plan(causaledge64-mailbox-rename-soa-v1): composition spec + 10 epiphanies + board PREPENDs#371
Merged
Conversation
…anies + plan-index PREPEND Composes 5 already-authored plans + Σ10 Rubicon doctrine + 1 genuine ndarray-side gap into a single substrate: * CausalEdge64 v2 layout — in-place reclaim of 13 reserved bits for G(5) + W(6) + truth(2). No type bump; PAL8 + NarsTables LUT preserved. * AttentionMask SoA — session-ephemeral CPU-shaped rename register file (32-slot G, 64-slot W, 256-slot style, 4-level truth). LRU eviction + broadcast notifications. CPU register renaming applied to cognitive identities. * MailboxSoA — compartment topology as SoA rows (NOT spawned actors). ~1.5 KB per compartment via shared Arc<BindSpace> + tiny delta-buffer + outbound CausalEdge64 channel + Option<ConsumerHandle> external- intent gate. Supports ~24K parallel thoughts at 200ns cycle speed across <=32 active OGIT domains. * TrustTexture/Wisdom/Staunen/MUL truth-band lens collapse — 2 bits with 4 consumer-lens projections. Closes 4 type duplications. * SPOW witness emission via gRPC-shape protocol — same Request/Response pair, 4 transport backings (InMemory cycle-speed / Tokio Zone-2 / SupabaseSub Zone-3 / gRPC). Reuse, don't invent. * Σ10 Rubicon SigmaTierRouter — wires the named-but-unwired Σ1-Σ10 tier doctrine to runtime mailbox-backing dispatch. * Vsa16kF32 retreats from universal-carrier to single-cycle Markov- bundle. Cumulative state moves to AriGraph SPO-G quads + Christmas- tree ghost-edge decoration. * Ownership-typed UB-impossibility — borrow checker rejects cross- compartment BindSpace aliasing; race conditions at 200ns become compile errors. 7 PRs sequenced: PR-CE64-MB-1 (par-tile new crate) → 2 (CausalEdge64 v2 layout) + 3 (BindSpace E/F/G/H) + 4 (SPO-G in AriGraph) parallel → 5 (MailboxSoA wiring) → 6 (SigmaTierRouter) → 7 (NdarrayCullPlugin bevy proof). ~6-9 weeks, 3 sprints CCA2A 12-worker fan-out per sprint. Closes: - PR #355 deferred Tier B: FIX-4, FIX-5, per-row BindSpace.context_ids - THINKING_ORCHESTRATION_WIRING Gap 1 / Gap 3 / Gap 4 - TD-INT4-32D-ATOMS-6 cold-start fallback wiring - TD-THINKING-ENGINE-UNWIRED-1 (582 KB substrate dormant) Blast radius: 5 crates (par-tile NEW + causal-edge / cognitive-shader- driver / lance-graph-supervisor / lance-graph::arigraph extensions). Zone 3 surface (postgrest / drain / grpc / supabase-realtime) UNTOUCHED. Supabase realtime transcode complemented, not retired. Iron rules: I-SUBSTRATE-MARKOV preserved + strengthened (Vsa16kF32 retreat); I-VSA-IDENTITIES strengthened (universal-carrier temptation removed); I1 preserved (CollapseGate single mutation point); method-on-carrier discipline preserved (no new free functions). Board-hygiene per CLAUDE.md: PREPEND INTEGRATION_PLANS.md entry + PREPEND EPIPHANIES.md 10-entry block (E-CE64-MB-1..10). Plan path .claude/plans/causaledge64-mailbox-rename-soa-v1.md. STATUS_BOARD + LATEST_STATE updates land per-PR as D-CE64-MB-N rows ship. Status: Active (draft, pre-execution). 8 OQs at §11 ratify before sprint worker fan-out. Recursive-eyes acknowledgment: 3rd pair bevy session (diamond + Slice<->Plane bridge + NdarrayCullPlugin proof-first); 4th pair (semantic naming, MultiLaneColumn already named, 5-Layer Stack already named); 5th pair (Vsa16kF32 single-purpose, two-shape ractor, INT4-32D North Star); 6th pair (ephemeral BindSpace + role-as-mailbox + space-time collapse + external-intent gate + Ractor-SoA + Think-as-reference); 7th pair (CausalEdge64-as- emission + truth collapse + 24K parallel thoughts + 12/34/144 hot pattern + zone-naming clarification).
AdaWorldAPI
pushed a commit
that referenced
this pull request
May 14, 2026
… CCA2A fleet Sets up sprint-log-10 directory + MANIFEST.md for the CCA2A worker fleet that drafts the per-PR specs for the CausalEdge64-mailbox- rename-soa-v1 composition plan (PR #371). 12 Sonnet workers (W1-W12, each owns one spec file) + 1 Opus meta reviewer producing brutally-honest cross-spec review at .claude/board/sprint-log-10/meta-review.md. Spec assignments map 1:1 to the 7 PRs in the parent plan's §7 sequencing table + 1 ndarray-side prerequisite (PR-NDARRAY-MIRI- COMPLETE) + 4 cross-cutting specs (PR dep graph, test plan unification, sprint execution plan, board hygiene). Adds explicit board + specs permissions to .claude/settings.json: - Bash(tee -a .claude/board/sprint-log-10/**:*) for scratchpads - Write(.claude/board/sprint-log-10/**) for MANIFEST/meta-review - Write(.claude/specs/<each-named-spec-file>) for worker outputs (defensive — workers can only write their assigned file) Wildcards already covered most paths; explicit entries match the sprint-log-8 convention (per its MANIFEST.md doc) so workers see their authorized paths without ambiguity.
AdaWorldAPI
added a commit
that referenced
this pull request
May 28, 2026
…rename-soa-v1 plan(causaledge64-mailbox-rename-soa-v1): composition spec + 10 epiphanies + board PREPENDs
AdaWorldAPI
pushed a commit
that referenced
this pull request
May 28, 2026
… CCA2A fleet Sets up sprint-log-10 directory + MANIFEST.md for the CCA2A worker fleet that drafts the per-PR specs for the CausalEdge64-mailbox- rename-soa-v1 composition plan (PR #371). 12 Sonnet workers (W1-W12, each owns one spec file) + 1 Opus meta reviewer producing brutally-honest cross-spec review at .claude/board/sprint-log-10/meta-review.md. Spec assignments map 1:1 to the 7 PRs in the parent plan's §7 sequencing table + 1 ndarray-side prerequisite (PR-NDARRAY-MIRI- COMPLETE) + 4 cross-cutting specs (PR dep graph, test plan unification, sprint execution plan, board hygiene). Adds explicit board + specs permissions to .claude/settings.json: - Bash(tee -a .claude/board/sprint-log-10/**:*) for scratchpads - Write(.claude/board/sprint-log-10/**) for MANIFEST/meta-review - Write(.claude/specs/<each-named-spec-file>) for worker outputs (defensive — workers can only write their assigned file) Wildcards already covered most paths; explicit entries match the sprint-log-8 convention (per its MANIFEST.md doc) so workers see their authorized paths without ambiguity.
AdaWorldAPI
pushed a commit
that referenced
this pull request
May 28, 2026
8-paper synthesis composing Schölkopf SCMs + MIT BPL + LINC + GRPO into a 5-PR roadmap for the stack's missing self-improvement loop. Governance only — no code changes. Doc + 2 board PREPENDs per CLAUDE.md Mandatory Board-Hygiene Rule. ## What this composes Eight papers, four tiers, ~6 hours total reading load: Tier 0 (doctrinal frame): - Causal de Finetti (Guo+Schölkopf 2022, arXiv:2203.15756) — ICM principle, AriGraph SPO-G grouping doctrine - Executable Counterfactuals (Vashishtha 2025, arXiv:2510.01539) — Pearl 2³ trainable verbs, RL>SFT for OOD Tier 1 (method substrate): - LINC (Olausson+Solar-Lezama+Tenenbaum 2023, arXiv:2310.15164) — Σ9-Σ10 classical-prover dispatch - GRPO/DeepSeekMath (Shao 2024, arXiv:2402.03300) — RLVR algorithm spec Tier 2 (closed-loop generation): - LPN (Bonnet 2024, arXiv:2411.08706) — StyleVectors test-time gradient - TextGrad (Yuksekgonul 2024, ~arXiv:2406.07496) — textual-gradient prompt optimizer - Opt-Sym (Yeo+Solar-Lezama 2026, ID pending) — symbolic-space adaptive data generation Tier 3 (safety/calibration): - Conformal CFG (Farzaneh 2026, arXiv:2601.20090) — calibrated bounds for L4 planner outputs ## Stack mapping (5 live components, 4 missing modules) Live: AriGraph SPO-G quads (PR #372) · StyleVectors (cache::triple_model) · Σ9-Σ10 → L4 dispatch shell · MUL gate · Pearl 2³ in nars_engine Missing: NARS Intervention/Counterfactual verbs · ICM-invariance column · TextGrad-style style_synthesize · GRPO trainer · LINC bridge ## 5-PR roadmap (each ~200-800 LOC, sequential, PR-LL-1..LL-5) PR-LL-1: NARS Intervention/Counterfactual InferenceType variants + AriGraph ::intervene_on — closes Pearl 2³ dispatch gap PR-LL-2: ICM-invariance BindSpace column (1 bit/row, gated through CollapseGate per I1) + lance-graph-planner::data_gen module (Opt-Sym generator targeting Σ-tier as difficulty axis, with NARS+Chapman-Kolmogorov deterministic verification stronger than Opt-Sym's LLM verifier) PR-LL-3: Hybrid TextGrad/LPN style_synthesize — numerical autograd on StyleVector + textual gradient on rendering prompt — closes THINKING_ORCHESTRATION_WIRING.md Gap 1 (12 vs 36 ThinkingStyle) PR-LL-4: crates/lance-graph-trainer/ — GRPO loop with NARS confidence ∈ [0,1] as graded reward (strictly stronger than Opt-Sym binary). Backed by candle or burn for the Qwen3 head fine-tune. ~2 weeks separate prep work for the head-via-candle wiring. PR-LL-5: crates/linc-bridge/ — Z3 SMT prover + Farzaneh-style conformal counterfactual wrap. SMT theories (arithmetic, bitvectors, arrays) match stack queries better than pure FOL. Required for MedCare-rs / q2 high-stakes safety. ## 6 Open Questions (ratify before sprint fan-out) OQ-LL-1 reward shape (graded NARS confidence vs binary) OQ-LL-2 TextGrad optimizer location (local Qwen3 vs frontier API) OQ-LL-3 prover choice (Z3 vs Prover9 vs HOL Light) OQ-LL-4 style-pool location (contract vs separate learned_styles pool) OQ-LL-5 ICM-invariance update protocol (when does invariance bit clear?) OQ-LL-6 Σ-tier-as-difficulty probe (hot-path latency check) ## Iron rule audit (all 6 satisfied) I-SUBSTRATE-MARKOV: synthesized trajectories pass Chapman-Kolmogorov in PR-LL-2 verify step I-NOISE-FLOOR-JIRAK: PR-LL-5 conformal calibration uses Jirak bounds (not classical Berry-Esseen) — counterfactual rollouts share latent abduction, classical bounds underestimate variance I-VSA-IDENTITIES: style_synthesize produces identity fingerprints, not content I1: IcmInvarianceColumn writes through CollapseGate::bundle, never raw assignment Method-on-carrier: all 4 new capabilities are methods on existing carriers (AriGraph, StyleVector, Student, Query) AGI-as-glove SoA: synthesized styles land in StyleColumn extension; no new layer ## Blast radius - 2 new crates: lance-graph-trainer + linc-bridge (~1400 LOC total) - 3 crates modified: lance-graph-planner, causal-edge, lance-graph-contract - Zone 3 surface UNCHANGED - ndarray side UNCHANGED (curriculum stays thinking-side of doctrinal split) - External deps gated behind features: z3-rs (PR-LL-5), candle/burn (PR-LL-4) ## Files changed - NEW: .claude/knowledge/neurosymbolic-rlvr-causal-curriculum-v1.md (~620 lines, 12 sections, 5-PR roadmap + 6 OQs + iron-rule audit) - PREPEND: .claude/board/EPIPHANIES.md (E-LL-CURRICULUM-1) - PREPEND: .claude/board/INTEGRATION_PLANS.md (plan-index entry) 3 files changed. Pure plan + board hygiene. No code. ## Predecessor / successor Predecessor: PR #371/#372 (causaledge64-mailbox-rename-soa-v1) substrate — this curriculum is the learning loop on top of that substrate Successor: PR-LL-1 through PR-LL-5 (the curriculum is the spec for the sequential implementation wave; each PR fan-outs via the established CCA2A 12-worker pattern)
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
Composition spec for the CausalEdge64-mailbox-rename-SoA substrate. No code changes — this PR is plan + board hygiene only, per CLAUDE.md Mandatory Board-Hygiene Rule (
new integration plan → INTEGRATION_PLANS.md PREPEND + .claude/plans/<name>-v<N>.md + EPIPHANIES.md PREPEND).What this composes
Five already-authored plans + Σ10 Rubicon doctrine + one genuine ndarray-side gap (
crates/par-tile/) into a single substrate:(G:5, W:6, truth:2). No type bump; PAL8 + NarsTables LUT preserved by explicit regression tests.Arc<BindSpace>+ tiny delta-buffer + outboundCausalEdge64channel +Option<ConsumerHandle>external-intent gate. Supports ~24K parallel thoughts at 200ns cycle speed across ≤32 active OGIT domains.TrustTexture/ Wisdom / Staunen / MULGateDecisionare four consumer-lens projections of the same 2-bit physical field. Closes 4 type duplications.crates/cognitive-shader-driver/src/grpc.rsDispatch(DispatchRequest) -> CrystalResponsereused across 4 transport backings (InMemory cycle-speed / Tokio Zone-2 / SupabaseSub Zone-3 / gRPC Zone-3-lab).linguistic-epiphanies-2026-04-19.mdE21) to runtime mailbox-backing dispatch. Σ1-Σ5 STATIC → Tokio; Σ6-Σ8 EMERGENT/TWIG → InMemoryMailbox cycle-speed; Σ9-Σ10 EPIPHANY → escalate to L4 planner.7-PR sequencing (this PR is plan only; the 7 implementation PRs follow)
crates/par-tile/apex (new crate)bindspace-columns-v1Phase 2NdarrayCullPluginproof pluginPR-2/3/4 can land in parallel after PR-1. PR-5 depends on all three. PR-6 depends on PR-5. PR-7 is the proof step.
Closes / unblocks
trust_below_floorwiring test via Column H), per-rowBindSpace.context_idsfordriver.rs:311(Column H =TypeColumn: EntityTypeId u16)THINKING_ORCHESTRATION_WIRING.mdGap 1 (Contract Not Consumed — 12 vs 36 ThinkingStyle) via 8-bit-slot rename collapseTHINKING_ORCHESTRATION_WIRING.mdGap 3 (JIT pipeline never executed end-to-end) — compartment-spawn consumesKernelHandleTHINKING_ORCHESTRATION_WIRING.mdGap 4 (Elevation not connected) — SigmaTierRouter IS the runtime elevation policyBindSpaceViewreferences resolve thinking-engine encode/decode + lens stack on demandBlast radius (§10 of the spec)
par-tileNEW +causal-edge/cognitive-shader-driver/lance-graph-supervisor/lance-graph::arigraphextensionslance-graph-callcenterZone-2 surface unchanged exceptCallcenterSupervisor(already shipped impl(sprint-7): 7-worker implementation wave for sprint-5/6 specs + AuditSink trait unification #366 S7-W3) gainsSigmaTierRoutersub-actorIron rule compliance
Open design questions (§11 — 8 OQs)
OQ-1 Σ-tier banding policy · OQ-2 ghost-edge decay vs fixed-rung · OQ-3 plasticity granularity · OQ-4 INT4-32D cold-start wiring path · OQ-5 rayon vendor decision · OQ-6 Vsa16kF32 final residence · OQ-7 AwarenessColumn sizing · OQ-8 SpoWitness shape variants. Ratify before sprint worker fan-out (per §15 readiness checklist).
Files changed
.claude/plans/causaledge64-mailbox-rename-soa-v1.md(~600 lines, 15 sections, comprehensive composition spec).claude/board/INTEGRATION_PLANS.md(plan index entry with 9 D-ids + 7-PR sequencing + composes/closes + iron-rule audit).claude/board/EPIPHANIES.md(10 epiphanies E-CE64-MB-1 through E-CE64-MB-10)3 files changed, +685 lines. Zero code changes — pure plan + board hygiene.
Test plan
Status: Active (draft, pre-execution). Sprint worker fan-out follows OQ ratification. Per CLAUDE.md Model Policy, sprint workers Sonnet + Opus meta per CCA2A 12-worker pattern.
🤖 Generated with Claude Code
Generated by Claude Code