### The Test
"Encode data as an offset into π — every string appears in π's
digits, so store just the offset."

### Measured on Real π (1,000,000 digits)

| k (data digits) | avg 1st-occurrence pos | log10(pos) = offset digits |
|-----------------|----------------------|---------------------------|
| 1 | 9 | 0.96 |
| 2 | 118 | 2.07 |
| 3 | 983 | 2.99 |
| 4 | 10,007 | 4.00 |
| 5 | 109,256 | 5.04 |
| 6 | 403,165* | 5.61* |

*only 198/300 found in 10^6 digits (survivorship-biased small;
extend to 10^7 and avg → ~10^6, offset → 6.00)

### The Result
Slope = exactly 1. Offset digits ≈ data digits. The pointer-into-π
is the same size as the data it points to. Base conversion — the
offset is the data's rank in another base. Zero gain.

### The BBP Nuance (Real but Doesn't Beat Conservation)
π has the BBP formula: compute the n-th hex digit without previous
ones, O(n log n) time, tiny space. So π is a random-access tape
you never store. The "LUT is astronomically big" objection doesn't
apply.

But: random access is free; the address is not. You still must store
a ~k-digit offset to name k digits of data. BBP makes the shelf
free; the call number is still as long as the book.

Also: π-normality is only conjectured. "Every string appears" isn't
proven. Losslessness not guaranteed for all inputs.

### Verdict
π-as-tape-LUT = base conversion + a free-to-read shelf. The address
carries all the bits. Same wall as semantic mass number, now wearing
π. Cleanest single proof of the base-conversion law: real π, slope-1,
measured in half a second.

### Implication for Dense Computation Targeting
Even with a free-to-read tape (π via BBP), the address IS the
information. The conservation law is substrate-independent — it
doesn't matter whether the tape is stored, computed, or given by
physics. The sparse structure you target must be genuinely sparse
(low-rank, k-sparse, RIP-compliant), not just "looked up from a
big table." Look-up = base conversion = no gain.

## SLOS Direction: What the Pipeline Actually Needs

### The Two Computation Paths

| Path | What it computes | Cost | Gives |
|------|-----------------|------|-------|
| SLOS (Perceval, K=2) | Full output distribution, all M_n states | O(n × M_n) | Exact Omega |
| QR/spectral (numpy, K=1) | Eigenvalue spectrum → first column of U | O(n³) | Approximate Omega |

K=2 (SLOS) gives the correct Sidon vs non-Sidon signal (4/4).
K=1 (spectral) gave an INVERTED signal initially.

### What Each Pipeline Query Actually Needs

| Query | Needs SLOS? | Why |
|-------|-------------|-----|
| Omega (exhaust mode probabilities) | YES (K=2) | K=1 approximation was wrong |
| Sidon property check | NO | is_sidon() = pure integer arithmetic |
| GCCL coherence gate | NO | IS the eigenvalue spectrum |
| QR rank/spectrum | NO | O(n³) eigendecomposition |
| Collision count | NO | Integer pairwise-sum check |

The pipeline runs SLOS to compute ONE number (Omega) from the
full M_n-state distribution. Everything else is computed without SLOS.

### The Conservation Law Applied to SLOS

The K=2 interference pattern IS the "residual" — the part the K=1
spectrum can't predict. This is the same conservation law:

  K=1 spectrum (program/model) + K=2 interference (residual) = full distribution

The spectrum captures the sparse structure (block-diagonal modes,
dominant eigenvalues). The K=2 interference captures the dense
correlations (photon-photon interactions, Hong-Ouandel interference).

You can't predict Omega from the spectrum alone because Omega depends
on the K=2 interference — which IS the incompressible residual.

### The Honest Application

The shortcut for the SLOS direction is NOT "skip SLOS using the spectrum."
The shortcut is: **don't run SLOS for queries that don't need it.**

- Sidon check: run is_sidon() (O(n²) integer arithmetic, no SLOS)
- GCCL gate: compute eigenvalue spectrum (O(n³), no SLOS)
- QR rank: compute eigenvalue spectrum (O(n³), no SLOS)
- Collision count: integer pairwise-sum check (O(n²), no SLOS)
- Omega: MUST run SLOS (K=2 interference is irreducible)

The pipeline was running SLOS for ALL queries when only Omega needed it.
The shortcut: skip SLOS for 4/5 queries, run it only for Omega.

### What This Saves

For an n=8, m=8 system:
- Full SLOS: O(8 × 6435) = O(51,480) operations
- Skip SLOS for 4/5 queries: save 4 × 51,480 = 205,920 operations
- Only run SLOS for Omega: 51,480 operations (1/5 of before)

This is a real 5x speedup — not from compression, but from not
running expensive quantum simulations for queries that only need
integer arithmetic or eigenvalue decomposition.

### What This Does NOT Save

The Omega computation itself still needs full SLOS (K=2). The K=2
interference pattern is the irreducible residual. The conservation
law says: you can't predict it from the K=1 spectrum. The spectrum
is the model; the K=2 interference is the tape. Both must be computed.

### The Quandela SLOS as "Weird Machine"

The SLOS IS a weird machine: a photonic quantum simulator repurposed
as a Sidon set classifier. The "weirdness" is that a quantum optics
experiment (designed for photon counting) is being used to detect
combinatorial properties (Sidon sets) of integer sets.

But the conservation law applies: the quantum computation (K=2 SLOS)
gives information that the classical computation (K=1 spectrum) can't
recover. This IS the quantum advantage — not in computation speed,
but in information content. The K=2 interference pattern contains
genuinely new information (the residual) that the K=1 spectrum lacks.

This is the honest quantum advantage: SLOS computes something the
spectrum can't predict. Not because of quantum speedup, but because
the K=2 correlations are fundamentally denser than the K=1 spectrum.

## SLOS is Linear Optical: The Shortcut Works (Revised)

### Why Linearity Changes Everything

The previous analysis said "K=2 interference is irreducible — the
spectrum can't predict it." That's WRONG for linear optical systems.

For a LINEAR optical system:
- The circuit IS a unitary matrix U (n×n)
- The output distribution is determined by U^(⊗m) — the m-fold tensor product
- U^(⊗m) is FULLY determined by U's eigenvalues + eigenvectors
- U's spectrum is O(n²) — already computed (no SLOS needed)

For a GENERAL quantum system:
- Interactions create genuinely new information
- The spectrum of the 1-particle Hamiltonian doesn't determine the
  2-particle output (interactions = nonlinear = new information)
- K=2 IS irreducible from K=1

For a LINEAR optical system:
- NO interactions (photons don't interact — they just interfere at beam splitters)
- U^(⊗m) IS determined by U (the single-particle unitary)
- The eigenvalue PRODUCTS (all m-fold products of U's eigenvalues)
  determine the spectral structure of the output
- The number of distinct products ≤ n^m (much smaller than M_n)

### The Shortcut (Only for Linear Optical)

1. Compute U's eigenvalue decomposition: O(n³)
2. Compute eigenvalue product distribution: O(n^m) — cheap for small m
3. If products are highly degenerate → output concentrated → H low
   → spectrum suffices → SKIP SLOS
4. If products are all distinct → output spread → H high
   → run SLOS for exact probabilities

### Why K=1 Failed but the Full Spectrum Should Work

The K=1 approximation used ONLY the first column of U (one eigenvector
projection). The full spectral prediction needs ALL eigenvalue products
— not just one column. K=1 threw away 15 of 16 spectral directions
for the Sidon crossing matrix.

The correct shortcut isn't "K=1 spectrum vs K=2 SLOS."
It's "FULL spectrum (all eigenvalue products) vs full SLOS."

For the Sidon crossing matrix (4 blocks of 2×2):
- U has 8 eigenvalues (4 pairs)
- U^(⊗2) has ≤ 4² = 16 distinct eigenvalue products
- M_n = C(15,7) = 6435 output states
- 16 spectral directions vs 6435 full states → 400x reduction
- The 16 products capture the block-diagonal structure
- SLOS computes all 6435 → overkill for this circuit

### The Conservation Law (Revised for Linear Systems)

For general quantum: spectrum + interactions = full output
  → interactions are irreducible → can't skip SLOS

For linear optical: spectrum = full output (no interactions)
  → U^(⊗m) is determined by U → CAN skip SLOS
  → the cost is COMPUTATIONAL (M_n is large), not INFORMATION-THEORETIC
  → the information IS in the spectrum, but computing it is expensive
  → the shortcut reduces the computation, not the information

### What This Means for the Pipeline

The photonic Sidon search used SLOS (K=2) because the K=1 approximation
(one column of U) was wrong. But the FULL spectrum (all eigenvalue
products) should give the same answer as SLOS — because SLOS is linear
optical and the output IS determined by U's spectrum.

The shortcut: replace SLOS with full eigenvalue product computation.
For block-diagonal circuits (Sidon): 16 products vs 6435 states.
For random circuits: 64 products vs 6435 states.

This is a REAL shortcut — not from compression, but from the linearity
of the optical system. The conservation law says you can't reduce
information. But for linear systems, the information IS already in the
spectrum — you just need to compute the right function of it (the
eigenvalue products, not just one column).

## Reconciliation: Linear Tool on Linear Problem vs Linear Tool on Nonlinear Problem

### The Apparent Contradiction

Attack 5: "Coherence (Pearson) is linear only — wrong for Sidon sets"
SLOS analysis: "Eigenvalue spectrum works for SLOS circuits"

Both are correct. They talk about different objects:

| Object | Structure | Linear tool works? | Why |
|--------|-----------|-------------------|-----|
| Sidon set | Nonlinear (pairwise sums) | NO | Pearson can't detect quadratic structure |
| SLOS circuit | Linear (unitary matrix) | YES | U^(⊗m) determined by U's spectrum |

### The Principle

The tool must match the problem's structure:
- Linear problem → linear tool (spectrum, QR) → works
- Nonlinear problem → nonlinear tool (is_sidon, permanent) → needed
- Universal linear tool on nonlinear problem → fails
- Universal nonlinear tool on linear problem → overkill

### Why the Coherence Gate Died (Attack 5)

The coherence gate (Pearson correlation) tried to be a UNIVERSAL
admissibility check. It's linear. Sidon structure is nonlinear
(pairwise sums = quadratic). Linear tool on nonlinear problem = fail.

### Why the SLOS Shortcut Works

SLOS is linear optical. The circuit IS a unitary matrix. U^(⊗m) is
determined by U's spectrum. The eigenvalue spectrum is a LINEAR tool
on a LINEAR problem. It works.

### The Conservation Law (Final Form)

For LINEAR systems: spectrum = full information (zero residual)
  → shortcut works (compute spectrum, skip full simulation)
  → the cost is computational (M_n is large), not information-theoretic

For NONLINEAR systems: spectrum + interactions = full information
  → interactions are irreducible residual
  → shortcut fails (must compute the full simulation)
  → the cost IS information-theoretic (interactions create new info)

### Problem-Specific Admissibility (Confirmed)

Each problem needs its own admissibility check matching its structure:
- SLOS (linear): eigenvalue product distribution
- Sidon (nonlinear): is_sidon() integer check
- cmix weights (linear): SVD singular values
- Unit-distance (nonlinear): count_unit_distances()

No universal check. The ManifoldShortcut framework's "problem-specific
admissibility" (from the 5-way attack refinement) is confirmed by
this linearity analysis.

## Capstone: The Octagon Principle

### "If you can't fit a square peg in a triangle hole, turn them both into octagons."

- **Square peg** = nonlinear data (Sidon set, combinatorial structure)
- **Triangle hole** = linear tool (eigenvalue spectrum, SLOS, QR)
- **They don't fit** = Attack 5: linear coherence can't detect nonlinear structure
- **Octagon** = the embedding (matrix) compatible with BOTH

The octagon has MORE sides than either — it's richer, not simpler.
The matrix carries the nonlinear property (Sidon structure) AND has a
linear spectrum (eigenvalue decomposition). Both data and tool are
transformed into the octagon (matrix) where they interface.

### This IS the Observerless Observer Protocol

The invariant (nonlinear property) survives the projection (matrix
embedding) because the spectral signature is preserved. The observer
(linear tool) and the observed (nonlinear data) meet at the octagon
level — the matrix — where the invariant is detectable.

DNA is the octagon carrier: structurally linear (compatible with the
pipeline) but carrying nonlinear meaning (compatible with the problem).
The p-adic valuations are the octagon's corners — linear (prime
factorization) but carrying nonlinear information (set structure).

### What the Session Proved (Measured)

| Problem | Octagon works? | Why |
|---------|----------------|-----|
| Sidon sets | YES (4/4) | Pairwise-sum matrix → eigenvalue degeneracy = spectral signature |
| GW ringdown | Partial (1.5x) | Mode coupling matrix → spectrum works for signal, noise is residual |
| Text | NO (3.088 b/B) | Co-occurrence matrix → spectrum doesn't capture language structure |
| Graph coloring | YES (known) | Adjacency matrix → Hoffman bound: chromatic ≥ max eigenvalue + 1 |

### The Principle

The octagon works when the nonlinear property HAS a linear spectral
signature. It fails when the property is genuinely non-spectral.

The conservation law governs INFORMATION (can't reduce it).
The octagon reduces COMPUTATION (finds the embedding where linear
tools apply). These are different axes:
- Compression: reduce information → blocked by conservation law
- Computation: reduce cost → enabled by the octagon embedding

### The Honest Split

- Compression: dead (conservation law, measured across 8 branches)
- Computation shortcut: alive (octagon embedding, measured for Sidon)
- The pipeline's real value: find the octagon for each problem
  (the matrix embedding where the nonlinear property becomes spectral)