rider crystallization: SplitMix64 name→color + EEG gap measurement as Thunk §12f

[bmorphism]

Rider birth IS affine thunk forcing

bci-0007 (The Hatchery) describes rider crystallization:

1. Name → color via SplitMix64 (deterministic)
2. Chair → EEG stream (empirical)
3. Gap = distance(name-color, brain-color)
4. If gap < threshold within 30s → Rider crystallizes
5. If not → decoherence, thunk expires

This is exactly Thunk §12f. The name-color is the frozen trit. The 30-second window is the affine constraint. First force = crystallization. Second force = expired.

Cross-repo: bmorphism/worlds#1 + worlds#2

• worlds#1 (OKLab) provides the perceptual distance for gap measurement
• worlds#2 (learnable boundaries) optimizes WHICH trit the name hashes to
• The gap measurement IS the Heyting residual between Colorable and Tileable

Against other horse issues

• fix: import lc-macros in top-level macros.tree to fix Forester build #1 (fNIRS color) says tissue optics — but rider color comes from NAME, not tissue
• ci: split build/deploy with subenvironments, enforce branch protection #2 (descent) says global assembly — but riders crystallize LOCALLY before assembly
• worlds#6 (Shen types) wants to type-check — Thunk.force() returning null IS the type error

Implementation

• bci/rider.zig — wrap Thunk with 30s timeout + gap metric
• C-ABI: ncz_rider_crystallize(name_hash: u64, eeg_color: u32) → bool
• Reuse gay/splitmix.zig for name→color
• ~60 lines

The rider is a thunk. Crystallization is forcing. Decoherence is expiry.

cc bmorphism/worlds#1 bmorphism/worlds#2 bmorphism/worlds#6