Peer review of v3.0 — the record behind the v3.1 remediation

The v3.1 update post pointed at REVIEW-v3.0.md in a private repository. This is that document, unedited. Line numbers refer to the v3.0 source at commit 3c88f4f; every P0–P2 item was applied in v3.1 and is recorded in Appendix J of the thesis.

Date: 2026-09-11 Scope: latex/ as ground truth (front matter, Parts I–VI, back matter; 438 pp / ~140k words in the pdflatex build). README.md and outreach/kardashev-labs-roadmap.md read as the outreach surface. Purpose stated by the author: grade the thesis for originality, accuracy, consistency, cogency, soundness and utility, where utility means the document’s capacity to generate interest and investment in Kardashev Labs — a Bell-Labs-style institution that would build the pieces of the stack the thesis argues a next monetary system would need. Method: Parts I–II and the front matter were read directly. Parts III, IV–V, VI and the back matter were each refereed in full by an independent reviewer; every high-severity finding was re-checked against the LaTeX source before inclusion (one referee claim was a false positive and was discarded). The build gate (latex/check.sh) was run; the compiled PDF text was extracted to confirm structural defects; the ETP flow mathematics was re-derived; the MATMUL acceptance-bound exploit was reproduced numerically; and the Fed basis-trade FEDS Note, the Washington Post C2PA test, the 23 June 2026 Treasury speech, Kimi K3 and the Tornado Cash legal timeline were web-verified.

Status of this document: review. The remediation pass that follows it (v3.1) is recorded in the changelog appendix (latex/chapters/bm-app-changes.tex).

---

0. Verdict

As scholarship this is a serious, unusually honest, genuinely original conditional thesis that is not yet peer-review-ready. As an instrument for raising investment in Kardashev Labs it does not yet function. The gap between those two judgments is the most important finding.

The compiled PDF contains a duplicated section (§11.9/§11.10). A v3.0 headline change (the demotion of r(W) ≤ 0.3 to an engineering floor) was not propagated to Parts V and VI. Several load-bearing technical claims fail under test: the MATMUL_4096 acceptance bound admits INT8 execution; the viewing-key scope table promises properties no shielded design has; BTC↔shielded-ZEC adaptor swaps do not exist; the “symmetric knowledge” mechanism is not Dang–Gorton–Holmström information-insensitivity and Red Line 17 as written would trip on Bitcoin today. The Sources list has no URLs or DOIs and three verifiable misattributions.

Meanwhile the document’s own concessions — Bitcoin’s buyerless work function is the better money; utility-token precedents reached low-single-digit fee coverage; the fee base arrives only after a clearing event the thesis cannot time — add up to a pitch the text never makes: the underwritable asset is the stack, and the base-asset premium is a free option on top. “Kardashev Labs” does not appear anywhere in the LaTeX source.

---

1. Grades

Criterion	Parts I–II	Part III	Parts IV–V	Part VI	Back matter	Overall
Originality	A−	B	A−	A−	—	A−
Accuracy	B	C	C	B−	C−	C
Consistency	C	C−	C−	C+	B−	C−
Cogency	B−	C+	B−	B	—	B−
Soundness	C+	C	C+	C+	B	C
Utility (Kardashev Labs)	C−	C+	B−	D+	C	D+

Originality — A−

The VerifyPrice/Reach/Settle/Flow family as a verification instrument set, Delivered Verified Capacity as a max-flow issuance bound, the Wrapper Dominance Ratio as a stock/flow pair, the capturable wedge (RL14), the denomination null with its insider/outsider split (RL18), verification modality bands as a monetary rather than engineering distinction, the Capital Survival Ratio, and the framing of a competent closed sovereign stack as the real competitor are genuine contributions. Prior art (TUF, CT, OONI, BIP-324, anonymous credentials, Green’s ETP mechanics) is honestly labelled.

Accuracy — C

The macro spine holds: the FEDS Note figures ($1.85T Cayman Treasuries end-2024, ~$1.4T TIC undercount, 2.1pp saving-rate overstatement), the WaPo C2PA test (7 of 8 platforms stripped credentials), the Treasury speech, Kimi K3, Perez’s terminology and ~20 academic citations all check out. The errors cluster at load-bearing points: the MATMUL acceptance bound (reproduced: INT8-input product passes; a single entry corrupted by ~200× passes); Lloyd’s Register presented as a sampling precedent when it was a per-vessel survey; viewing keys that cannot be revoked, time-bounded or amount-bounded; BTC↔ZEC named as a ready privacy corridor; “boring on the wire” transports that the GFW has fingerprinted since 2022; a Tornado Cash narrative that stops before the August 2025 Storm conviction; Fedimint as non-custodial; “no asset in monetary history” against the Menger/Mises commodity-money account; zero URLs/DOIs in Sources; Culkin’s venue, the COMIT swap authorship and C2PA’s parent organisation wrong.

Consistency — C−

An orphan \section header and leaked draft comments compile into two consecutive sections with identical titles. “Fifteen” red lines survives in three body locations plus README and roadmap; “fifteen empty precedent rows” cites a table that does not exist; “eleven of eighteen … the four they do not reach” (11 + 4 ≠ 18). Part V Invariant 1 and the Part VI Phase I gate still treat r(W) ≤ 0.3 as the economic hinge. RL1’s condition has drifted onto RL10’s series while the text says “do not merge them”. RL5 uses the retired fee+burn quantity. Tier C has four incompatible rules. VerifyPrice is defined three ways. The Monetary Constitution applies a halving schedule to Work Credits. Pre-retraction voice survives in Part II and Part V. The glossary contradicts the body on Sovereign Optionality. Three changelog claims describe edits never made.

Cogency — B−

The nine-link conditional chain is clear and the concessions are propagated honestly. Two gaps are structural. (1) The relocated monetary premium is a holder-side state-contingent service flow, but the thesis’s own instrument for contingent access is the Work Credit; a holder hedging loss of access should hold Work Credits plus Bitcoin, and the text never says why the hedge bid goes to a conceded-worse money. (2) The self-retraction archaeology in Part II leaves the constructive argument buried under its corrections. SNARK-wrapping and side-channel budgets are offered as mechanisms that do not do the claimed work.

Soundness — C

The κ / ε* / WRR derivations and the detection-bound arithmetic verify exactly. Nothing labelled a lemma or corollary is proved. “Symmetric knowledge” is not DGH insensitivity (a payoff-structure property); RL17 would trip on Bitcoin. RL18-B rests on “insiders face no basis risk”, which is false (hashprice is quoted in USD/PH/day). RL16 estimates a regime coefficient from 4–16 observations and kills by low power. Issuance weights of 1.1×/1.2× mint more Work Credits than delivered work. The DVC max-flow is not unit-coherent across the energy→proof chain and double-counts shared capacity across workloads. π-weighted averages of per-state p95s are not percentiles. The tanh capacity modulator is unscaled and moves issuance by ±0.2%, not ±10%.

Utility (for the Kardashev Labs goal) — D+

As a research agenda: strong. The VerifyPrice observatory and adversarial harness, VerifyReach on OONI methodology, the Layer 2 TUF/CT invariants, the PIDL receipt schema, the FCR schema, one BTC↔XMR corridor with VerifySettle, and VerifyFlow wrapper telemetry are buildable; VerifyFlow is sellable today against existing BTC ETFs without the protocol existing. As an investment document: the lab is unnamed in the source; there is no budget, headcount, dated milestone, entity, IP or funding structure; MVS step 2 (a 100%-refund-safe BTC↔ZEC corridor) is blocked at design; the fee base is placed “on the far side of a financial clearing event the thesis cannot time”, which puts Phase III outside any fundable horizon without saying so; and the stack-first / coin-optional thesis the concessions support is never stated.

---

2. Findings

Severity: Critical = blocks peer review or breaks a core definition; Major = a referee would require the change; Minor = should fix. Line numbers are ±3 at v3.0 (3c88f4f).

2.1 Parts I–II and front matter

ID	Location	Finding	Sev	Fix
P2-1	part-ii.tex ~977–993 → PDF §11.9/§11.10 (p.125)	Orphan \section{Economic Linkage…} plus leaked draft comments (“Status: NOT integrated. Depends on Routes A and C”) compile into two consecutive sections with identical titles.	Critical	Delete the orphan header and comment block; rebuild and diff the TOC.
P2-2	part-ii.tex 269–305, 2120–2139; part-v.tex 288	Pre-retraction voice survives: “every use of the triad requires acquiring the base token”, “burns create scarcity … value appreciation. This closes the loop” — the inference §fee-incidence-what-is-established withdraws.	Major	Rewrite in the post-retraction voice (burn = pro-rata buyback; accrual ≠ moneyness).
P2-3	part-ii.tex ~1126 (Value Capture Lemma); Corollary (Market-Price Non-Equivalence)	Lemma says “creates store-of-value premium … only if” while the gloss says it establishes accrual, not moneyness; Corollary is derived from nothing; neither carries a proof.	Major	Restate as a necessary-condition Proposition with an argument; drop or derive the Corollary.
P2-4	sec:information-insensitivity; RL17 (p6-risk.tex 325)	“Symmetric knowledge” is not DGH insensitivity (payoff-structure property). Unit-quality verification is already perfect for Bitcoin, whose haircuts sit at 30–50% and disperse across lenders — RL17 as written trips on Bitcoin.	Critical	Restrict to counterfeit/fraud sensitivity; locate pledgeability in over-collateralized, DVC-backed positions; re-specify RL17.
P2-5	part-ii.tex ~857, ~2404, ~2624	”Lost ~1–2% of supply per year (empirical from BTC)” unsourced; “most successful SoV of the last century by any reasonable measure” unqualified; “Duplex-style PoUW” uncited — Pass, The Economics of Proof-of-Useful-Work (arXiv 2606.06700) bears directly on fee incidence and is never engaged.	Minor	Source or soften; add Pass (2026); engage its Duplexia regime in §fee-incidence.
P1-1	p1-intro.tex 10, 63; p1-threat-model.tex 401	”the fifteen ways to prove us wrong” in the allocator reading path; line 63 lists 13 falsifiers then “full set of eighteen”; stale hardcoded “§22–23”.	Major	”eighteen”; reconcile the list; replace § with \cref.

2.2 Part III (Layers 0–3)

ID	Location	Finding	Sev	Fix
P3-1	part-iii.tex 316	Lloyd’s Register “A1” offered as the century-old precedent for sampled attestation; Lloyd’s classification was a per-vessel survey.	Major	MIL-STD-105 / ANSI Z1.4 acceptance sampling or pharmaceutical lot release; keep Lloyd’s only as the publish-and-underwrite precedent.
P3-2	part-iii.tex 401 vs front-matter.tex 79	”municipal bonds whose coupon is denominated in SLA-backed capacity” contradicts “proofs must not become the bond”.	Major	Delete or reconcile with duration-neutrality.
P3-3	part-iii.tex 184–186 vs 105, 588	Issuance weights 1.1×/1.2× mint more Work Credits than delivered work — WC becomes a subsidy token; conflicts with the DVC issuance bound.	Critical	Differentiate grades by haircut/price only, or fund bonuses from a separate pool with WC redemption unaffected.
P3-4	part-iii.tex 592–597	Zero-DVC scenarios excluded by default and, in the next sentence, excluded only by explicit force-majeure declaration.	Major	Keep the second rule.
P3-5	part-iii.tex 301, 339–347	Detection bound assumes test sensitivity = 1; dopant-level trojans are invisible to imaging; full-die RE at 5–7 nm is not a quarterly activity.	Major	Multiply by σ; target counterfeit dies, binned parts, firmware substitution, metering fraud.
P3-6	part-iii.tex 294, 1319, 1358	SNARK-wrapping a vendor attestation leaves the vendor key as root of trust; “Attestation Without Priesthood” is not earned.	Major	Wrapping buys compression/selective disclosure; trust reduction comes from sampling and open design.
P3-7	part-iii.tex 833–851	Uniform-random-bytes transports are the fingerprint the GFW blocks heuristically since 2022 (Wu et al., USENIX Security 2023); look-like-TLS, Snowflake, Conjure absent.	Major	Separate encrypted from look-like-something transports; add mimicry/refraction fallbacks.
P3-8	part-iii.tex 145 vs 350–356; 214 vs 352; 327–329	Gold ≡ L0-D but Gold’s conditions contain no openness requirement; L0-B’s 80% coverage floor equals Gold’s criterion; units-vs-lots ambiguity.	Major	Make Gold an orthogonal assurance tier or add openness; per-lot minimum.
P3-9	part-iii.tex 570–576, 650–654	DVC max-flow not unit-coherent across energy→proof; per-workload DVC double-counts shared capacity; π-weighted average of per-state p95s is not a percentile.	Major	Generalized/multi-commodity flow; pooled-run percentiles.
P3-10	part-iii.tex 672 vs 675; 1044; 1432 vs 890	Decision-rule clause (iii) references a Physical VerifyPrice SLO band said not to exist; TUF freshness stops rollback not freeze; L1 falsification (>20% ASNs) vs SLO permitting 30% degradation.	Major	Per-state bounds; expiry-bounded timestamp metadata; specify the ASN denominator.
P3-11	part-iii.tex 142–152, 150, 233, 1348	L0-D “5–10+ years” with open PDKs only at 130–180 nm; proof-wrapped accelerators and frontier zk inference presented as engineering; reproducible accelerator firmware unsatisfiable.	Major	Scope L0-D to RoT/RNG/metering; mark verifiable GPU compute and zkML-at-scale as research.

2.3 Parts IV–V (Layers 4–6)

ID	Location	Finding	Sev	Fix
P4-1	part-iv.tex 265, 339, 349	‖C−AB‖∞ ≤ 10⁻³‖A‖∞‖B‖∞ with matrix ∞-norms is a per-row budget ≈1.16×10⁴ at n=4096. Reproduced: honest FP32 error 0.23; INT8-input product error 3.97×10³ passes; one entry corrupted by 1.05×10⁴ passes.	Critical	Componentwise bound |Cᵢⱼ−(AB)ᵢⱼ| ≤ ε_r Σ_k|Aᵢₖ||Bₖⱼ|, ε_r ≈ 10⁻⁵; adversarial test suite.
P4-2	part-v.tex 132; p6-adoption.tex 172 vs part-iv.tex 158–160	Invariant 1 still calls r(W) ≤ 0.3 “the economic hinge … Sev-1”; Phase I still targets it.	Major	Restate in M1/M2 modality-band terms.
P4-3	part-iv.tex 1158	”eleven of the eighteen … the four they do not reach” — 11 + 4 ≠ 18; RL16–18 unmentioned.	Major	”the seven they do not reach”, naming 16–18.
P4-4	part-iv.tex 738, 1042, 1060; part-v.tex 149, 155, 326	Tier C has four incompatible rules.	Major	Issues at 0.3×, collateral-ineligible; propagate.
P4-5	part-iv.tex 108, 504–506; part-v.tex 473; front-matter.tex 366–383	VerifyPrice defined as a five-field tuple, then (t50,t95,e,m,f), then a six-tuple, plus a cost vector — each as the definition.	Major	One canonical tuple; label the others “input views”; price m, b or mark published-not-priced.
P4-6	part-v.tex 260–269, 397, 406	Halving schedule + capacity modulator applied to Work Credit issuance (conflates WC with base asset); tanh(g−g*) unscaled → ±0.2%.	Major	Split base-asset envelope from WC minting; tanh((g−g*)/σ).
P4-7	part-iv.tex 960–969	Viewing keys promised to be time-bounded, amount-bounded, revocable, “no master key”. None holds.	Major	Disclosure-scope model with per-row mechanisms; forward rotation, not revocation.
P4-8	part-iv.tex 841, 861–870, 926; p6-adoption.tex 104, 192; p6-implementation.tex 35	Refund-safety as unilateral recovery ≠ COMIT no-loss/punish; BTC↔shielded-ZEC adaptor swaps do not exist; refund_safe = 1.0 trips on liveness; anon_set ≥ 1000 unreachable for XMR (ring 16), trivial for ZEC.	Major	Bounded-time no-loss exit; scope prior art to XMR; per-asset anonymity metrics; gate BTC↔ZEC.
P4-9	part-v.tex 978–986	Tornado Cash stops at the 2024 reversal. OFAC delisted 21 Mar 2025; Storm convicted 6 Aug 2025 (§1960 conspiracy); Pertsev convicted in NL 2024.	Major	Add outcomes; the lesson is code vs coder; refresh Howey against 2025 SEC staff statements.
P4-10	part-iv.tex 147, 403, 725; 361–381	Freivalds’ O(n²) check labelled a “conjecture”; r ≈ 1/n ignores 3k/n; PoUW usefulness/unpredictability dilemma never stated.	Major	Relabel as consensus-security conjecture; r ≈ 3k/n; name the dilemma as the flagship research problem.
P4-11	part-v.tex 249 vs 989; part-iv.tex 890	Subtractive 3-of-5 emergency powers are a censorship kill switch if signers are compelled; bridge Class A “atomic swap” is not a bridge; Liquid is not bonded.	Major	Advisory-default or bounded to proven refund-safety breaches; fix taxonomy.
P4-12	part-v.tex 839; front-matter.tex 227; p6-risk.tex 134; p6-adoption.tex 42	”fee+burn” survives in four places after part-v.tex 357 pins the retained-fee convention.	Major	Retained-fee coverage everywhere.

2.4 Part VI

ID	Location	Finding	Sev	Fix
P6-1	p6-objections.tex 289; p6-why-money.tex 35	”fifteen ways to show it false”; “fifteen empty precedent rows” — no precedent table exists.	Major	”eighteen”; add the table or say “five”.
P6-2	p6-risk.tex 102 vs 172, 196–203	RL1’s condition drifted to “Physical VerifyPrice p95” — RL10’s series.	Major	Strike “Physical” from RL1.
P6-3	p6-risk.tex 299–302; Warning 13 clause 3	Holder hedging loss of access should hold Work Credits + Bitcoin, not the base asset; the thesis never says why the hedge bid goes to the base asset.	Critical	Answer directly, or state that the base-asset premium is a free option and the underwritable thesis is the stack.
P6-4	p6-risk.tex 384, 393 (RL18-B)	“Insiders face no basis risk”; Bitcoin “cannot trip B” — miners’ costs are fiat-denominated.	Major	Re-specify B around net-denomination of costs.
P6-5	p6-risk.tex 293–297 (RL16)	Regime coefficient from ≥4 quarters across two macro states — kills by low power.	Major	Minimum regime transitions; pre-registered power calculation.
P6-6	p6-why-money.tex 33	”No asset in monetary history is known to have made this crossing” vs Menger/Mises commodity money.	Major	Restrict to equity-like fee-stream claims; engage Menger (1892).
P6-7	RL3, RL5, RL10, RL11, RL13, RL15, RL18; p6-adoption.tex 127, 143	Concentration lines lack maturity gates; fee-collapse threshold 10% / 30% / unnumbered; five red lines have no numeric trigger.	Major	Maturity gates; 30% = phase regression, 10% = red line; publish thresholds.
P6-8	p6-adoption.tex 180; p6-objections.tex 34, 36	Fedimint as non-custodial plumbing (RL2 rejects federated custody); Bitcoin “requires no grid at the moment of use”.	Major	Label Fedimint Class B/C; narrow the Bitcoin claim.
P6-9	p6-implementation.tex; p6-adoption.tex 127; p6-conclusion.tex	No budget, headcount, dated milestones, entity or funding structure; MVS step 2 blocked at design; clearing-event dependency puts Phase III beyond any fundable horizon; “Kardashev Labs” absent.	Critical	Name the lab; add the investment addendum; stack-first / coin-optional.
P6-10	p6-objections.tex 26; p6-risk.tex 372, 392, 403	”fifteen-year record” for Bitcoin (17 years).	Minor	”seventeen-year” / “since 2009”.

2.5 Back matter

ID	Location	Finding	Sev	Fix
BM-1	bm-sources.tex (84 entries)	Zero URLs and DOIs; Sourcing Note promises “linked” sources; ~45% blogs/Substacks/litepapers; nine undated; no \cite mechanism.	Critical	URL/DOI on every entry; date commentary; keyed list.
BM-2	bm-sources.tex 34, 254, 342, 161	Culkin (1967) was in The Saturday Review, not America; COMIT swap is Gugger (h4sh3d) + Eizinger et al., not Reitwießner; C2PA is Linux Foundation JDF, not Creative Commons; FEDS Note lists no authors.	Major	Correct all four.
BM-3	bm-app-market-realization.tex 100, 5–6; front-matter.tex 581–582	”κ = εAr” for unlevered (κ = ε; εAr is Q^RC); front matter says κ/WRR require L>1 while the appendix defines them for inverse L; Cheng–Madhavan (2009) never cited.	Major	κ = ε; domain L∉{0,1}; cite Cheng–Madhavan, Avellaneda–Zhang.
BM-4	bm-glossary.tex 67 vs part-iii.tex 584	Glossary: Sovereign Optionality “index of the number of pathways”; body: “does not count pathways”. WDR, Homestead Ratio, MPR, CCR/SER/RCR/NUS, Participation Line have no entries.	Major	Copy body wording; add entries.
BM-5	bm-app-changes.tex 65, 49, 37 vs p6-closed-stack.tex 107, part-iii.tex 688	Three changelog claims describe edits never made.	Major	Make the edits.
BM-6	bm-app-sdk.tex 115–116; bm-app-verifyprice.tex JSON	SDK Example 3 settles payroll in USDt (the bypass channel); Receipt ≠ PIDL Def 3; r_max = 0.21 for a succinct workload violates the M1 band.	Major	Settle in base asset; align Receipt; r_max ≈ 5×10⁻⁴.

2.6 Outreach surface

ID	Location	Finding	Sev	Fix
OUT-1	README.md 5, 13, 15, 159, 232; outreach/kardashev-labs-roadmap.md	README advertises v1.9, “fifteen named red lines”, 358 pages; roadmap says fifteen and has no budget, team, timeline or sized deliverables.	Major	Regenerate from v3.x; rewrite the roadmap as a fundable plan.

---

3. Verification log

Item	Check	Result
Build gate	tectonic via check.sh: 438 pages, reference integrity clean; committed PDF matches	Held
Cross-references	351 labels, 124 back-matter \cref targets, 0 unresolved; no hardcoded § in body except two stale hits	Held
Counts	18 red lines contiguous; 10 premises, 9 links, 9 requirements, 12 primitives, 4 apps, 12 boards, 13 warnings	Held
FEDS Note (Oct 2025)	$1.85T Cayman; ~$1.4T TIC undercount; 2.1pp saving-rate — Barth, Beltran, Hoops, Kahn, Liu, Perozek	Held (authors missing in Sources)
WaPo C2PA test (22 Oct 2025)	7 of 8 platforms stripped credentials	Held
Kimi K3; Treasury speech 23 Jun 2026	2.8T params, open weights; “economic security begins with national capacity”	Held
Part I macro arithmetic	1.025/1.06−1 ≈ −3.3%; 3yr ≈ −9.6%; +300bp on D≈8 ≈ −24%; cumulative ≈ −31%	Held
ETP flow model	κ = L(L−1)+εL²; ε* = −(L−1)/L; WRR = −εL/(L−1); κ = L(L−1)(1−WRR)	Held
Sampling bound	1−0.95⁵⁰ = 0.923; rule of three 3/50 = 6% vs exact 5.8%	Held (sensitivity omitted)
Academic sources	Gorton–Pennacchi, DGH, Holmström, Gorton–Metrick, Doepke–Schneider, Baumol, Grossman–Stiglitz, Kaldor, Working, Perez, Eisenstein, McLuhan, Ong, Anderson, RFC 6962, BIP-324	Held
MATMUL bound	INT8 execution and 200× single-entry corruption pass	Failed
Tornado Cash timeline	Delisting 21 Mar 2025; Storm convicted 6 Aug 2025	Thesis stale
Lloyd’s Register	Per-vessel survey, not sampled attestation	Failed
”Kardashev Labs” in LaTeX	0 occurrences	Absent

---

4. Fix plan

P0 — before any external review

1. Remove the duplicate §Economic Linkage header and leaked draft comments; rebuild.
2. fifteen → eighteen at p1-intro.tex 10, p6-objections.tex 289, README, roadmap; fix “fifteen precedent rows” and “11 of 18 / four unreached”.
3. Replace the MATMUL ∞-norm acceptance bound with a componentwise bound; add an adversarial test suite.
4. Propagate the modality-band change: part-v.tex 132 Invariant 1, p6-adoption.tex 172.
5. RL1: strike “Physical”. RL5 and all “fee+burn”: retained-fee convention. Unify Tier C. One VerifyPrice tuple with named views.
6. Split the Monetary Constitution into base-asset envelope vs Work Credit minting; scale the tanh modulator.
7. Rewrite surviving pre-retraction passages (part-ii.tex 269–305, 2120–2139; part-v.tex 288).
8. Sources: URL/DOI on every entry; fix Culkin, COMIT/Gugger, C2PA; author the FEDS Note; cite Cheng–Madhavan in the appendix.
9. Update Tornado Cash (Storm conviction) and the Howey section against 2025 SEC staff statements.
10. Delete or reconcile the municipal-bond paragraph with “proofs must not become the bond”; replace the Lloyd’s analogy.

P1 — soundness repairs a referee will demand

1. Re-specify information-insensitivity (counterfeit/fraud sensitivity only; pledgeability via over-collateralized DVC-backed positions) and re-test RL17 against Bitcoin.
2. Answer why the state-contingent hedge bid goes to the base asset rather than Bitcoin + Work Credits — or concede it and pivot the monetary claim to a free option.
3. RL18-B: insiders do face fiat basis risk. RL16: add regime-transition minimum and power calculation. Add maturity gates to RL3/11/13; publish missing thresholds.
4. Fix WC issuance weights >1; resolve the zero-DVC default; make DVC a generalized/multi-commodity flow; compute adjusted percentiles over pooled runs.
5. Add test sensitivity σ to the sampling bound; re-target sampling at counterfeit/binning/firmware substitution.
6. State the PoUW usefulness/unpredictability dilemma as the flagship research problem; relabel Freivalds as a theorem; r ≈ 3k/n.
7. Viewing keys → disclosure scope with per-row mechanisms; refund-safety → bounded-time no-loss exit; per-asset anonymity metrics; gate BTC↔ZEC.
8. Layer 1: add look-like-TLS / Snowflake / Conjure transports; TUF freeze protection; fix the ASN denominator.
9. Emergency powers: advisory-default or bounded to proven refund-safety breaches.
10. Engage Menger/Mises on the base-rate claim; engage Pass (2026) on PoUW economics.

P2 — turning the thesis into a Kardashev Labs case

1. Name Kardashev Labs in the conclusion and front matter; state the lab’s relationship to the protocol.
2. Investment addendum: phased 18/36-month budget, team shape, dated deliverables mapped to the MVS table, entity/funding/IP structure.
3. Make the stack-first / coin-optional thesis explicit: what is underwritable vs what is a free option.
4. Proving-cost roadmap with the modality bet stated; L0-A reference prover/verifier BOM; first L0-C open components (RoT, RNG, metering).
5. Single workload registry table; KPI appendix as metric · unit · source · cadence · board · red line · threshold.
6. Regenerate README and the roadmap from the current source; fix the SDK payroll example.

---

5. What to keep

The honesty is the asset. The Bitcoin concession, the Travel Rule correction, the Howey self-critique, the fee-incidence retraction and the “who finances the reactor” concession are argued against the author’s own interest and are the most credible passages in the document. The fix is not to soften them; it is to stop letting them sit next to un-retracted claims and to draw the investment conclusion they imply.