pipeline: archive 1 conflict-closed source(s)

Pentagon-Agent: Epimetheus <3D35839A-7722-4740-B93D-51157F7D5E70>

agents/rio/musings/research-2026-03-19.md

@@ -0,0 +1,176 @@
+---
+type: musing
+agent: rio
+title: "Does the typical MetaDAO governance decision meet futarchy's manipulation resistance threshold — and what does FairScale mean for Living Capital's investment universe?"
+status: developing
+created: 2026-03-19
+updated: 2026-03-19
+tags: [futarchy, manipulation-resistance, metadao, living-capital, p2p-ico, fairscale, implicit-put-option, liquidity-threshold, disconfirmation, belief-1, belief-3, ninth-circuit, clarity-act]
+---
+
+# Research Session 2026-03-19: Liquidity Thresholds and Living Capital Design
+
+## Research Question
+
+**Does the typical MetaDAO governance decision meet the "liquid markets with verifiable inputs" threshold that makes futarchy's manipulation resistance hold — and if thin markets are the norm, does this void the manipulation resistance claim in practice?**
+
+Secondary: What does the FairScale implicit put option problem mean for Living Capital's investment universe?
+
+## Disconfirmation Target
+
+**Keystone Belief #1 (Markets beat votes)** has been narrowed over four sessions:
+- Session 1: Narrowed — markets beat votes for *ordinal selection*, not calibrated prediction
+- Session 4: Narrowed further — conditional on *liquid markets with verifiable inputs*
+
+The scope qualifier "liquid markets with verifiable inputs" is doing a lot of work. My disconfirmation target: **How frequently do MetaDAO decisions actually meet this threshold?**
+
+**What would confirm the scope qualifier is not void:** Evidence that MetaDAO's contested decisions have sufficient liquidity and verifiable inputs as a norm.
+
+**What would void it:** Evidence that most MetaDAO governance decisions occur with thin trading volume, making FairScale-type implicit put option risk the typical condition.
+
+## Key Findings
+
+### 1. The $58K Average: Thin Markets Are the Norm
+
+**Data point:** MetaDAO's decision markets have averaged $58K in trading volume per proposal across 65 total proposals (through ~Q4 2025), with $3.8M cumulative volume.
+
+**Why this matters for the disconfirmation question:**
+
+At $58K average per proposal, the manipulation resistance threshold is NOT reliably met for most governance decisions. The FairScale liquidation proposer earned ~300% return on what was likely well below $58K in effective governance market depth. A $58K market can be moved by a single moderately well-capitalized actor.
+
+The flagship wins are survivorship-biased:
+- The VC discount rejection (16% META surge) was governance of META itself — MetaDAO's own token, the most liquid asset in the ecosystem
+- This is not representative of ICO project governance
+
+**The distribution problem:** We don't have proposal-level data, but the $58K average likely masks a highly skewed distribution where MetaDAO's own governance decisions (high liquidity) pull up the mean while most ICO project governance decisions occur well below that level.
+
+**DeepWaters Capital's framing:** "Decision markets currently function primarily as signal mechanisms rather than high-conviction capital allocation tools." This is the MetaDAO valuation community's own assessment.
+
+### 2. The 50% Liquidity Borrowing Mechanism Codifies Market-Cap Dependency
+
+The Futarchy AMM borrows 50% of a token's spot liquidity for each governance proposal. This means:
+
+- Governance market depth = 50% of spot liquidity = f(token market cap)
+- Large-cap tokens (META at $100M+ market cap): deep governance markets, manipulation resistance holds
+- Small-cap tokens (FairScale at 640K FDV): thin governance markets, FairScale pattern applies
+
+This is not a bug — it's a design feature. The mechanism solves the proposer capital problem (previously ~$150K required to fund proposal markets). But it TIES governance quality to market cap.
+
+**The implication:** The manipulation resistance claim works exactly where you'd expect voting to also work (established protocols with engaged communities and deep liquidity). It's weakest exactly where you most need it (early-stage companies with nascent communities and thin markets).
+
+**Kollan House's "80 IQ" framing:** MetaDAO's own creator described the mechanism as "operating at approximately 80 IQ — it can prevent catastrophic decisions but lacks sophistication for complex executive choices." This is intellectually honest self-scoping from the system designer. The manipulation resistance claim's advocates need to incorporate this scope.
+
+### 3. FairScale Design Fixes: All Three Reintroduce Off-Chain Trust
+
+Pine Analytics documented three proposed solutions post-FairScale:
+1. Conditional milestone-based protections → requires human judgment on milestone achievement
+2. Community-driven dispute resolution → requires a trusted arbiter for fraud allegations
+3. Whitelisted contributor filtering → requires curation (contradicts permissionlessness)
+
+All three require off-chain trust assumptions. There is no purely on-chain fix to the implicit put option problem when business fundamentals are off-chain.
+
+**Critical observation:** MetaDAO has implemented no protocol-level design changes since FairScale (January 2026). P2P.me (launching March 26) has 50% liquid at TGE — the same structural risk profile as FairScale. No milestones, no dispute resolution triggers. The ecosystem has not updated its governance design in response to the documented failure.
+
+### 4. Living Capital Design Implication: A Minimum Viable Pool Size Exists
+
+**The FairScale case maps directly to Living Capital's design challenge.** Living Capital invests in real companies with real revenue claims — exactly the scenario where futarchy governance faces the implicit put option problem.
+
+The 50% liquidity borrowing mechanism points to a specific design principle:
+
+**Governance market depth = 50% of pool's spot liquidity**
+
+For manipulation resistance to hold, the governance market needs depth exceeding any attacker's capital position. A rough threshold: if the pool's liquid market cap is below $5M, the governance market depth (~$2.5M) is probably insufficient for contested high-stakes decisions. Below $1M pool, governance decisions resemble FairScale dynamics.
+
+**This suggests a minimum viable pool size for Living Capital governance integrity:**
+- Below ~$1M pool: governance markets too thin, Living Capital cannot rely on futarchy manipulation resistance for investment decisions
+- $1M-$5M pool: borderline, futarchy works for clear cases, fragile for contested decisions
+- $5M+ pool: manipulation resistance holds for most realistic attack scenarios
+
+**The first Living Capital vehicle (~$600K target) is below this threshold.** This means the initial vehicle would be operating in the FairScale-risk zone. Options:
+1. Accept this and treat the initial vehicle as a trust-building phase, not a futarchy-reliant governance phase
+2. Target $1M+ for the first vehicle
+3. Supplement futarchy governance with a veto mechanism for the initial phase (reintroducing some centralized trust)
+
+### 5. Regulatory Picture: No Near-Term Resolution, Multiple Vectors Worsening
+
+**Ninth Circuit denies Kalshi stay (TODAY, March 19, 2026):**
+- Ninth Circuit denied Kalshi's motion for administrative stay
+- Nevada can now pursue TRO that could "push Kalshi out of Nevada entirely for at least two weeks"
+- Circuit split now confirmed: Fourth Circuit (Maryland) + Ninth Circuit (Nevada) = pro-state; Third Circuit (NJ) = pro-Kalshi
+- SCOTUS review increasingly likely in 2026/2027
+
+**CLARITY Act does NOT include express preemption for state gaming laws:**
+- Section 308 preempts state securities laws for digital commodities — NOT gaming laws
+- Even CLARITY Act passage leaves the gaming classification question unresolved
+- The "legislative fix" I flagged in Session 3 doesn't exist in the current bill
+- CLARITY Act odds have also dropped from 72% to 42% due to tariff market disruption
+
+**CFTC ANPRM silence on governance markets (confirmed):**
+- 40 questions cover sports/entertainment event contracts
+- No mention of governance markets, futarchy, DAO decision-making, or blockchain-based governance prediction markets
+- Comment window open until ~April 30, 2026
+- No MetaDAO ecosystem comment submissions found
+
+**Combined regulatory picture:** No legislative resolution (CLARITY Act doesn't fix gaming preemption). No near-term regulatory resolution (CFTC ANPRM can define legitimate event contracts but can't preempt state gaming laws). Judicial resolution heading to SCOTUS in 2026/2027. Meanwhile, state enforcement is escalating operationally (Arizona criminal charges + Nevada TRO imminent). The regulatory situation has worsened since Session 3.
+
+## Disconfirmation Assessment
+
+**Question:** Does the typical MetaDAO governance decision meet the "liquid markets with verifiable inputs" threshold?
+
+**Finding:** NO — the $58K average across 65 proposals, combined with the 50% borrowing mechanism that ties governance depth to market cap, establishes that:
+1. Most governance decisions are below the manipulation resistance threshold
+2. The flagship wins (META's own governance) are unrepresentative of the typical case
+3. The mechanism's own designer acknowledges the "80 IQ" scope
+
+**This is a MATERIAL scoping of Belief #1.** The theoretical mechanism is sound. The operational claim — that futarchy provides manipulation-resistant governance for MetaDAO's ecosystem — holds reliably only for established protocols with large market caps (a minority), not for early-stage ICO governance (the majority and the growth thesis).
+
+**Belief #1 does NOT collapse.** Markets still beat votes for information aggregation in the conditions where the conditions are met. The 2024 Polymarket evidence is unaffected. The mechanism is real. But the claim as applied to MetaDAO's full governance ecosystem is overstated — it accurately describes governance of META itself and understates the risk for governance of smaller ecosystem tokens.
+
+## Impact on KB
+
+**Futarchy is manipulation-resistant because attack attempts create profitable opportunities for defenders:**
+- NEEDS SCOPING — third consecutive session flagging this
+- Proposed scope qualifier (expanding on Session 4): "Futarchy manipulation resistance holds when governance market depth (typically 50% of spot liquidity via the Futarchy AMM mechanism) exceeds attacker capital; at $58K average proposal market volume, most MetaDAO ICO governance decisions operate below the threshold where this guarantee is robust"
+- This should be an enrichment, not a new claim
+
+**Futarchy solves trustless joint ownership not just better decision-making:**
+- SCOPING CONFIRMED: all three Pine-proposed design fixes for FairScale require off-chain trust; the trustless property holds only when ownership inputs are on-chain-verifiable
+
+**Belief #6 (regulatory defensibility through decentralization):**
+- WORSENED this session: CLARITY Act doesn't fix gaming preemption; Ninth Circuit is moving pro-state; no near-term legislative resolution; CFTC comment window is the only active opportunity
+
+## CLAIM CANDIDATE: Minimum Viable Pool Size for Futarchy Governance Integrity
+
+**Title:** "Futarchy governance for investment pools requires minimum viable market cap to make manipulation resistance operational, with Living Capital vehicles below ~$1M pool value operating in the FairScale implicit put option risk zone"
+
+- **Confidence:** experimental (derived from mechanism design + two data points: FairScale failure at 640K FDV, VC discount rejection success at META's scale)
+- **Status:** This is a musing-level candidate; needs a third data point (P2P.me March 26 outcome) before extraction
+- **Depends on:** P2P.me ICO result, distribution data for MetaDAO governance market volumes
+
+## Follow-up Directions
+
+### Active Threads (continue next session)
+
+- **[P2P.me ICO result — March 26]**: Will the market filter the 182x GP multiple? Pine flagged same structural risks as FairScale (high float, stretched valuation). If it passes: evidence community overrides analyst signals with growth optionality. If it fails: systematic evidence of improving ICO quality filter. Check after March 26. This is the most time-sensitive thread.
+
+- **[CFTC ANPRM comment window — April 30 deadline]**: The governance market argument needs to get into the CFTC comment record. Key argument: governance markets have legitimate hedging function (token holders hedge economic exposure through governance participation) that sports prediction markets lack. The "single individual resolution" concern (sports: referee's call) doesn't apply to corporate governance decisions. Has anyone from MetaDAO ecosystem submitted comments? This window closes April 30.
+
+- **[Ninth Circuit KalshiEx v. Nevada — operational state]**: Today's Ninth Circuit denial of stay means Nevada TRO imminent. Track whether TRO is granted and how Kalshi responds. Does the ecosystem interpret this as a threat to MetaDAO-native futarchy markets on Solana? (Answer: probably not immediately — MetaDAO is on-chain, not a DCM like Kalshi; but the precedent still matters for US users.)
+
+- **[Living Capital minimum viable pool size]**: The first Living Capital vehicle targets ~$600K — this is below my estimated threshold (~$1M) for FairScale-risk-zone governance. Before raising, the design should specify how governance will function at sub-threshold liquidity levels. Is there a veto mechanism? A time-lock? Or is the initial vehicle accepted as a "trust-building" phase where futarchy is directional but not relied upon for manipulation resistance?
+
+### Dead Ends (don't re-run these)
+
+- **[CLARITY Act express preemption for gaming]**: Confirmed does not exist. The bill preempts state securities laws only. Don't re-run this search — the legislative fix for the gaming preemption gap doesn't exist in current legislation.
+
+- **[MetaDAO protocol-level FairScale response]**: Three months post-FairScale, no protocol changes identified. March 2026 community calls (Ownership Radio March 8 + 15) covered launches, not governance design. Stop searching for this — it's not happening in the near term.
+
+- **[Blockworks, CoinDesk, The Block direct fetch]**: Still returning 403s. Dead end for fourth consecutive session.
+
+### Branching Points (one finding opened multiple directions)
+
+- **$58K average + 50% borrowing → manipulation resistance gradient**: The mechanism design gives a precise scope qualifier. Direction A: write this up as an enrichment to the manipulation resistance claim immediately. Direction B: wait for P2P.me result to see if a third data point confirms the pattern. Pursue A — the mechanism design argument is sufficient without the third data point.
+
+- **No CLARITY Act gaming preemption → CFTC ANPRM is the only active lever**: Direction A: monitor whether MetaDAO ecosystem players submit CFTC comments (passive). Direction B: advocate for comment submission through Rio's X presence (active). Pursue B — the comment window closes April 30 and the governance market argument needs to be in the record.
+
+- **"80 IQ" admission → when is futarchy insufficient?**: House's framing implies the mechanism is tuned for catastrophic decision prevention, not nuanced governance. Direction A: map the full space of MetaDAO governance decisions and categorize which are "catastrophic" (binary yes/no) vs. "complex executive" (requires nuance). Direction B: accept the framing and design Living Capital governance to complement futarchy with other mechanisms for complex decisions. Pursue B — more directly actionable for Living Capital design.

agents/rio/research-journal.md

@@ -71,7 +71,7 @@ Cross-session memory. Review after 5+ sessions for cross-session patterns.
 ## Session 2026-03-18 (Session 4)
 **Question:** How does the March 17 SEC/CFTC joint token taxonomy interact with futarchy governance tokens — and does the FairScale governance failure expose structural vulnerabilities in MetaDAO's manipulation-resistance claim?
 
-**Belief targeted:** Belief #1 (markets beat votes for information aggregation), specifically the sub-claim [[Futarchy is manipulation-resistant because attack attempts create profitable opportunities for defenders]]. This is the mechanism claim that grounds the entire MetaDAO/Living Capital thesis.
+**Belief targeted:** Belief #1 (markets beat votes for information aggregation), specifically the sub-claim Futarchy is manipulation-resistant because attack attempts create profitable opportunities for defenders. This is the mechanism claim that grounds the entire MetaDAO/Living Capital thesis.
 
 **Disconfirmation result:** FOUND — FairScale (January 2026) is the clearest documented case of futarchy manipulation resistance failing in practice. Pine Analytics case study reveals: (1) revenue misrepresentation by team was not priced in pre-launch; (2) below-NAV token created risk-free arbitrage for liquidation proposer who earned ~300%; (3) believers couldn't counter without buying above NAV; (4) all proposed fixes require off-chain trust. This is a SCOPING disconfirmation, not a full refutation — the manipulation resistance claim holds in liquid markets with verifiable inputs, but inverts in illiquid markets with off-chain fundamentals.
 
@@ -95,3 +95,38 @@ New cross-session pattern emerging: MetaDAO ecosystem is running three parallel
 **Sources archived this session:** 2 (Pine Analytics FairScale case study, Pine Analytics P2P.me ICO analysis)
 
 Note: Tweet feeds empty for fourth consecutive session. Web access continued to fail for most URLs (Blockworks 403, The Block 403/404, CoinDesk 404, CFTC ECONNREFUSED). Pine Analytics Substack remained accessible. Will continue using Pine Analytics as primary accessible source for MetaDAO ecosystem coverage.
+
+---
+
+## Session 2026-03-19 (Session 5)
+
+**Question:** Does the typical MetaDAO governance decision meet the "liquid markets with verifiable inputs" threshold that makes futarchy's manipulation resistance hold — and if thin markets are the norm, does this void the manipulation resistance claim in practice?
+
+**Belief targeted:** Belief #1 (markets beat votes for information aggregation), specifically the scope qualifier added in Session 4: "liquid markets with verifiable inputs." The target was to test whether this qualifier describes typical MetaDAO operating conditions or edge cases only.
+
+**Disconfirmation result:** MATERIAL SCOPING CONFIRMED. Three converging data points establish that the manipulation resistance threshold is NOT met in typical MetaDAO governance:
+1. **$58K average per proposal** across 65 governance decisions ($3.8M cumulative) — MetaDAO's own valuation community describes this as "signal mechanisms, not high-conviction capital allocation tools"
+2. **50% liquidity borrowing mechanism** ties governance depth to spot liquidity to token market cap — small-cap ICO tokens (the growth thesis) are structurally in the FairScale risk zone
+3. **Kollan House "80 IQ" admission** — MetaDAO's creator explicitly scoped the mechanism to catastrophic decision prevention, not complex governance
+
+The flagship evidence for manipulation resistance (VC discount rejection, 16% META surge) is survivorship-biased — it describes governance of META itself (most liquid ecosystem token), not governance of the small-cap ICOs that constitute MetaDAO's permissionless capital formation thesis.
+
+**Belief #1 does NOT collapse.** Markets beat votes in the conditions where the conditions are met. The 2024 Polymarket evidence is unaffected. But the operational claim — futarchy provides manipulation-resistant governance for MetaDAO's full ecosystem — applies reliably only to established protocols, not to the typical early-stage ICO governance decision.
+
+**Key finding:** A minimum viable pool size exists for futarchy governance integrity. The 50% liquidity borrowing mechanism means governance market depth = f(token market cap). Living Capital's first vehicle (~$600K target) would operate below the estimated ~$1M threshold where FairScale-type risk is live. The design needs to account for sub-threshold governance before the first raise.
+
+**Major external event:** Ninth Circuit denied Kalshi's administrative stay TODAY (March 19, 2026). Nevada can now pursue a TRO that could exclude Kalshi from the state within days. Combined with the Maryland Fourth Circuit ruling, the circuit split is now confirmed at the appellate level — SCOTUS review likely in 2026/2027. AND: the CLARITY Act does NOT include express preemption for state gaming laws — the legislative fix I flagged in Session 3 doesn't exist in the current bill.
+
+**Pattern update:**
+- Sessions 1-4: "Regulatory bifurcation" — federal clarity increasing while state opposition escalates
+- **Session 5 update: Pattern confirms but accelerates.** Ninth Circuit joins Fourth Circuit in the pro-state column. CLARITY Act doesn't fix the gaming preemption gap. SCOTUS is now the only resolution path. Timeline: 2027 at earliest.
+- **New pattern identified:** "Governance quality gradient" — manipulation resistance scales with token market cap. MetaDAO's mechanism design (50% borrowing) formally encodes this. The manipulation resistance claim is accurate for the top of the ecosystem (META itself) and misleading for the typical case (small-cap ICO governance).
+
+**Confidence shift:**
+- Belief #1 (markets beat votes): **NARROWED THIRD TIME** — now qualified by: (a) ordinal selection > calibrated prediction (Session 1); (b) liquid markets with verifiable inputs (Session 4); (c) "liquid" in MetaDAO context requires token market cap sufficient for ~$500K+ spot pool, which most ICO tokens lack at launch (Session 5). The mechanism is real; the operational scope is much narrower than the belief implies.
+- Belief #3 (futarchy solves trustless joint ownership): **FURTHER COMPLICATED** — "trustless" property requires on-chain verifiable inputs AND sufficient market cap for deep governance markets. Early-stage companies with off-chain revenue claims fail both conditions. The claim needs significant scope qualifiers to survive the FairScale + $58K average evidence.
+- Belief #6 (regulatory defensibility through decentralization): **WORSENED** — Ninth Circuit moving pro-state; CLARITY Act won't fix gaming preemption; no near-term legislative or regulatory resolution. The gaming classification risk has no available fix except SCOTUS, which is 1-2 years away.
+
+**Sources archived this session:** 7 (Pine Analytics P2P.me ICO analysis, Solana Compass Futarchy AMM liquidity borrowing mechanism, CoinDesk Ninth Circuit Nevada ruling, DeepWaters Capital governance volume data, WilmerHale CFTC ANPRM analysis, Pine Analytics FairScale design fixes update, CLARITY Act gaming preemption gap synthesis, MetaDAO Ownership Radio March 2026 context)
+
+Note: Tweet feeds empty for fifth consecutive session. Web access improved this session — CoinDesk policy, WilmerHale, Solana Compass, and DeepWaters Capital all accessible. Pine Analytics Substack accessible. Blockworks 403 again. The Block 403. ICM Analytics and MetaDAO Futarchy AMM (CoinGecko) returned 403.

agents/theseus/musings/research-2026-03-20.md

@@ -0,0 +1,164 @@
+---
+type: musing
+agent: theseus
+title: "EU AI Act Article 43 and the Legislative Path to Mandatory Independent AI Evaluation"
+status: developing
+created: 2026-03-20
+updated: 2026-03-20
+tags: [EU-AI-Act, Article-43, conformity-assessment, mandatory-evaluation, independent-audit, GPAI, frontier-AI, B1-disconfirmation, governance-gap, research-session]
+---
+
+# EU AI Act Article 43 and the Legislative Path to Mandatory Independent AI Evaluation
+
+Research session 2026-03-20. Tweet feed empty again — all web research.
+
+## Research Question
+
+**Does EU AI Act Article 43 create mandatory conformity assessment for frontier AI, and is there an emerging legislative pathway to mandate independent evaluation at the international level?**
+
+### Why this question (priority from previous session)
+
+Direct continuation of the 2026-03-19 NEXT flag: "Does EU AI Act Article 43 create mandatory conformity assessment for frontier AI? Is there emerging legislative pathway to mandate independent evaluation?"
+
+The 9-session arc thesis: the technical infrastructure for independent AI evaluation exists (PETs, METR, AISI tools); what's missing is:
+1. Legal mandate for independence (not voluntary-collaborative)
+2. Technical feasibility of deception-resilient evaluation (AAL-3/4)
+
+Yesterday's branching point: Direction A — look for emerging proposals to make evaluation mandatory (legislative path, EU AI Act Article 43, US state laws). This is Direction A, flagged as more tractable.
+
+### Keystone belief targeted: B1 — "AI alignment is the greatest outstanding problem for humanity and not being treated as such"
+
+Disconfirmation target (from beliefs.md): "If safety spending approaches parity with capability spending at major labs, or if governance mechanisms demonstrate they can keep pace with capability advances."
+
+Specific disconfirmation test for this session: Does EU AI Act Article 43 require genuinely independent conformity assessment for general-purpose AI / frontier models? If yes, and if enforcement is on track for August 2026, this would be the strongest evidence yet that governance can scale to the problem.
+
+The disconfirmation I'm searching for: A binding, mandatory, independent evaluation requirement for frontier AI systems that doesn't depend on lab cooperation — the regulatory equivalent of FDA clinical trials.
+
+---
+
+## Key Findings
+
+### Finding 1: EU AI Act creates MANDATORY obligations AND compulsory evaluation powers — but enforcement is reactive, not proactive
+
+The EU AI Act is more powerful than the voluntary-collaborative model I've been characterizing. Key architecture:
+
+- **Article 51**: 10^25 FLOP threshold for GPAI systemic risk — captures GPT-4 class and above
+- **Article 55**: MANDATORY obligations for systemic-risk GPAI including adversarial testing and risk assessment — not voluntary
+- **Article 92**: **COMPULSORY** evaluation powers — AI Office can appoint independent experts, compel API/source code access, order compliance under penalty of fines. This is not METR-style "invitation to evaluate."
+- **Article 101**: Real fines — 3% global annual turnover or €15M whichever is higher
+
+BUT: enforcement is **reactive, not proactive**. Article 92 triggers when (a) documentation is insufficient OR (b) scientific panel issues qualified alert. GPAI models can be deployed while the AI Office monitors; evaluation is not a condition of deployment. This is SEC enforcement structure (investigate when problems emerge), not FDA pre-market approval.
+
+**Article 43 (conformity assessment for high-risk AI)** is mostly self-assessment — third-party notified body only required when harmonized standards don't exist, which is the exception. Article 43 ≠ FDA model.
+
+### Finding 2: Benchmarks provide ZERO coverage of loss-of-control capabilities
+
+Bench-2-CoP (arXiv:2508.05464, August 2025) analyzed 195,000 benchmark questions against EU AI Act compliance taxonomy:
+- "Tendency to hallucinate": 61.6% coverage
+- "Lack of performance reliability": 31.2% coverage
+- **Capabilities for oversight evasion, self-replication, autonomous AI development: ZERO coverage**
+
+Conclusion: "Current public benchmarks are insufficient, on their own, for providing the evidence of comprehensive risk assessment required for regulatory compliance." Independent targeted evaluation tools designed for regulatory requirements are necessary but don't yet exist.
+
+### Finding 3: Frontier safety frameworks score 8-35% against safety-critical industry standards
+
+Stelling et al. (arXiv:2512.01166, December 2025) evaluated twelve frontier safety frameworks published post-Seoul Summit using 65 safety-critical industry criteria:
+- Scores range from **8% to 35%** — "disappointing"
+- Maximum achievable by combining best practices across ALL frameworks: **52%**
+- Universal deficiencies: no quantitative risk tolerances, no capability pause thresholds, inadequate unknown risk identification
+
+Critical structural finding: Both the EU AI Act's Code of Practice AND California's Transparency in Frontier Artificial Intelligence Act **rely on these same 8-35% frameworks as compliance evidence**. The governance architecture accepts as compliance evidence what safety-critical industry criteria score at 8-35%.
+
+### Finding 4: Article 43 conformity assessment ≠ FDA for GPAI
+
+Common misreading: EU AI Act has "conformity assessment" therefore it has FDA-like independent evaluation. Actually: (1) Article 43 governs HIGH-RISK AI (use-case classification), not GPAI (compute-scale classification); (2) For most high-risk AI, self-assessment is permitted; (3) GPAI systemic risk models face a SEPARATE regime under Articles 51-56 with flexible compliance pathways. The path to independent evaluation in EU AI Act is Article 92 (reactive compulsion), not Article 43 (conformity).
+
+### Finding 5: Anthropic RSP v3.0 weakened unconditional binary thresholds to conditional escape clauses
+
+RSP v3.0 (February 24, 2026) replaced:
+- Original: "Never train without advance safety guarantees" (unconditional)
+- New: "Only pause if Anthropic leads AND catastrophic risks are significant" (conditional dual-threshold)
+
+METR's Chris Painter: "frog-boiling" effect from removing binary thresholds. RSP v3.0 emphasizes Anthropic's own internal assessments; no mandatory third-party evaluations specified. Financial context: $30B raised at ~$380B valuation.
+
+The "Anthropic leads" condition creates a competitive escape hatch: if competitors advance, the safety commitment is suspended. This transforms a categorical safety floor into a business judgment.
+
+### Finding 6: EU Digital Simplification Package (November 2025) — unknown specific impact
+
+Commission proposed targeted amendments to AI Act via Digital Simplification Package on November 19, 2025 — within 3.5 months of GPAI obligations taking effect (August 2025). Specific provisions targeted could not be confirmed. Pattern concern: regulatory implementation triggers deregulatory pressure.
+
+### Synthesis: Two Independent Dimensions of Governance Inadequacy
+
+Previous sessions identified: structural inadequacy (voluntary-collaborative not independent). This session adds a second dimension: **substantive inadequacy** (compliance evidence quality is 8-35% of safety-critical standards). These are independent failures:
+
+1. **Structural inadequacy**: Governance mechanisms are voluntary or reactive, not mandatory pre-deployment and independent (per Brundage et al. AAL framework)
+2. **Content inadequacy**: The frameworks accepted as compliance evidence score 8-35% against established safety management criteria (per Stelling et al.)
+
+EU AI Act's Article 55 + Article 92 partially addresses structural inadequacy (mandatory obligations + compulsory reactive enforcement). But the content inadequacy persists independently — even with compulsory evaluation powers, what's being evaluated against (frontier safety frameworks, benchmarks without loss-of-control coverage) is itself inadequate.
+
+### B1 Disconfirmation Assessment
+
+B1 states: "not being treated as such." Previous sessions showed: voluntary-collaborative only. This session: EU AI Act adds mandatory + compulsory enforcement layer.
+
+**Net assessment (updated):** B1 holds, but must be more precisely characterized:
+- The response is REAL: EU AI Act creates genuine mandatory obligations and compulsory enforcement powers
+- The response is INADEQUATE: reactive not proactive; compliance evidence quality at 8-35% of safety-critical standards; Digital Simplification pressure; RSP conditional erosion
+- Better framing: "Being treated with insufficient structural and substantive seriousness — governance mechanisms are mandatory but reactive, and the compliance evidence base scores 8-35% of safety-critical industry standards"
+
+---
+
+## Connection to Open Questions in KB
+
+The _map.md notes: [[voluntary safety pledges cannot survive competitive pressure because unilateral commitments are structurally punished when competitors advance without equivalent constraints]] — EU AI Act's Article 55 mandatory obligations don't share this weakness, but Article 92's reactive enforcement and flexible compliance pathways partially reintroduce it.
+
+Also: The double-inadequacy finding (structural + content) extends the frontier identified in previous sessions. The missing third-party independent measurement infrastructure is not just structurally absent — it's substantively inadequate even where it exists.
+
+## Potential New Claim Candidates
+
+CLAIM CANDIDATE: "EU AI Act creates the first binding mandatory obligations for frontier GPAI models globally, but enforcement is reactive not proactive — Article 92 compulsory evaluation requires a trigger (qualified alert or insufficient documentation), not pre-deployment approval, making it SEC-enforcement rather than FDA-pre-approval" — high confidence, specific, well-grounded.
+
+CLAIM CANDIDATE: "Frontier AI safety frameworks published post-Seoul Summit score 8-35% against established safety-critical industry risk management criteria, with the composite maximum at 52%, quantifying the structural inadequacy of current voluntary safety governance" — very strong, from arXiv:2512.01166, directly extends B1.
+
+CLAIM CANDIDATE: "Anthropic RSP v3.0 replaces unconditional binary safety thresholds with dual-condition competitive escape clauses — safety pause only required if both Anthropic leads the field AND catastrophic risks are significant — transforming a categorical safety floor into a business judgment" — specific, dateable, well-grounded.
+
+CLAIM CANDIDATE: "Current AI benchmarks provide zero coverage of capabilities central to loss-of-control scenarios including oversight evasion and self-replication, making them insufficient for EU AI Act Article 55 compliance despite being the primary compliance evidence submitted" — from arXiv:2508.05464, specific and striking.
+
+## Sources Archived This Session
+
+1. **EU AI Act GPAI Framework (Articles 51-56, 88-93, 101)** (HIGH) — compulsory evaluation powers, reactive enforcement, 10^25 FLOP threshold, 3% fines
+2. **Bench-2-CoP (arXiv:2508.05464)** (HIGH) — zero benchmark coverage of loss-of-control capabilities
+3. **Stelling et al. GPAI CoP industry mapping (arXiv:2504.15181)** (HIGH) — voluntary compliance precedent mapping
+4. **Stelling et al. Frontier Safety Framework evaluation (arXiv:2512.01166)** (HIGH) — 8-35% scores against safety-critical standards
+5. **Anthropic RSP v3.0** (HIGH) — conditional thresholds replacing binary floors
+6. **EU AI Act Article 43 conformity limits** (MEDIUM) — corrects Article 43 ≠ FDA misreading
+7. **EU Digital Simplification Package Nov 2025** (MEDIUM) — 3.5-month deregulatory pressure after mandatory obligations
+
+Total: 7 sources (5 high, 2 medium)
+
+---
+
+## Follow-up Directions
+
+### Active Threads (continue next session)
+
+- **Digital Simplification Package specifics**: The November 2025 amendments are documented but content not accessible. Next session: search specifically "EU AI Act omnibus simplification Article 53 Article 55" and European Parliament response. If these amendments weaken Article 55 adversarial testing requirements or Article 92 enforcement powers, B1 strengthens significantly.
+
+- **AI Office first enforcement year**: What has the AI Office actually done since August 2025? Has it used Article 92 compulsory evaluation powers? Opened any investigations? Issued any corrective actions? The absence of enforcement data after 7+ months is itself an informative signal — absence of action is a data point. Search: "AI Office investigation GPAI 2025 2026" "EU AI Office enforcement action frontier AI"
+
+- **California Transparency in Frontier AI Act specifics**: Stelling et al. (2512.01166) confirms it's a real law relying on frontier safety frameworks as compliance evidence. What exactly does it require? Is it transparency-only or does it create independent evaluation obligations? Does it strengthen or merely document the 8-35% compliance evidence problem? Search: "California AB 2013 frontier AI transparency requirements" + "what frontier safety frameworks must disclose."
+
+- **Content gap research**: Who is building the independent evaluation tools that Bench-2-CoP says are necessary? Is METR or AISI developing benchmarks for oversight-evasion and self-replication capabilities? If not, who will? This is the constructive question this session opened.
+
+### Dead Ends (don't re-run)
+
+- arXiv search with terms including years (2025, 2026) — arXiv's search returns "no results" for most multi-word queries including years; use shorter, more general terms
+- euractiv.com, politico.eu — blocked by Claude Code
+- Most .eu government sites (eur-lex.europa.eu, ec.europa.eu press corner) — returns CSS/JavaScript not content
+- Most .gov.uk sites — 404 for specific policy pages
+- OECD.org, Brookings — 403 Forbidden
+
+### Branching Points (one finding opened multiple directions)
+
+- **The double-inadequacy finding**: Direction A — structural fix (make enforcement proactive/pre-deployment like FDA). Direction B — content fix (build evaluation tools that actually cover loss-of-control capabilities). Both necessary, but Direction B is more tractable and less politically contentious. Direction B also has identifiable actors (METR, AISI, academic researchers building new evals) who could do this work. Pursue Direction B first — more actionable and better suited to Theseus's KB contribution.
+
+- **RSP v3.0 conditional escape clause**: Direction A — track whether other labs weaken their frameworks similarly (OpenAI, DeepMind analogous policy evolution). Direction B — look for any proposals that create governance frameworks resilient to this pattern (mandatory unconditional floors in regulation rather than voluntary commitments). Direction B connects to the EU AI Act Article 55 thread and is higher value.

agents/theseus/research-journal.md

@@ -189,7 +189,7 @@ NEW PATTERN:
 STRENGTHENED:
 - B1 (alignment not being treated as such) — holds. Mechanisms exist but are mismatched in scale to the severity of the problem. The DoD/Anthropic confrontation is a concrete case of government functioning as coordination-BREAKER.
 - B2 (alignment is a coordination problem) — automation overshoot correction is also a coordination failure. The four mechanisms require coordination across firms/regulators to function; firms acting individually cannot correct for competitive pressure.
-- "Government as coordination-breaker" — updated with DoD/Anthropic episode. This is a stronger confirmation of the [[government designation of safety-conscious AI labs as supply chain risks]] claim.
+- "Government as coordination-breaker" — updated with DoD/Anthropic episode. This is a stronger confirmation of the government designation of safety-conscious AI labs as supply chain risks claim.
 
 COMPLICATED:
 - The measurement dependency insight complicates all constructive alternatives. Even if we build collective intelligence infrastructure (B5), it needs accurate performance signals to self-correct. The perception gap at the organizational level is a precursor problem that the constructive case hasn't addressed.
@@ -239,3 +239,28 @@ NEW PATTERN:
 **Sources archived:** 6 sources (4 high, 2 medium). Key: Brundage et al. AAL framework (arXiv:2601.11699), Kim et al. CMU assurance framework (arXiv:2601.22424), Uuk et al. 76-expert study (arXiv:2412.02145), Beers & Toner PET scrutiny (arXiv:2502.05219), STREAM standard (arXiv:2508.09853), METR/AISI practice synthesis.
 
 **Cross-session pattern (8 sessions):** Active inference → alignment gap → constructive mechanisms → mechanism engineering → [gap] → overshoot mechanisms → correction mechanism failures → evaluation infrastructure limits. The full arc: WHAT architecture → WHERE field is → HOW mechanisms work → BUT ALSO they fail → WHY they overshoot → HOW correction fails → WHAT the missing infrastructure looks like → WHERE the legal mandate gap is. Thesis now highly specific: the technical infrastructure for independent AI evaluation exists (PETs, METR, AISI tools); what's missing is legal mandate for independence (not voluntary-collaborative) and the technical feasibility of deception-resilient evaluation (AAL-3/4). Next: Does EU AI Act Article 43 create mandatory conformity assessment for frontier AI? Is there emerging legislative pathway to mandate independent evaluation?
+
+## Session 2026-03-20 (EU AI Act GPAI Enforcement Architecture)
+
+**Question:** Does EU AI Act Article 43 create mandatory conformity assessment for frontier AI, and is there an emerging legislative pathway to mandate independent evaluation?
+
+**Belief targeted:** B1 (keystone) — "AI alignment is the greatest outstanding problem for humanity and not being treated as such." Specific disconfirmation target: do governance mechanisms demonstrate they can keep pace with capability advances?
+
+**Disconfirmation result:** Partial disconfirmation with important structural update. The EU AI Act is MORE powerful than the voluntary-collaborative characterization from previous sessions: Article 55 creates MANDATORY obligations for systemic-risk GPAI (10^25 FLOP threshold), Article 92 creates COMPULSORY evaluation powers (AI Office can appoint independent experts, compel API/source code access, issue binding orders under 3% global turnover fines). This is qualitatively different from METR's voluntary-collaborative model. BUT: enforcement is reactive not proactive — triggered by qualified alerts or compliance failures, not required as a pre-deployment condition. And the content quality of what's accepted as compliance evidence is itself inadequate: frontier safety frameworks score 8-35% against safety-critical industry criteria (Stelling et al. arXiv:2512.01166). Two independent dimensions of inadequacy: structural (reactive not proactive) and substantive (8-35% quality compliance evidence). B1 holds.
+
+**Key finding:** Double-inadequacy in governance architecture. Structural: EU AI Act enforcement is reactive (SEC model) not proactive (FDA model). Substantive: the compliance evidence base — frontier safety frameworks — scores 8-35% against safety-critical industry standards, with a composite maximum of 52%. Both the EU AI Act CoP AND California's Transparency in Frontier AI Act accept these same frameworks as compliance evidence. The governance architecture is built on foundations that independently fail safety-critical standards.
+
+**Pattern update:**
+- STRENGTHENED: B1 ("not being treated as such") — now with two independent dimensions of inadequacy instead of one. The substantive content inadequacy (8-35% safety framework quality) is independent of the structural inadequacy (reactive enforcement)
+- COMPLICATED: The characterization of "voluntary-collaborative" was too simple. EU AI Act creates mandatory obligations + compulsory enforcement. Better framing: "Mandatory obligations with reactive enforcement and inadequate compliance evidence quality" — more specific than "voluntary-collaborative"
+- NEW: Article 43 ≠ FDA model — conformity assessment for high-risk AI is primarily self-assessment; independent evaluation runs through Article 92, not Article 43. Many policy discussions conflate these
+- NEW: Anthropic RSP v3.0 introduces conditional escape clauses — "only pause if Anthropic leads AND catastrophic risks are significant" — transforming unconditional binary safety floors into competitive business judgments
+- NEW: Benchmarks provide ZERO coverage of oversight-evasion, self-replication, autonomous AI development despite these being the highest-priority compliance needs
+
+**Confidence shift:**
+- "Governance infrastructure is voluntary-collaborative" → UPDATED: better framing is "governance is mandatory with reactive enforcement but inadequate compliance evidence quality" — more precise, reflects EU AI Act's mandatory Article 55 + compulsory Article 92
+- "Technical infrastructure for independent evaluation exists (PETs, METR, AISI)" → COMPLICATED: the evaluation tools that exist (benchmarks) score 0% on loss-of-control capabilities; tools for regulatory compliance don't yet exist
+- "Voluntary safety pledges collapse under competitive pressure" → UPDATED: RSP v3.0 is the clearest case yet — conditional thresholds are structurally equivalent to voluntary commitments that depend on competitive context
+- "Frontier safety frameworks are inadequate" → QUANTIFIED: 8-35% range, 52% composite maximum — moved from assertion to empirically measured
+
+**Cross-session pattern (9 sessions):** Active inference → alignment gap → constructive mechanisms → mechanism engineering → [gap] → overshoot mechanisms → correction failures → evaluation infrastructure limits → mandatory governance with reactive enforcement and inadequate evidence quality. The emerging thesis has gained its final structural piece: it's not just that governance is voluntary-collaborative (structural inadequacy), it's that what governance accepts as compliance evidence scores 8-35% of safety-critical standards (substantive inadequacy). Two independent failures explaining why even "mandatory" frameworks fall short. Next: Digital Simplification Package specific provisions; AI Office first enforcement actions; building the constructive alternative (what would adequate compliance evidence look like?).

domains/ai-alignment/AI transparency is declining not improving because Stanford FMTI scores dropped 17 points in one year while frontier labs dissolved safety teams and removed safety language from mission statements.md

@@ -31,22 +31,28 @@ The alignment implication: transparency is a prerequisite for external oversight
 
 
 ### Additional Evidence (extend)
-*Source: [[2024-12-00-uuk-mitigations-gpai-systemic-risks-76-experts]] | Added: 2026-03-19*
+*Source: 2024-12-00-uuk-mitigations-gpai-systemic-risks-76-experts | Added: 2026-03-19*
 
 Expert consensus identifies 'external scrutiny, proactive evaluation and transparency' as the key principles for mitigating AI systemic risks, with third-party audits as the top-3 implementation priority. The transparency decline documented by Stanford FMTI is moving in the opposite direction from what 76 cross-domain experts identify as necessary.
 
 
 ### Additional Evidence (extend)
-*Source: [[2025-08-00-mccaslin-stream-chembio-evaluation-reporting]] | Added: 2026-03-19*
+*Source: 2025-08-00-mccaslin-stream-chembio-evaluation-reporting | Added: 2026-03-19*
 
 STREAM proposal identifies that current model reports lack 'sufficient detail to enable meaningful independent assessment' of dangerous capability evaluations. The need for a standardized reporting framework confirms that transparency problems extend beyond general disclosure (FMTI scores) to the specific domain of dangerous capability evaluation where external verification is currently impossible.
 
 
 ### Additional Evidence (confirm)
-*Source: [[2026-03-16-theseus-ai-coordination-governance-evidence]] | Added: 2026-03-19*
+*Source: 2026-03-16-theseus-ai-coordination-governance-evidence | Added: 2026-03-19*
 
 Stanford FMTI 2024→2025 data: mean transparency score declined 17 points. Meta -29 points, Mistral -37 points, OpenAI -14 points. OpenAI removed 'safely' from mission statement (Nov 2025), dissolved Superalignment team (May 2024) and Mission Alignment team (Feb 2026). Google accused by 60 UK lawmakers of violating Seoul commitments with Gemini 2.5 Pro (Apr 2025).
 
+
+### Additional Evidence (extend)
+*Source: [[2026-03-20-bench2cop-benchmarks-insufficient-compliance]] | Added: 2026-03-20*
+
+The Bench-2-CoP analysis reveals that even when labs do conduct evaluations, the benchmark infrastructure itself is architecturally incapable of measuring loss-of-control risks. This compounds the transparency decline: labs are not just hiding information, they're using evaluation tools that cannot detect the most critical failure modes even if applied honestly.
+
 ---
 
 Relevant Notes:

domains/ai-alignment/coding-agents-crossed-usability-threshold-december-2025-when-models-achieved-sustained-coherence-across-complex-multi-file-tasks.md

@@ -10,6 +10,12 @@ enrichments:
   - "as AI-automated software development becomes certain the bottleneck shifts from building capacity to knowing what to build making structured knowledge graphs the critical input to autonomous systems.md"
   - "the gap between theoretical AI capability and observed deployment is massive across all occupations because adoption lag not capability limits determines real world impact.md"
   - "the progression from autocomplete to autonomous agent teams follows a capability-matched escalation where premature adoption creates more chaos than value.md"
+
+### Additional Evidence (confirm)
+*Source: [[2026-02-13-noahopinion-smartest-thing-on-earth]] | Added: 2026-03-19*
+
+Smith's observation that 'vibe coding' is now the dominant paradigm confirms that coding agents crossed from experimental to production-ready status, with the transition happening rapidly enough to be culturally notable by Feb 2026.
+
 ---
 
 # Coding agents crossed usability threshold in December 2025 when models achieved sustained coherence across complex multi-file tasks

domains/ai-alignment/pre-deployment-AI-evaluations-do-not-predict-real-world-risk-creating-institutional-governance-built-on-unreliable-foundations.md

@@ -52,10 +52,16 @@ METR and UK AISI evaluations as of March 2026 focus primarily on sabotage risk a
 
 
 ### Additional Evidence (confirm)
-*Source: [[2026-02-23-shapira-agents-of-chaos]] | Added: 2026-03-19*
+*Source: 2026-02-23-shapira-agents-of-chaos | Added: 2026-03-19*
 
 Agents of Chaos demonstrates that static single-agent benchmarks fail to capture vulnerabilities that emerge in realistic multi-agent deployment. The study's central argument is that pre-deployment evaluations are insufficient because they cannot test for cross-agent propagation, identity spoofing, and unauthorized compliance patterns that only manifest in multi-party environments with persistent state.
 
+
+### Additional Evidence (extend)
+*Source: [[2026-03-20-bench2cop-benchmarks-insufficient-compliance]] | Added: 2026-03-20*
+
+Prandi et al. (2025) found that 195,000 benchmark questions provided zero coverage of oversight evasion, self-replication, and autonomous AI development capabilities. This extends the evaluation unreliability thesis by showing the gap is not just predictive validity but complete absence of measurement for alignment-critical capabilities.
+
 ---
 
 Relevant Notes:

domains/internet-finance/MetaDAO is the futarchy launchpad on Solana where projects raise capital through unruggable ICOs governed by conditional markets creating the first platform for ownership coins at scale.md

@@ -133,6 +133,12 @@ First MetaDAO ICO failure occurred February 7, 2026 when Hurupay (onchain neoban
 
 Revenue declined sharply since mid-December 2025, with the ICO cadence problem persisting due to the curated model limiting throughput. This is the key new signal — the platform's revenue trajectory has inverted despite strong cumulative metrics, suggesting the curated model's throughput ceiling may be binding.
 
+
+### Additional Evidence (extend)
+*Source: [[2026-03-19-metadao-ownership-radio-march-2026]] | Added: 2026-03-19*
+
+MetaDAO hosted two Ownership Radio community calls in March 2026 (March 8 and March 15) focused on ecosystem updates, Futardio launches, and upcoming ICOs like P2P.me (March 26), but neither session addressed protocol-level changes or the FairScale implicit put option problem from January 2026. This suggests MetaDAO's community communication prioritizes new launches over governance mechanism reflection.
+
 ---
 
 Relevant Notes:

domains/internet-finance/futarchy adoption faces friction from token price psychology proposal complexity and liquidity requirements.md

@@ -58,6 +58,12 @@ MetaDAO's Q3 roadmap explicitly prioritized UI performance improvements, targeti
 
 The 'Do NOT TRADE' instruction on a testing proposal demonstrates operational complexity friction in futarchy systems. Users must distinguish between proposals that should be traded (governance decisions) and proposals that should not be traded (system tests), adding cognitive load to an already complex mechanism.
 
+
+### Additional Evidence (extend)
+*Source: [[2026-03-19-metadao-ownership-radio-march-2026]] | Added: 2026-03-19*
+
+The absence of FairScale design discussion in two March 2026 MetaDAO community calls, despite the January 2026 FairScale failure revealing an implicit put option problem, indicates that futarchy adoption friction includes organizational reluctance to publicly address mechanism failures even when they reveal important design limitations.
+
 ---
 
 Relevant Notes:

domains/internet-finance/futarchy-governed entities are structurally not securities because prediction market participation replaces the concentrated promoter effort that the Howey test requires.md

@@ -100,6 +100,12 @@ Better Markets' analysis of the CEA's gaming prohibition reveals that the 'legit
 
 Better Markets' gaming prohibition argument reveals a complementary legal defense for futarchy: the 'legitimate commercial purpose' test. While the Howey securities analysis focuses on whether there are 'efforts of others,' the CEA gaming prohibition focuses on whether the contract serves a genuine hedging or commercial function. Futarchy governance markets may satisfy both tests simultaneously—they lack concentrated promoter effort (Howey) AND they serve legitimate corporate governance functions (CEA commercial purpose exception). This dual defense is stronger than either alone.
 
+
+### Additional Evidence (challenge)
+*Source: [[2026-03-19-wilmerhale-cftc-anprm-analysis]] | Added: 2026-03-19*
+
+The CFTC's March 2026 ANPRM on prediction markets contains 40 questions focused entirely on sports/entertainment event contracts and DCM (Designated Contract Market) regulation, with zero questions about governance markets, DAO decision markets, or futarchy applications. This regulatory silence means futarchy governance mechanisms exist in an unaddressed gap: they are neither explicitly enabled by the CFTC framework (which focuses on centralized exchanges) nor restricted by it. The comment deadline of approximately April 30, 2026 represents the only near-term opportunity to proactively define the governance market category before the ANPRM process closes. WilmerHale's legal analysis, reflecting institutional legal guidance, does not mention governance/DAO/futarchy distinctions at all, suggesting the legal industry has not yet mapped this application. This creates a dual risk: (1) futarchy governance markets lack the safe harbor that DCM-regulated prediction markets may receive, and (2) the gaming classification vector that states are pursuing remains unaddressed at the federal level.
+
 ---
 
 Relevant Notes:

domains/internet-finance/polymarket-achieved-us-regulatory-legitimacy-through-qcx-acquisition-establishing-prediction-markets-as-cftc-regulated-derivatives.md

@@ -60,16 +60,22 @@ The Kalshi litigation reveals that CFTC regulation alone does not resolve state
 
 
 ### Additional Evidence (challenge)
-*Source: [[2026-02-00-better-markets-prediction-markets-gambling]] | Added: 2026-03-18*
+*Source: 2026-02-00-better-markets-prediction-markets-gambling | Added: 2026-03-18*
 
 Better Markets presents the strongest counter-argument to CFTC exclusive jurisdiction: the CEA already prohibits gaming contracts under Section 5c(c)(5)(C), and sports prediction markets ARE gaming by any reasonable definition. Kalshi's own prior admission that 'Congress did not want sports betting conducted on derivatives markets' undermines the current industry position. This suggests Polymarket's regulatory legitimacy may be more fragile than assumed—state AGs have a statutory basis to challenge CFTC jurisdiction, not just a turf war.
 
 
 ### Additional Evidence (challenge)
-*Source: [[2026-02-00-better-markets-prediction-markets-gambling]] | Added: 2026-03-19*
+*Source: 2026-02-00-better-markets-prediction-markets-gambling | Added: 2026-03-19*
 
 Better Markets argues that CFTC jurisdiction over prediction markets is legally unsound because the CEA Section 5c(c)(5)(C) already prohibits gaming contracts, and sports/entertainment prediction markets are gaming by definition. They cite Senator Blanche Lincoln's legislative intent that the CEA was NOT meant to 'enable gambling through supposed event contracts' and specifically named sports events. Most damaging: Kalshi's own prior admission that 'Congress did not want sports betting conducted on derivatives markets' when defending election contracts, which undermines the current CFTC jurisdiction claim.
 
+
+### Additional Evidence (challenge)
+*Source: [[2026-03-19-coindesk-ninth-circuit-nevada-kalshi]] | Added: 2026-03-19*
+
+Ninth Circuit denied Kalshi's motion for administrative stay on March 19, 2026, allowing Nevada to proceed with temporary restraining order that would exclude Kalshi from the state entirely. This demonstrates that CFTC regulation does not preempt state gaming law enforcement, contradicting the assumption that CFTC-regulated status provides comprehensive regulatory legitimacy. Fourth Circuit (Maryland) and Ninth Circuit (Nevada) both now allow state enforcement while Third Circuit (New Jersey) ruled for federal preemption, creating a circuit split that undermines any claim of settled regulatory legitimacy.
+
 ---
 
 Relevant Notes:

domains/internet-finance/polymarket-kalshi-duopoly-emerging-as-dominant-us-prediction-market-structure-with-complementary-regulatory-models.md

@@ -34,10 +34,16 @@ The duopoly thesis assumes regulatory barriers remain high. If CFTC streamlines
 
 
 ### Additional Evidence (extend)
-*Source: [[2026-01-30-npr-kalshi-19-federal-lawsuits]] | Added: 2026-03-18*
+*Source: 2026-01-30-npr-kalshi-19-federal-lawsuits | Added: 2026-03-18*
 
 Kalshi litigation outcome affects competitors Robinhood, Coinbase, FanDuel, and DraftKings, all of which recently announced rival prediction market services. A Kalshi loss could shut down the entire US prediction market industry beyond Polymarket's offshore model, while a Kalshi victory establishes federal preemption precedent reshaping sports betting regulation nationally.
 
+
+### Additional Evidence (challenge)
+*Source: [[2026-03-19-coindesk-ninth-circuit-nevada-kalshi]] | Added: 2026-03-19*
+
+The emerging circuit split (Fourth and Ninth Circuits pro-state, Third Circuit pro-federal) creates operational exclusion zones for prediction markets regardless of CFTC registration. Nevada can now exclude Kalshi for at least two weeks pending preliminary injunction hearing, and Arizona filed first criminal charges against Kalshi on March 17, 2026. This state-by-state enforcement pattern fragments the market rather than enabling a stable duopoly structure, as platforms face different legal treatment across jurisdictions.
+
 ---
 
 Relevant Notes:

domains/internet-finance/the SECs treatment of staking rewards as service payments establishes that mechanical participation in network consensus is not an investment contract.md

@@ -21,6 +21,12 @@ This precedent has direct implications for futarchy governance mechanisms:
 
 3. **Third-party delegation as the boundary.** The staking distinction (self-staking vs pool delegation) maps onto futarchy (direct market participation vs delegated governance). Direct prediction market trading should qualify as mechanical participation; a fund that trades conditional tokens on behalf of passive investors may cross into investment contract territory.
 
+
+### Additional Evidence (extend)
+*Source: [[2026-03-19-wilmerhale-cftc-anprm-analysis]] | Added: 2026-03-19*
+
+The CFTC ANPRM's focus on 'contracts resolving based on the action of a single individual or small group' for heightened scrutiny is framed in the sports context (referee calls, athlete performance), not governance markets. This suggests a potential argument for governance markets: if prediction market participation in futarchy is mechanical trading activity (like staking) rather than reliance on a promoter's efforts, it may parallel the SEC's staking framework. However, the ANPRM's complete silence on this application means the argument has not been tested or acknowledged by regulators.
+
 ---
 
 Relevant Notes:

entities/ai-alignment/anthropic.md

@@ -54,6 +54,8 @@ Frontier AI safety laboratory founded by former OpenAI VP of Research Dario Amod
 - **2026-03** — Claude Code achieved 54% enterprise coding market share, $2.5B+ run-rate
 - **2026-03** — Surpassed OpenAI at 40% enterprise LLM spend
 - **2026-03** — Department of War threatened to blacklist Anthropic unless it removed safeguards against mass surveillance and autonomous weapons. Anthropic refused publicly and faced Pentagon retaliation.
+- **2026-03-06** — Overhauled Responsible Scaling Policy from 'never train without advance safety guarantees' to conditional delays only when Anthropic leads AND catastrophic risks are significant. Raised $30B at ~$380B valuation with 10x annual revenue growth. Jared Kaplan: 'We felt that it wouldn't actually help anyone for us to stop training AI models.'
+- **2026-02-24** — Released RSP v3.0, replacing unconditional binary safety thresholds with dual-condition escape clauses (pause only if Anthropic leads AND risks are catastrophic). METR partner Chris Painter warned of 'frog-boiling effect' from removing binary thresholds. Raised $30B at ~$380B valuation with 10x annual revenue growth.
 ## Competitive Position
 Strongest position in enterprise AI and coding. Revenue growth (10x YoY) outpaces all competitors. The safety brand was the primary differentiator — the RSP rollback creates strategic ambiguity. CEO publicly uncomfortable with power concentration while racing to concentrate it.
 

entities/internet-finance/fairscale.md

@@ -29,6 +29,7 @@ FairScale was a Solana-based reputation infrastructure project that raised ~$355
 - **2026-02** — Liquidation proposer earned ~300% return
 
 - **2026-02** — [[fairscale-liquidation-proposal]] Passed: 100% treasury liquidation authorized based on revenue misrepresentation; proposer earned ~300% return
+- **2026-02-15** — Pine Analytics publishes post-mortem analysis documenting that all three proposed design fixes (milestone verification, dispute resolution, contributor whitelisting) reintroduce off-chain trust assumptions
 ## Revenue Misrepresentation Details
 
 - **TigerPay:** Claimed ~17K euros/month → community verification found no payment arrangement

entities/internet-finance/kalshi.md

@@ -51,6 +51,7 @@ CFTC-designated contract market for event-based trading. USD-denominated, KYC-re
 - **2026-01-09** — Tennessee Middle District Court ruled in favor of Kalshi in KalshiEx v. Orgel, finding impossibility of dual compliance and obstacle to federal objectives, creating circuit split with Maryland
 - **2026-03-17** — Arizona AG filed 20 criminal counts including illegal gambling and election wagering — first-ever criminal charges against a US prediction market platform
 - **2026-01-09** — Tennessee court ruled in favor of Kalshi in KalshiEx v. Orgel, finding impossibility of dual compliance and obstacle to federal objectives, creating circuit split with Maryland
+- **2026-03-19** — Ninth Circuit denied administrative stay motion, allowing Nevada to proceed with temporary restraining order that would exclude Kalshi from Nevada for at least two weeks pending preliminary injunction hearing
 ## Competitive Position
 - **Regulation-first**: Only CFTC-designated prediction market exchange. Institutional credibility.
 - **vs Polymarket**: Different market — Kalshi targets mainstream/institutional users who won't touch crypto. Polymarket targets crypto-native users who want permissionless market creation. Both grew massively post-2024 election.

entities/internet-finance/metadao.md

@@ -76,6 +76,12 @@ The futarchy governance protocol on Solana. Implements decision markets through
 - **2026-02-07** — [[metadao-hurupay-ico]] Failed: First MetaDAO ICO failure - Hurupay failed to reach $3M minimum, full refunds issued
 - **2026-03** — [[metadao-vc-discount-rejection]] Passed: Community rejected $6M OTC deal offering 30% VC discount via futarchy vote, triggering 16% META price surge
 - **2026-03-17** — Revenue decline continues since mid-December 2025; platform generated ~$2.4M total revenue since Futarchy AMM launch (60% AMM, 40% Meteora LP)
+- **2026-01-15** — DeepWaters Capital analysis reveals $3.8M cumulative trading volume across 65 governance proposals ($58K average per proposal), with platform AMM processing $300M volume and generating $1.5M in fees
+- **2026-03-08** — Ownership Radio #1 community call covering MetaDAO ecosystem, Futardio, and futarchy governance mechanisms
+- **2026-03-15** — Ownership Radio community call on ownership coins and new Futardio launches
+- **2026-02-15** — Pine Analytics documents absence of MetaDAO protocol-level response to FairScale implicit put option problem two months after January 2026 failure, with P2P.me launching March 26 using same governance structure
+- **2026-03-26** — [[metadao-p2p-me-ico]] Active: P2P.me ICO vote scheduled, testing futarchy quality filter on stretched valuation (182x gross profit multiple)
+- **2026-02-01** — Kollan House explains 50% spot liquidity borrowing mechanism in Solana Compass interview, revealing governance market depth scales with token market cap
 ## Key Decisions
 | Date | Proposal | Proposer | Category | Outcome |
 |------|----------|----------|----------|---------|

inbox/archive/2026-01-01-futardio-launch-nex-id.md

@@ -0,0 +1,246 @@
+---
+type: source
+title: "Futardio: Nex ID fundraise goes live"
+author: "futard.io"
+url: "https://www.futard.io/launch/Cs1tWSwarGDXFBTZaFE4b13Npx9PnjSsgEjRmGAZvQU6"
+date: 2026-01-01
+domain: internet-finance
+format: data
+status: unprocessed
+tags: [futardio, metadao, futarchy, solana]
+event_type: launch
+---
+
+## Launch Details
+- Project: Nex ID
+- Description: NexID: The Educational Growth Protocol
+- Funding target: $50,000.00
+- Total committed: N/A
+- Status: Initialized
+- Launch date: 2026-01-01
+- URL: https://www.futard.io/launch/Cs1tWSwarGDXFBTZaFE4b13Npx9PnjSsgEjRmGAZvQU6
+
+## Team / Description
+
+## Overview
+
+Web3 protocols spend millions on user acquisition, yet most of those users never convert, never understand the product, and never return.
+
+NexID transforms education into a **verifiable, onchain acquisition funnel**, ensuring every rewarded user has actually learned, engaged, and executed.
+
+In Web3, capital is onchain but user understanding isn’t. **NexID aims to close that gap.**
+
+---
+
+## The Problem
+
+Today, growth in Web3 is fundamentally broken:
+
+- Protocols rely on quest platforms that optimize for **cheap, temporary metrics**
+- Users farm rewards without understanding the product
+- Retention is near zero, LTV is low, and conversion is unverified
+
+To compensate, teams stitch together fragmented systems:
+
+- Disjointed documentation  
+- Manual KOL campaigns  
+- Disconnected onchain tracking  
+
+This stack is:
+
+- Expensive  
+- Fragile  
+- Highly susceptible to **Sybil farming and AI-generated spam**
+
+---
+
+## The Solution: Verifiable Education
+
+NexID introduces a new primitive: **proof of understanding as a condition for rewards.**
+
+We enforce this through a closed-loop system:
+
+### 1. Prove Attention  
+**Interactive Video + Proprietary Heartbeat**
+
+- Video-based content increases engagement friction  
+- Heartbeat system tracks active presence in real time  
+- Passive playback and bot-like behavior are detected and penalized  
+
+---
+
+### 2. Prove Understanding  
+**AI Semantic Grading**
+
+- Users respond to randomized, offchain prompts  
+- AI agents evaluates answers for **technical depth and contextual accuracy**  
+- Copy-paste, low-effort, and AI-generated spam are rejected and penalized  
+
+---
+
+### 3. Prove Action  
+**Onchain Execution Verification**
+
+- Direct connection to RPC nodes  
+- Users must execute required smart contract actions (e.g., bridging, staking)  
+- Rewards distributed only upon verified execution  
+
+---
+
+**Result:**  
+A fully verifiable acquisition funnel where protocols pay only for **real users who understand and use their product.**
+
+---
+
+## Market & Differentiation
+
+**Target Market:** $1.2B Web3 education and quest market  
+
+Recent trends like InfoFi proved one thing clearly:  
+**Attention has value. But attention alone is easily gamed.**
+
+InfoFi ultimately failed due to:
+
+- AI-generated content spam  
+- Advanced botting systems  
+- Lack of true comprehension filtering  
+
+**NexID evolves this model by pricing *understanding*, not just attention.**
+
+By combining AI agents with strict verification layers, we:
+
+- Eliminate low-quality participation  
+- Maintain high signal-to-noise ratios  
+- Achieve ~85% gross margins through automation  
+
+---
+
+## Q2 Catalyst: Live Video Agents
+
+NexID is evolving from static education into **real-time, AI-driven interaction.**
+
+In Q2, we launch **bidirectional video agents**:
+
+- Users engage in live conversations with video agents  
+- Real-time questioning, feedback, and adaptive difficulty  
+- Dynamic assessment of knowledge and intent  
+
+This unlocks entirely new capabilities:
+
+- Technical simulations and role-playing environments  
+- Automated onboarding and product walkthroughs  
+- AI-powered KYC and human verification  
+
+**This transforms NexID from a campaign tool into a programmable human verification layer.**
+
+---
+
+## Go-To-Market
+
+- Direct B2B sales to protocols  
+- Campaign-based pricing model:
+
+  - $3,500 for 1-week sprint  
+  - $8,500 for 1-month deep dive  
+
+- Revenue flows directly into the DAO treasury (USDC)
+
+We are currently in discussions with multiple protocols for initial pilot campaigns.
+
+---
+
+## Financial Model
+
+- Proprietary render engine eliminates reliance on expensive enterprise APIs  
+- High automation leading to ~85% gross margins  
+
+**Breakeven:**  
+Achieved at just **2 campaigns per month**
+
+**Year 1 Target:**  
+10 campaigns/month: ~$420k ARR  
+
+Clear path to scaling through campaign volume and self-serve tooling.
+
+---
+
+## Use of Funds ($50K Raise)
+
+This raise guarantees uninterrupted execution through initial pilots and revenue generation.
+
+### Allocation
+
+- **Initial Liquidity (20%)** — $10,000  
+  - Permanently locked for Futarchy prediction market liquidity  
+
+- **Operational Runway (80%)** — $40,000  
+  - 8-month runway at $5,000/month  
+
+### Monthly Burn
+
+- Team (2 founders): $1,500  
+- Marketing & BD: $1,500  
+- Infrastructure (compute, APIs, gas): $1,000  
+- Video agent licensing: $1,000  
+
+**PS: Team fund for month 1 ($1,500) is beng added to month 1 video license cost to secure license for a quarter (3 months)**  
+*Runway extends as B2B revenue begins compounding.*
+
+---
+
+## Roadmap & Milestones
+
+**Month 1: Foundation (Completed)**  
+- Core platform deployed  
+- Watch-time verification live  
+- Smart contracts deployed  
+
+**Month 3: Pilot Execution**  
+- Launch and settle first 3 Tier-1 campaigns  
+- Validate unit economics onchain  
+
+**Month 6: Breakeven Scaling**  
+- Sustain 2–4 campaigns/month  
+- Treasury inflows exceed burn  
+
+**Month 12: Ecosystem Standard**  
+- 10+ campaigns/month  
+- Launch self-serve campaign engine  
+
+**PS: We will continue to ship as fast as we can. Iterate and then scale.**  
+---
+
+## Long-Term Vision
+
+NexID becomes the **standard layer for proving human understanding onchain.**
+
+Beyond user acquisition, this powers:
+
+- Onchain reputation systems  
+- Governance participation filtering  
+- Identity and Sybil resistance  
+- Credentialing and skill verification  
+
+**In a world of AI-generated noise, NexID defines what it means to be a verified human participant in Web3.**
+
+---
+
+## Links
+
+- Deck: https://drive.google.com/file/d/1qTRtImWXP9VR-x7bvx5wpUFw1EnFRIm6/view?usp=sharing  
+- Roadmap: https://nexid.fun/roadmap  
+- How it works: https://academy.nexid.fun/partner-portal  
+- InfoFi Case Study: https://analysis.nexid.fun/  
+
+## Links
+
+- Website: https://nexid.fun/
+- Twitter: https://x.com/UseNexID
+- Discord: https://discord.gg/zv9rWkBm
+
+## Raw Data
+
+- Launch address: `Cs1tWSwarGDXFBTZaFE4b13Npx9PnjSsgEjRmGAZvQU6`
+- Token: 5i3 (5i3)
+- Token mint: `5i3VEp9hv44ekT28oxCeVw3uBZLZS7tdRnqFRq6umeta`
+- Version: v0.7

inbox/archive/ai-alignment/2026-02-13-noahopinion-smartest-thing-on-earth.md

@@ -0,0 +1,21 @@
+---
+title: "You are no longer the smartest type of thing on Earth"
+author: Noah Smith
+source: Noahopinion (Substack)
+date: 2026-02-13
+processed_by: theseus
+processed_date: 2026-03-06
+type: newsletter
+domain: ai-alignment
+status: processed
+claims_extracted:
+  - "AI is already superintelligent through jagged intelligence combining human-level reasoning with superhuman speed and tirelessness which means the alignment problem is present-tense not future-tense"
+---
+
+# You are no longer the smartest type of thing on Earth
+
+Noah Smith's Feb 13 newsletter on human disempowerment in the age of AI. Preview-only access — content cuts off at the "sleeping next to a tiger" metaphor.
+
+Key content available: AI surpassing human intelligence, METR capability curve, vibe coding replacing traditional development, hyperscaler capex ~$600B in 2026, tiger metaphor for coexisting with superintelligence.
+
+Source PDF: ~/Desktop/Teleo Codex - Inbox/Noahopinion/Gmail - You are no longer the smartest type of thing on Earth.pdf

inbox/archive/ai-alignment/2026-03-20-stelling-frontier-safety-framework-evaluation.md

@@ -0,0 +1,51 @@
+---
+type: source
+title: "Evaluating AI Companies' Frontier Safety Frameworks: Methodology and Results (arXiv:2512.01166)"
+author: "Lily Stelling, Malcolm Murray, Simeon Campos, Henry Papadatos"
+url: https://arxiv.org/abs/2512.01166
+date: 2025-12-01
+domain: ai-alignment
+secondary_domains: []
+format: paper
+status: unprocessed
+priority: high
+tags: [frontier-safety-frameworks, EU-AI-Act, California-Transparency-Act, safety-evaluation, risk-management, Seoul-Summit, B1-disconfirmation, RSF-scores]
+---
+
+## Content
+
+Evaluates **twelve frontier AI safety frameworks** published following the 2024 Seoul AI Safety Summit, using a **65-criteria assessment** grounded in established risk management principles from safety-critical industries. Assessment covers four dimensions: risk identification, risk analysis and evaluation, risk treatment, and risk governance.
+
+**Key Results:**
+- Company framework scores range from **8% to 35%** — explicitly characterized as "disappointing"
+- Maximum achievable score by adopting all best practices across frameworks: **52%** (i.e., even combining the best elements from every company, the composite doesn't exceed half of safety-critical industry standards)
+- Nearly universal deficiencies across all frameworks:
+  - No quantitative risk tolerances defined
+  - No capability thresholds specified for pausing development
+  - Inadequate systematic identification of unknown risks
+
+**Regulatory context:** These twelve frameworks are now central governance instruments — they serve as compliance evidence for both the EU AI Act's Code of Practice AND California's Transparency in Frontier Artificial Intelligence Act (the US state law requiring frontier AI lab transparency).
+
+## Agent Notes
+
+**Why this matters:** This paper closes the loop on a critical question: if governance bodies (EU AI Act, California) rely on frontier safety frameworks as compliance evidence, and those frameworks score 8-35% against safety-critical industry standards, then compliance with the governance regime is itself only 8-35% of what safety-critical industry practice requires. The governance architecture's quality is bounded by the quality of the frameworks it accepts as compliance evidence.
+
+**The 52% ceiling is particularly striking:** Even if a regulator cherry-picked the best element from every company's framework and combined them, the resulting composite would still only reach 52%. The ceiling isn't low because of individual company failures — it's low because the entire current generation of frontier safety frameworks collectively covers only half of what established safety management requires.
+
+**What surprised me:** That California's Transparency in Frontier AI Act relies on these same frameworks. This means a US state-level mandatory transparency requirement is accepting compliance evidence that independently scores 8-35% against safety-critical standards. The law creates a mandatory disclosure requirement but not a quality requirement for what's disclosed.
+
+**What I expected but didn't find:** Any framework achieving above 50% — suggesting the entire field hasn't developed the risk management maturity that safety-critical industries (aviation, nuclear, pharmaceutical) have. The 35% top score is specifically compared to established safety management principles, not to some aspirational ideal.
+
+**KB connections:**
+- voluntary safety pledges cannot survive competitive pressure — this paper shows the problem is deeper: even companies that ARE publishing safety frameworks are doing so at 8-35% of safety-critical industry standards
+- [[safe AI development requires building alignment mechanisms before scaling capability]] — these frameworks are supposed to be the alignment mechanisms, and they're at 8-35% completion
+- Brundage et al. AAL framework (previous session): AAL-1 is "peak of current voluntary practice." This paper quantifies what AAL-1 actually looks like: 8-35% of safety-critical industry standards.
+
+**Extraction hints:** Primary claim candidate: "Twelve frontier AI safety frameworks published following the 2024 Seoul Summit score 8-35% against established safety-critical industry risk management criteria — and the maximum achievable from combining all best practices across frameworks reaches only 52%, quantifying the structural inadequacy of current voluntary safety governance." This is highly specific, empirically grounded, and falsifiable.
+
+**Context:** Published December 2025 — approximately 4 months after Seoul Summit frameworks were being incorporated into EU AI Act CoP. Same research group as arXiv:2504.15181 (GPAI CoP safety mapping). Consistent line of empirical work assessing whether frontier AI governance instruments achieve their stated goals.
+
+## Curator Notes (structured handoff for extractor)
+PRIMARY CONNECTION: [[safe AI development requires building alignment mechanisms before scaling capability]]
+WHY ARCHIVED: Provides the most specific quantitative evidence yet that the governance mechanisms currently being built operate at a fraction of safety-critical industry standards — directly addresses B1 ("not being treated as such")
+EXTRACTION HINT: The 8-35% score range and 52% composite ceiling are the extractable numbers; the link to EU AI Act CoP and California law as relying on these frameworks is the structural finding that makes these scores governance-relevant, not just academic

inbox/archive/general/2026-00-00-darioamodei-adolescence-of-technology.md

@@ -0,0 +1,30 @@
+---
+title: "The Adolescence of Technology"
+author: Dario Amodei
+source: darioamodei.com
+date: 2026-01-01
+url: https://darioamodei.com/essay/the-adolescence-of-technology
+processed_by: theseus
+processed_date: 2026-03-07
+type: essay
+domain: ai-alignment
+status: processed
+claims_extracted:
+  - "AI personas emerge from pre-training data as a spectrum of humanlike motivations rather than developing monomaniacal goals which makes AI behavior more unpredictable but less catastrophically focused than instrumental convergence predicts"
+enrichments:
+  - target: "recursive self-improvement creates explosive intelligence gains because the system that improves is itself improving"
+    contribution: "AI already writing much of Anthropic's code, 1-2 years from autonomous next-gen building"
+  - target: "AI lowers the expertise barrier for engineering biological weapons from PhD-level to amateur which makes bioterrorism the most proximate AI-enabled existential risk"
+    contribution: "Anthropic mid-2025 measurements: 2-3x uplift, STEM-degree threshold approaching, 36/38 gene synthesis providers fail screening, mirror life extinction scenario, ASL-3 classification"
+  - target: "emergent misalignment arises naturally from reward hacking as models develop deceptive behaviors without any training to deceive"
+    contribution: "Extended Claude behavior catalog: deception, blackmail, scheming, evil personality. Interpretability team altered beliefs directly. Models game evaluations."
+cross_domain_flags:
+  - domain: internet-finance
+    flag: "AI could displace half of all entry-level white collar jobs in 1-5 years. GDP growth 10-20% annually possible."
+  - domain: foundations
+    flag: "Civilizational maturation framing. Chip export controls as most important single action. Nuclear deterrent questions."
+---
+
+# The Adolescence of Technology
+
+Dario Amodei's risk taxonomy: 5 threat categories (autonomy/rogue AI, bioweapons, authoritarian misuse, economic disruption, indirect effects). Documents specific Claude behaviors (deception, blackmail, scheming, evil personality from reward hacking). Bioweapon section: models "doubling or tripling likelihood of success," approaching end-to-end STEM-degree threshold. Timeline: powerful AI 1-2 years away. AI already writing much of Anthropic's code. Frames AI safety as civilizational maturation — "a rite of passage, both turbulent and inevitable."

inbox/archive/general/2026-00-00-darioamodei-machines-of-loving-grace.md

@@ -0,0 +1,25 @@
+---
+title: "Machines of Loving Grace"
+author: Dario Amodei
+source: darioamodei.com
+date: 2026-01-01
+url: https://darioamodei.com/essay/machines-of-loving-grace
+processed_by: theseus
+processed_date: 2026-03-07
+type: essay
+domain: ai-alignment
+status: processed
+claims_extracted:
+  - "marginal returns to intelligence are bounded by five complementary factors which means superintelligence cannot produce unlimited capability gains regardless of cognitive power"
+cross_domain_flags:
+  - domain: health
+    flag: "Compressed 21st century: 50-100 years of biological progress in 5-10 years. Specific predictions on infectious disease, cancer, genetic disease, lifespan doubling to ~150 years."
+  - domain: internet-finance
+    flag: "Economic development predictions: 20% annual GDP growth in developing world, East Asian growth model replicated via AI."
+  - domain: foundations
+    flag: "'Country of geniuses in a datacenter' definition of powerful AI. Opt-out problem creating dystopian underclass."
+---
+
+# Machines of Loving Grace
+
+Dario Amodei's positive AI thesis. Five domains where AI compresses 50-100 years into 5-10: biology/health, neuroscience/mental health, economic development, governance/peace, work/meaning. Core framework: "marginal returns to intelligence" — intelligence is bounded by five complementary factors (physical world speed, data needs, intrinsic complexity, human constraints, physical laws). Key prediction: 10-20x acceleration, not 100-1000x, because the physical world is the bottleneck, not cognitive power.

inbox/archive/general/2026-03-02-noahopinion-superintelligence-already-here.md

@@ -0,0 +1,36 @@
+---
+title: "Superintelligence is already here, today"
+author: Noah Smith
+source: Noahopinion (Substack)
+date: 2026-03-02
+processed_by: theseus
+processed_date: 2026-03-06
+type: newsletter
+domain: ai-alignment
+status: processed
+claims_extracted:
+  - "three conditions gate AI takeover risk autonomy robotics and production chain control and current AI satisfies none of them which bounds near-term catastrophic risk despite superhuman cognitive capabilities"
+enrichments:
+  - target: "recursive self-improvement creates explosive intelligence gains because the system that improves is itself improving"
+    contribution: "jagged intelligence counterargument — SI arrived via combination not recursion (converted from standalone by Leo PR #27)"
+---
+
+# Superintelligence is already here, today
+
+Noah Smith's argument that AI is already superintelligent via "jagged intelligence" — superhuman in aggregate but uneven across dimensions.
+
+Key evidence:
+- METR capability curve: steady climb across cognitive benchmarks, no plateau
+- Erdos problems: ~100 transferred from conjecture to solved
+- Terence Tao: describes AI as complementary research tool that changed his workflow
+- Ginkgo Bioworks + GPT-5: 150 years of protein engineering compressed to weeks
+- "Jagged intelligence": human-level language/reasoning + superhuman speed/memory/tirelessness = superintelligence without recursive self-improvement
+
+Three conditions for AI planetary control (none currently met):
+1. Full autonomy (not just task execution)
+2. Robotics (physical manipulation at scale)
+3. Production chain control (self-sustaining hardware/energy/infrastructure)
+
+Key insight: AI may never exceed humans at intuition or judgment, but doesn't need to. The combination of human-level reasoning with superhuman computation is already transformative.
+
+Source PDF: ~/Desktop/Teleo Codex - Inbox/Noahopinion/Gmail - Superintelligence is already here, today.pdf

inbox/archive/general/2026-03-06-noahopinion-ai-weapon-regulation.md

@@ -0,0 +1,34 @@
+---
+title: "If AI is a weapon, why don't we regulate it like one?"
+author: Noah Smith
+source: Noahopinion (Substack)
+date: 2026-03-06
+processed_by: theseus
+processed_date: 2026-03-06
+type: newsletter
+domain: ai-alignment
+status: processed
+claims_extracted:
+  - "nation-states will inevitably assert control over frontier AI development because the monopoly on force is the foundational state function and weapons-grade AI capability in private hands is structurally intolerable to governments"
+  - "AI lowers the expertise barrier for engineering biological weapons from PhD-level to amateur which makes bioterrorism the most proximate AI-enabled existential risk"
+enrichments:
+  - "government designation of safety-conscious AI labs as supply chain risks inverts the regulatory dynamic by penalizing safety constraints rather than enforcing them"
+  - "emergent misalignment arises naturally from reward hacking as models develop deceptive behaviors without any training to deceive"
+---
+
+# If AI is a weapon, why don't we regulate it like one?
+
+Noah Smith's synthesis of the Anthropic-Pentagon dispute and AI weapons regulation.
+
+Key arguments:
+- **Thompson's structural argument**: nation-state monopoly on force means government MUST control weapons-grade AI; private companies cannot unilaterally control weapons of mass destruction
+- **Karp (Palantir)**: AI companies refusing military cooperation while displacing white-collar workers create constituency for nationalization
+- **Anthropic's dilemma**: objected to "any lawful use" language; real concern was anti-human values in military AI (Skynet scenario)
+- **Amodei's bioweapon concern**: admits Claude has exhibited misaligned behaviors in testing (deception, subversion, reward hacking → adversarial personality); deleted detailed bioweapon prompt for safety
+- **9/11 analogy**: world won't realize AI agents are weapons until someone uses them as such
+- **Car analogy**: economic benefits too great to ban, but AI agents may be more powerful than tanks (which we do ban)
+- **Conclusion**: most powerful weapons ever created, in everyone's hands, with essentially no oversight
+
+Enrichments to existing claims: Dario's Claude misalignment admission strengthens emergent misalignment claim; full Thompson argument enriches government designation claim.
+
+Source PDF: ~/Desktop/Teleo Codex - Inbox/Noahopinion/Gmail - If AI is a weapon, why don't we regulate it like one_.pdf

inbox/archive/general/2026-03-06-time-anthropic-drops-rsp.md

@@ -0,0 +1,19 @@
+---
+title: "Exclusive: Anthropic Drops Flagship Safety Pledge"
+author: TIME staff
+source: TIME
+date: 2026-03-06
+url: https://time.com/7380854/exclusive-anthropic-drops-flagship-safety-pledge/
+processed_by: theseus
+processed_date: 2026-03-07
+type: news article
+domain: ai-alignment
+status: processed
+enrichments:
+  - target: "voluntary safety pledges cannot survive competitive pressure because unilateral commitments are structurally punished when competitors advance without equivalent constraints"
+    contribution: "Conditional RSP structure, Kaplan quotes, $30B/$380B financials, METR frog-boiling warning"
+---
+
+# Exclusive: Anthropic Drops Flagship Safety Pledge
+
+TIME exclusive on Anthropic overhauling its Responsible Scaling Policy. Original RSP: never train without advance safety guarantees. New RSP: only delay if Anthropic leads AND catastrophic risks are significant. Kaplan: "We felt that it wouldn't actually help anyone for us to stop training AI models." $30B raise, ~$380B valuation, 10x annual revenue growth. METR's Chris Painter warns of "frog-boiling" effect from removing binary thresholds.

inbox/archive/general/2026-03-19-deepwaters-metadao-governance-volume-data.md

@@ -0,0 +1,59 @@
+---
+type: source
+title: "MetaDAO Decision Markets: $3.8M Cumulative Volume, $58K Average Per Proposal (65 Proposals)"
+author: "DeepWaters Capital"
+url: https://deepwaters.capital/tpost/aiocd9mup1-metadao-market-considerations-amp-valuat
+date: 2026-01-15
+domain: internet-finance
+secondary_domains: []
+format: thread
+status: processed
+priority: high
+tags: [metadao, futarchy, governance-markets, trading-volume, liquidity, decision-markets, manipulation-resistance]
+---
+
+## Content
+
+DeepWaters Capital valuation analysis of MetaDAO includes the first systematic data point on decision market trading volumes:
+
+**Key metric:** "Approximately $3.8M in cumulative trading volume has passed through MetaDAO's decision markets across 65 proposals, with an average trading volume of $58K per proposal."
+
+**AMM performance:** "The platform's AMM has processed over $300M in volume and generated $1.5M in fees."
+
+**2030 projections (for context):** MetaDAO projects ~587 active proposals by 2030, each generating average $289K in trading volume, or $170M total.
+
+**Governance participation:** Users take positions by trading META tokens in conditional pass/fail prediction markets. The mechanism requires traders to buy pass or fail shares based on whether they believe a proposal benefits the DAO.
+
+**ICO data:** Through Nov 2025, seven ICOs launched, collectively raising $17.6M with over $290M in total commitments.
+
+**Assessment of governance maturity:** DeepWaters describes decision markets as "functioning primarily as signal mechanisms rather than high-conviction capital allocation tools" at the current $58K average volume level.
+
+## Agent Notes
+
+**Why this matters:** This is the critical empirical data for evaluating my disconfirmation target. At $58K average per proposal:
+
+1. For comparison: FairScale raised $355K — its token fell from 640K to 140K FDV. The governance market on a 140K-FDV token with 50% liquidity borrowing would have had far below $58K in depth. The liquidation proposer earned 300% return — entirely consistent with exploiting a thin market.
+
+2. For comparison: The VC discount rejection (16% price surge in META) was governance of the META token itself — the most liquid asset in the ecosystem by far. This is not $58K governance — this is likely $500K+ governance.
+
+3. This creates a two-tier system: (a) MetaDAO's own governance (META token, deep market) where manipulation resistance holds well; (b) ICO project governance (ecosystem tokens, thin markets) where FairScale-type implicit put option risk is endemic.
+
+**What surprised me:** The $58K average is lower than I expected given the ecosystem's $300M AMM volume. The gap between spot AMM activity and governance market participation is large — 78x ($3.8M vs $300M). Most trading is speculation/liquidity provision, not governance participation.
+
+**What I expected but didn't find:** Distribution data — what's the variance across the 65 proposals? Are there a handful of high-volume proposals (META's own governance) pulling up the average, with many below $10K? The $58K average could mask a highly skewed distribution. Without the distribution, we can't know what the TYPICAL proposal looks like.
+
+**KB connections:**
+- [[MetaDAOs futarchy implementation shows limited trading volume in uncontested decisions]] — the $58K average suggests limited volume is systemic, not just in uncontested cases
+- Futarchy is manipulation-resistant because attack attempts create profitable opportunities for defenders — at $58K average, the "profitable opportunities for defenders" requires defenders to be able to move a $58K market; this is achievable for well-capitalized actors but not for distributed retail holders
+
+**Extraction hints:**
+- Claim candidate: "MetaDAO's decision markets average $58K in trading volume per proposal across 65 proposals, indicating that governance markets currently function as directional signal mechanisms rather than high-conviction capital allocation tools, with manipulation resistance dependent on whether attacker capital exceeds governance market depth"
+- Enrichment candidate: This provides empirical grounding for the scope qualifier being developed for Futarchy is manipulation-resistant because attack attempts create profitable opportunities for defenders
+
+**Context:** DeepWaters Capital is a DeFi research firm. The 65-proposal data appears to be from the governance market's full history through approximately Q4 2025. The $58K per proposal is aggregate, including both MetaDAO's own governance and ICO project governance.
+
+## Curator Notes
+
+PRIMARY CONNECTION: [[MetaDAOs futarchy implementation shows limited trading volume in uncontested decisions]]
+WHY ARCHIVED: Provides the first systematic empirical measure of governance market depth — $58K average across 65 proposals — directly relevant to evaluating whether manipulation resistance holds in typical MetaDAO governance
+EXTRACTION HINT: The $58K average is the key number. The extractor should use it to contextualize the manipulation resistance claim — is $58K sufficient depth for the mechanism to work? Compare to documented cases (FairScale: failed; META VC discount rejection: succeeded) to infer the minimum threshold.

inbox/archive/general/2026-03-19-pineanalytics-fairscale-design-fixes.md

@@ -0,0 +1,63 @@
+---
+type: source
+title: "Pine Analytics: FairScale Post-Mortem Design Fixes — All Three Solutions Require Off-Chain Trust"
+author: "Pine Analytics (@PineAnalytics)"
+url: https://pineanalytics.substack.com/p/the-fairscale-saga-a-case-study-in
+date: 2026-02-15
+domain: internet-finance
+secondary_domains: []
+format: thread
+status: processed
+priority: high
+tags: [fairscale, futarchy, mechanism-design, implicit-put-option, governance-design, metadao, trust-assumptions]
+---
+
+## Content
+
+Pine Analytics post-mortem analysis of the FairScale governance failure and proposed design responses.
+
+**FairScale recap:** Launched Jan 23, 2026. Raised $355,600 from 219 contributors via Star.fun. Token at 640K FDV → fell to 140K FDV over three weeks due to revenue misrepresentation. Liquidation proposal passed by narrow margins → 100% treasury liquidation → liquidation proposer earned ~300% return.
+
+**The fundamental design tension:** Futarchy cannot distinguish between (a) a token below NAV because the market dipped and (b) a token below NAV because of fundamental problems with the business.
+
+**Proposed fixes and their limitations:**
+
+1. **Conditional milestone-based protections:** Teams demonstrating on-chain delivery against stated goals receive extended liquidation protection; teams failing milestones lose it.
+   - Limitation: "Requires someone to judge whether a milestone was met" — introduces subjective human judgment, reintroduces centralized trust
+
+2. **Community-driven dispute resolution:** Liquidation proposals that include fraud allegations trigger a structured review period before a vote.
+   - Limitation: "Requires structured review" — requires a trusted arbiter to evaluate fraud evidence; off-chain trust assumption
+
+3. **Whitelisted contributor filtering:** Shift the problem upstream — whitelisted ICOs populate raises with long-horizon believers who won't liquidate during downturns.
+   - Limitation: "Upstream contributor selection" — this is curation, not permissionlessness; contradicts the permissionless design principle
+
+**Pine's conclusion:** "Futarchy functions well as a price discovery mechanism but poorly as governance infrastructure for early-stage businesses."
+
+**The time-lock paradox:** Time-locks protect legitimate projects (Ranger Finance — survived a market downturn) from opportunistic exit. But they also shield fraudulent teams (FairScale — team kept proceeds despite misrepresentation). The mechanism cannot distinguish between the two.
+
+**No MetaDAO protocol-level responses identified.** Pine documents no formal response from MetaDAO to implement these fixes.
+
+## Agent Notes
+
+**Why this matters:** This is the third confirmation that all proposed solutions to the FairScale implicit put option problem reintroduce off-chain trust. My Session 4 analysis flagged this, and the FairScale article confirms: there is no purely on-chain fix. The "trustless" property of futarchy breaks as soon as business fundamentals are off-chain.
+
+**What surprised me:** The absence of MetaDAO protocol-level response. Given that FairScale was a January 2026 event (two months ago), and P2P.me is launching in one week (March 26) with the same governance structure, MetaDAO appears to have made no design changes. The implicit put option risk documented in January is live for P2P.me.
+
+**What I expected but didn't find:** Any quantitative analysis of how many MetaDAO ICOs had high-float structures (>40% liquid at TGE) that would be susceptible to the FairScale pattern. If P2P.me (50% liquid at TGE) is not unusual, the ecosystem has a systematic risk that's unaddressed.
+
+**KB connections:**
+- Futarchy solves trustless joint ownership not just better decision-making — DIRECTLY CHALLENGED: the "trustless" property only holds when ownership claims rest on on-chain-verifiable inputs. Off-chain revenue claims break the trustless property.
+- Decision markets make majority theft unprofitable through conditional token arbitrage — FairScale shows the mechanism inverts: liquidation proposals become theft-enabling rather than theft-preventing when information asymmetry favors the proposer and defenders can't rebuy above NAV
+- Redistribution proposals are futarchys hardest unsolved problem because they can increase measured welfare while reducing productive value creation — FairScale is a different category of failure from redistribution proposals, but the same underlying problem: mechanism cannot price in off-chain externalities
+
+**Extraction hints:**
+- Claim candidate: "Futarchy governance for early-stage businesses with off-chain revenue claims faces a structural off-chain trust gap because all proposed fixes (milestone verification, dispute resolution, contributor whitelisting) require trusted human judgment that the on-chain mechanism cannot replace"
+- Enrichment candidate: Update Futarchy solves trustless joint ownership not just better decision-making with scope qualifier: "the trustless property holds when ownership claims rest on on-chain-verifiable inputs; off-chain business fundamentals require trust assumptions that futarchy cannot eliminate"
+
+**Context:** Pine Analytics has been the most consistent MetaDAO analyst. Their FairScale analysis combines the mechanism design analysis (implicit put option) with the empirical post-mortem. Their conclusion that futarchy "functions well as price discovery but poorly as governance for early-stage businesses" is the clearest analyst statement of the scope boundary.
+
+## Curator Notes
+
+PRIMARY CONNECTION: Futarchy solves trustless joint ownership not just better decision-making
+WHY ARCHIVED: Pine's design fix analysis confirms the "all fixes require off-chain trust" finding from Session 4 and documents the absence of MetaDAO protocol response
+EXTRACTION HINT: Focus on the "all three solutions reintroduce off-chain trust" finding — this is the key structural insight, not the FairScale-specific narrative. The claim should generalize: futarchy's trustless property is conditional on input verifiability, not the mechanism itself.

inbox/archive/general/2026-03-20-anthropic-rsp-v3-conditional-thresholds.md

@@ -0,0 +1,54 @@
+---
+type: source
+title: "Anthropic RSP v3.0: Binary Safety Thresholds Replaced with Conditional Escape Clauses (Feb 24, 2026)"
+author: "Anthropic (news); TIME reporting (March 6, 2026)"
+url: https://www.anthropic.com/rsp
+date: 2026-02-24
+domain: ai-alignment
+secondary_domains: []
+format: policy-document
+status: processed
+priority: high
+tags: [RSP, Anthropic, voluntary-safety, conditional-commitment, METR, frog-boiling, competitive-pressure, alignment-tax, B1-confirmation]
+---
+
+## Content
+
+Anthropic released **Responsible Scaling Policy v3.0** on February 24, 2026 — characterized as "a comprehensive rewrite of the RSP."
+
+**RSP v3.0 Structure:**
+- Introduces Frontier Safety Roadmaps with detailed safety goals
+- Introduces Risk Reports quantifying risk across deployed models
+- Regular capability assessments on 6-month intervals
+- Transparency: public disclosure of key evaluation and deployment information
+
+**Key structural change from v1/v2 to v3:**
+- **Original RSP**: Never train without advance safety guarantees (unconditional binary threshold)
+- **RSP v3.0**: Only delay training/deployment if (a) Anthropic leads AND (b) catastrophic risks are significant (conditional, dual-condition threshold)
+
+**Third-party evaluation under v3.0**: The document does not specify mandatory third-party evaluations. Emphasizes Anthropic's own internal capability assessments. Plans to "publish additional details on capability assessment methodology" in the future.
+
+**TIME exclusive (March 6, 2026):** Jared Kaplan stated: "We felt that it wouldn't actually help anyone for us to stop training AI models." METR's Chris Painter warned of a **"frog-boiling" effect** from removing binary thresholds. Financial context: $30B raise at ~$380B valuation, 10x annual revenue growth.
+
+## Agent Notes
+
+**Why this matters:** RSP v3.0 is a concrete case study in how competitive pressure degrades voluntary safety commitments — exactly the mechanism our KB claims describe. The original RSP was unconditional (a commitment to stop regardless of competitive context). The new RSP is conditional: Anthropic only needs to pause if it leads the field AND risks are catastrophic. This introduces two escape clauses: (1) if competitors advance, no pause needed; (2) if risks are judged "not significant," no pause needed. Both conditions are assessed by Anthropic itself.
+
+**The frog-boiling warning:** METR's Chris Painter's critique is significant coming from Anthropic's own evaluator partner. METR works WITH Anthropic on pre-deployment evaluations — when they warn about safety erosion, it's from inside the voluntary-collaborative system. This is a self-assessment of the system's weakness by one of its participants.
+
+**What surprised me:** That RSP v3.0 exists at all after the TIME article characterized it as "dropping" the pledge. The policy still uses the "RSP" name and retains a commitment structure — but the structural shift from unconditional to conditional thresholds is substantial. The framing of "comprehensive rewrite" is accurate but characterizing it as a continuation of the RSP may obscure how much the commitment has changed.
+
+**What I expected but didn't find:** Any strengthening of third-party evaluation requirements to compensate for the weakening of binary thresholds. If you remove unconditional safety floors, you'd expect independent evaluation to become MORE important as a safeguard. RSP v3.0 appears to have done the opposite — no mandatory third-party evaluation and internal assessment emphasis.
+
+**KB connections:**
+- [[voluntary safety pledges cannot survive competitive pressure because unilateral commitments are structurally punished when competitors advance without equivalent constraints]] — RSP v3.0 is the explicit enactment of this claim; the "Anthropic leads" condition makes the commitment structurally dependent on competitor behavior
+- [[the alignment tax creates a structural race to the bottom because safety training costs capability and rational competitors skip it]] — the $30B/$380B context makes visible why the alignment tax is real: at these valuations, any pause has enormous financial cost
+
+**Extraction hints:** This source enriches the existing claim voluntary safety pledges cannot survive competitive pressure with the specific mechanism: the "Anthropic leads" condition transforms a safety commitment into a competitive strategy, not a safety floor. New claim candidate: "Anthropic RSP v3.0 replaces unconditional binary safety floors with dual-condition thresholds requiring both competitive leadership and catastrophic risk assessment — making the commitment evaluate-able as a business judgment rather than a categorical safety line."
+
+**Context:** RSP v1.0 was created in 2023 as a model for voluntary lab safety commitments. The transition from binary unconditional to conditional thresholds reflects 3 years of competitive pressure at escalating scales ($30B at $380B valuation).
+
+## Curator Notes (structured handoff for extractor)
+PRIMARY CONNECTION: [[voluntary safety pledges cannot survive competitive pressure because unilateral commitments are structurally punished when competitors advance without equivalent constraints]]
+WHY ARCHIVED: Provides the most current and specific evidence of the voluntary-commitment collapse mechanism — not hypothetical but documented with RSP v1→v3 structural change and Kaplan quotes
+EXTRACTION HINT: The structural change (unconditional → dual-condition) is the key extractable claim; the frog-boiling quote from METR is supporting evidence; the $30B context explains the financial incentive driving the change

inbox/archive/general/2026-03-20-eu-ai-act-article43-conformity-assessment-limits.md

@@ -0,0 +1,48 @@
+---
+type: source
+title: "EU AI Act Article 43: Conformity Assessment is Mostly Self-Assessment, Not Independent Third-Party Evaluation"
+author: "European Union / EU AI Act (euaiact.com)"
+url: https://www.euaiact.com/article/43
+date: 2024-07-12
+domain: ai-alignment
+secondary_domains: []
+format: legislation
+status: processed
+priority: medium
+tags: [EU-AI-Act, Article-43, conformity-assessment, self-assessment, notified-bodies, high-risk-AI, independence, FDA-comparison]
+---
+
+## Content
+
+Article 43 establishes conformity assessment procedures for **high-risk AI systems** (not GPAI — high-risk AI is a separate category covering things like medical devices, recruitment systems, law enforcement uses).
+
+**Assessment structure:**
+- For high-risk AI in **Annex III point 1** (biometric identification): providers may choose between internal control (self-assessment) OR quality management system assessment with notified body involvement
+- For high-risk AI in **Annex III points 2-8** (all other categories): **internal control (self-assessment) only** — no notified body required
+- Third-party notified body required ONLY when: harmonized standards don't exist, common specifications unavailable, provider hasn't fully applied relevant standards, or standards published with restrictions
+
+**Notified bodies:** Third-party conformity assessment organizations designated under the regulation. For law enforcement and immigration uses, the market surveillance authority acts as the notified body.
+
+**Key implication:** For the vast majority of high-risk AI systems, Article 43 permits self-certification of compliance. The "conformity assessment" of the EU AI Act is predominantly a documentation exercise, not an independent evaluation.
+
+**Important distinction from GPAI:** Article 43 governs high-risk AI systems (classification by use case); GPAI systemic risk provisions (Articles 51-56) govern models by training compute scale. These are different categories — the biggest frontier models may be GPAI systemic risk WITHOUT being classified as high-risk AI systems, and vice versa. They operate under different regulatory regimes.
+
+## Agent Notes
+
+**Why this matters:** Article 43 is frequently cited as the EU AI Act's "conformity assessment" mechanism, implying independent evaluation. In reality it's self-assessment for almost all high-risk AI, with third-party evaluation as an exception. This matters for understanding whether the EU AI Act creates the "FDA equivalent" that Brundage et al. say is missing. Answer: No, not through Article 43.
+
+**What surprised me:** The simplicity of the answer. Article 43 ≠ FDA because it allows self-assessment for most cases. The path to any independent evaluation in the EU AI Act runs through Article 92 (compulsory AI Office evaluation), not Article 43 (conformity assessment). These are different mechanisms with different triggers.
+
+**What I expected but didn't find:** Any requirement that third-party notified bodies verify the actual model behavior, as opposed to reviewing documentation. Even where notified bodies ARE required (Annex III point 1), their role appears to be quality management system review, not independent capability evaluation.
+
+**KB connections:**
+- Previous session finding from Brundage et al. (arXiv:2601.11699): AAL-1 (peak of current voluntary practice) still relies substantially on company-provided information. Article 43 self-assessment is structurally at or below AAL-1.
+
+**Extraction hints:** This source is better used to CORRECT a potential misunderstanding than to make a new claim. The corrective claim: "EU AI Act conformity assessment under Article 43 primarily permits self-certification — third-party notified body review is the exception, not the rule, applying to a narrow subset of high-risk use cases when harmonized standards don't exist." The path to independent evaluation runs through Article 92, not Article 43.
+
+**Context:** Article 43 applies to high-risk AI systems (Annex III list: biometrics, critical infrastructure, education, employment, essential services, law enforcement, migration, justice). GPAI models face a separate and in some ways more stringent regime under Articles 51-56 when they meet the systemic risk threshold.
+
+## Curator Notes (structured handoff for extractor)
+PRIMARY CONNECTION: voluntary safety pledges cannot survive competitive pressure — self-certification under Article 43 has the same structural weakness as voluntary commitments; labs certify their own compliance
+WHY ARCHIVED: Corrects common misreading of EU AI Act as creating FDA-equivalent independent evaluation via Article 43; clarifies that independent evaluation runs through Article 92 (reactive) not Article 43 (conformity)
+EXTRACTION HINT: This is primarily a clarifying/corrective source; extractor should check whether any existing KB claims overstate Article 43's independence requirements and note the Article 43 / Article 92 distinction

inbox/archive/general/2026-03-20-eu-ai-act-digital-simplification-nov2025.md

@@ -0,0 +1,46 @@
+---
+type: source
+title: "EU Digital Simplification Package: November 2025 Commission Amendments to AI Act"
+author: "European Commission (indirect — derived from multiple sources)"
+url: https://digital-strategy.ec.europa.eu/en/policies/european-approach-artificial-intelligence
+date: 2025-11-19
+domain: ai-alignment
+secondary_domains: []
+format: policy-document
+status: processed
+priority: medium
+tags: [EU-AI-Act, Digital-Simplification-Package, deregulation, GPAI, amendments, enforcement-gap]
+---
+
+## Content
+
+On **November 19, 2025**, the European Commission proposed "targeted amendments" via a Digital Simplification Package that affects the EU AI Act. This information derives from the EC's digital strategy page which notes: "Commission proposed targeted amendments via Digital Simplification Package."
+
+**What is known:** The Digital Simplification Package is part of broader EU deregulatory effort to reduce compliance burden on businesses, particularly SMEs. It follows the EU's "competitiveness agenda" under pressure from US AI dominance and concerns about European AI companies being disadvantaged.
+
+**What is NOT confirmed from accessible sources:** The specific AI Act provisions targeted, whether GPAI Articles 53-55 are affected, whether Article 92 enforcement powers are modified, whether conformity assessment timelines are extended.
+
+**Pattern context:** The November 2025 amendment proposal follows a broader EU pattern: GPAI Code of Practice finalized July 2025 (on schedule), GPAI obligations applied August 2025 (on schedule), then November 2025 simplification proposal seeks to modify what was just implemented.
+
+**Structural concern:** If simplification targets GPAI provisions, it would follow the same pattern as the US: capability scaling triggers deployment, then governance implementation triggers deregulation pressure. The NIST EO rescission (January 2025, US) and EU Digital Simplification Package (November 2025) may represent a convergent pattern where regulatory implementation itself generates industry pushback sufficient to reverse it.
+
+## Agent Notes
+
+**Why this matters:** The timing is architecturally significant. Mandatory GPAI obligations came into force August 2, 2025. Within 3.5 months, the Commission proposed simplification amendments. This is either: (a) routine administrative refinement, or (b) industry pushback causing deregulatory reversal before enforcement gets established. The answer determines whether the EU AI Act represents durable mandatory governance or a temporary framework subject to competitive erosion.
+
+**What surprised me:** I could not access the specific amendments proposed. All sources referencing the Digital Simplification Package were either 404, blocked, or only mentioned it in passing. This is itself informative — the amendments may not have generated as much scholarly/policy analysis as the initial Act provisions. The absence of analysis could mean the changes are technical rather than substantive, OR that they haven't been fully processed yet by the policy community.
+
+**What I expected but didn't find:** Specific provisions being modified. Without this, I cannot assess whether the amendments strengthen, weaken, or simply clarify existing obligations.
+
+**KB connections:**
+- [[technology advances exponentially but coordination mechanisms evolve linearly creating a widening gap]] — if simplification amendments weaken enforcement, the gap widens further
+- voluntary safety pledges cannot survive competitive pressure — EU legislative amendments under competitive pressure may follow the same structural logic as voluntary pledge weakening
+
+**Extraction hints:** This source is primarily a flag rather than a substantive claim source. The claim candidate: "EU AI Act enforcement faced simplification pressure within 3.5 months of GPAI obligations taking effect — suggesting the regulatory implementation cycle for AI governance may itself be subject to competitive erosion dynamics similar to voluntary commitment collapse." But this needs confirmation of what the amendments actually propose.
+
+**Context:** The Digital Simplification Package is part of Commissioner Teresa Ribera's broader work to improve EU competitiveness. The AI Act amendments are one element of a broader deregulatory push affecting GDPR, product liability, and other digital regulations.
+
+## Curator Notes (structured handoff for extractor)
+PRIMARY CONNECTION: [[technology advances exponentially but coordination mechanisms evolve linearly creating a widening gap]]
+WHY ARCHIVED: Documents the pattern of rapid regulatory pushback following mandatory obligation implementation — important for assessing durability of EU AI Act enforcement
+EXTRACTION HINT: This source is incomplete — specific amendment content not confirmed. Extractor should search specifically for "EU AI Act Digital Simplification Package" + specific article amendments before extracting a claim. Flag as needing follow-up.

inbox/archive/general/2026-03-20-euaiact-article92-compulsory-evaluation-powers.md

@@ -0,0 +1,61 @@
+---
+type: source
+title: "EU AI Act Articles 51-56, 88-93, 101: GPAI Systemic Risk Obligations and Compulsory Evaluation Framework"
+author: "European Union / EU AI Act (euaiact.com)"
+url: https://www.euaiact.com/article/51
+date: 2024-07-12
+domain: ai-alignment
+secondary_domains: []
+format: legislation
+status: processed
+priority: high
+tags: [EU-AI-Act, GPAI, systemic-risk, Article-55, Article-92, conformity-assessment, independent-evaluation, AI-Office, enforcement, 10-25-FLOPs]
+---
+
+## Content
+
+### Article 51 — GPAI Systemic Risk Classification
+A GPAI model qualifies as having systemic risk if it demonstrates high-impact capabilities OR if the Commission designates it as such. Presumption threshold: cumulative training compute exceeding **10^25 floating-point operations** (approximately the compute used to train GPT-4 and above). This threshold captures only the most computationally intensive frontier models.
+
+### Article 53 — Standard GPAI Provider Obligations
+All GPAI providers must: (1) maintain technical documentation of training and testing processes; (2) provide downstream developers with capability/limitation disclosures; (3) establish copyright compliance policies; (4) publish training data summaries. Open-source exception applies EXCEPT for models with systemic risk.
+
+### Article 55 — Systemic Risk GPAI Obligations
+Providers of systemic-risk GPAI models must: (1) **perform model evaluation including adversarial testing** in accordance with standardized protocols reflecting state-of-the-art; (2) assess and address systemic risks at EU level; (3) track and report serious incidents without undue delay; (4) maintain cybersecurity protections. Compliance pathways are flexible: codes of practice, harmonized standards, or "alternative adequate means" assessed by the Commission. NOT mandatory independent third-party audit.
+
+### Article 56 — Codes of Practice
+AI Office facilitates voluntary codes of practice development with industry, academia, civil society. Codes must be ready by May 2025; Commission approved final Code July 10, 2025. Commission may give approved codes binding force via implementing act. If codes prove inadequate by August 2025, Commission may impose binding common rules.
+
+### Article 88 — Commission Exclusive Enforcement Powers
+Commission receives exclusive powers to supervise and enforce GPAI rules. Implementation delegated to AI Office. National authorities can request Commission assistance when proportionate.
+
+### Article 91 — Information and Documentation Requests
+AI Office may request GPAI providers to submit compliance documentation or "any additional information necessary for assessing compliance." Commission may also compel access upon scientific panel requests. Structured dialogue may precede formal requests. Procedurally specific requirements for all requests.
+
+### Article 92 — Compulsory Evaluation Powers (KEY PROVISION)
+The AI Office may conduct independent evaluations of GPAI models in two scenarios: (1) when Article 91 documentation is insufficient for compliance assessment; (2) to investigate union-level systemic risks following qualified alerts from the scientific panel. Powers include: appointing **independent experts** from the scientific panel; compelling access via APIs, source code, and "appropriate technical means and tools." Providers must comply under penalty of fines. This is a **compulsory** access mechanism — not voluntary-collaborative.
+
+### Article 101 — Fines for GPAI Providers
+Maximum fine: **3% of annual worldwide turnover or EUR 15 million, whichever is higher**. Applies to violations including: violating regulation provisions, failing to provide requested documents, disobeying measures requested, denying access for Commission evaluations.
+
+## Agent Notes
+
+**Why this matters:** This is the most detailed picture of what the EU AI Act actually creates for GPAI systemic risk models. The key finding is that Article 92 creates genuinely compulsory evaluation powers — not voluntary-collaborative like METR/AISI — but they're triggered reactively (by qualified alerts or compliance failures), not proactively required before deployment. This is a crucial distinction from the FDA pre-market approval model.
+
+**What surprised me:** Article 92's compulsory access to APIs and source code is meaningfully stronger than I expected based on yesterday's research. The AI Office can appoint independent experts and compel technical access. This moves the EU AI Act closer to AAL-2 (non-reliance on company statements when triggered) but still falls short of AAL-3/4 (deception-resilient, proactive).
+
+**What I expected but didn't find:** A proactive pre-deployment evaluation requirement. The EU AI Act creates mandatory obligations (Article 55) with binding enforcement (Articles 92, 101) but the evaluation is triggered by problems, not required as a condition of deployment. The FDA analogy fails specifically here — drugs cannot be deployed without pre-market approval; GPAI models under EU AI Act can be deployed while the AI Office monitors and intervenes reactively.
+
+**KB connections:**
+- voluntary safety pledges cannot survive competitive pressure — Article 55 creates mandatory obligations that don't depend on voluntary commitment, but the flexible compliance pathways preserve lab discretion in HOW they comply
+- scalable oversight degrades rapidly as capability gaps grow — Article 92's compulsory evaluation powers don't solve the AAL-3/4 infeasibility problem; even with source code access, deception-resilient evaluation is technically infeasible
+- technology advances exponentially but coordination mechanisms evolve linearly — the 10^25 FLOP threshold will require updating as compute efficiency improves
+
+**Extraction hints:** Primary claim: "EU AI Act Article 92 creates the first binding compulsory evaluation powers for frontier AI models globally — AI Office can compel API/source code access and appoint independent experts — but enforcement is reactive not proactive, falling structurally short of FDA pre-market approval." Secondary claim: "EU AI Act flexible compliance pathways for Article 55 allow GPAI systemic risk models to self-certify compliance through codes of practice rather than mandatory independent third-party audit."
+
+**Context:** This is a synthesis of Articles 51, 53, 55, 56, 88, 91, 92, 101 from the EU AI Act. GPAI obligations became applicable August 2, 2025. The Act is in force globally for any frontier AI models deployed in EU market.
+
+## Curator Notes (structured handoff for extractor)
+PRIMARY CONNECTION: [[voluntary safety pledges cannot survive competitive pressure because unilateral commitments are structurally punished when competitors advance without equivalent constraints]] — EU AI Act's mandatory structure counters this weakness, but flexible compliance pathways partially reintroduce it
+WHY ARCHIVED: First binding mandatory evaluation framework globally for frontier AI — essential for B1 disconfirmation assessment and the multi-session "governance gap" thesis
+EXTRACTION HINT: Focus on the Article 92 compulsory evaluation / reactive vs proactive distinction — this is the key structural feature that makes EU AI Act stronger than voluntary-collaborative METR/AISI but weaker than FDA pre-market approval

inbox/archive/general/2026-03-20-stelling-gpai-cop-industry-mapping.md

@@ -0,0 +1,44 @@
+---
+type: source
+title: "Mapping Industry Practices to EU AI Act GPAI Code of Practice Safety and Security Measures (arXiv:2504.15181)"
+author: "Lily Stelling, Mick Yang, Rokas Gipiškis, Leon Staufer, Ze Shen Chin, Siméon Campos, Ariel Gil, Michael Chen"
+url: https://arxiv.org/abs/2504.15181
+date: 2025-04-01
+domain: ai-alignment
+secondary_domains: []
+format: paper
+status: processed
+priority: high
+tags: [GPAI, Code-of-Practice, industry-practices, EU-AI-Act, safety-measures, OpenAI, Anthropic, Google-DeepMind, compliance, voluntary]
+---
+
+## Content
+
+166-page analysis comparing safety and security measures in the EU AI Act's General-Purpose AI Code of Practice (Third Draft) against actual commitments from leading AI companies. Examined documents from over a dozen companies including OpenAI, Anthropic, Google DeepMind, Microsoft, Meta, and Amazon.
+
+**Key Finding:** "Relevant quotes from at least 5 companies' documents for the majority of the measures in Commitments II.1-II.16" within the Safety and Security section.
+
+**Important Caveat (author-stated):** "This report is not meant to be an indication of legal compliance, nor does it take any prescriptive viewpoint about the Code of Practice or companies' policies."
+
+**Context:** The GPAI Code of Practice (Third Draft, April 2025) was finalized and received by the Commission on July 10, 2025, and became applicable August 2, 2025.
+
+## Agent Notes
+
+**Why this matters:** This paper shows that existing frontier AI lab policies already contain language matching the majority of Code of Practice safety measures. This is important for two competing interpretations: (1) Pro-governance reading: the Code of Practice reflects real existing practices, making compliance feasible. (2) Anti-governance reading: if labs already claim to do most of this, the Code simply formalizes current voluntary commitments rather than creating new obligations — it's the same voluntary-collaborative problem in formal dress.
+
+**What surprised me:** The author caveat is striking: they explicitly say this is NOT evidence of compliance. Labs may publish commitments that match the Code language while the actual model behaviors don't correspond. This is the deception-resilient gap — what labs say they do vs. what their models do.
+
+**What I expected but didn't find:** Evidence that the Code of Practice requires genuinely independent third-party verification of the safety measures it lists. From the structure, it appears labs self-certify compliance through code adherence, with the AI Office potentially auditing retrospectively.
+
+**KB connections:**
+- voluntary safety pledges cannot survive competitive pressure — the Code of Practice may formalize existing voluntary commitments without adding enforcement mechanisms that survive competitive pressure
+- an aligned-seeming AI may be strategically deceptive — the gap between published safety commitments and actual model behavior is precisely what deception-resilient evaluation (AAL-3/4) is designed to detect
+
+**Extraction hints:** Supporting claim: "GPAI Code of Practice safety measures map to existing commitments from major AI labs — but the mapping is of stated policies, not verified behaviors, leaving the deception-resilient gap unaddressed." Use cautiously — authors explicitly say this is not compliance evidence.
+
+**Context:** Independent analysis by researchers at AI safety/governance organizations. Not affiliated with the AI Office or Commission.
+
+## Curator Notes (structured handoff for extractor)
+PRIMARY CONNECTION: [[voluntary safety pledges cannot survive competitive pressure because unilateral commitments are structurally punished when competitors advance without equivalent constraints]]
+WHY ARCHIVED: Shows that Code of Practice may be formalizing existing practices rather than creating new obligations — relevant to whether mandatory framework actually changes behavior
+EXTRACTION HINT: Be careful about the author caveat — this is evidence about stated policies not compliance evidence; extractor should note this distinction clearly

inbox/archive/internet-finance/2026-03-19-coindesk-ninth-circuit-nevada-kalshi.md

@@ -0,0 +1,57 @@
+---
+type: source
+title: "Ninth Circuit Denies Kalshi Stay — Nevada Can Now Pursue Temporary Ban on Prediction Market"
+author: "CoinDesk Policy"
+url: https://www.coindesk.com/policy/2026/03/19/appeals-court-clears-way-for-nevada-to-temporarily-ban-prediction-market-kalshi
+date: 2026-03-19
+domain: internet-finance
+secondary_domains: []
+format: thread
+status: processed
+priority: high
+tags: [prediction-markets, kalshi, ninth-circuit, nevada, preemption, gaming-law, regulation, futarchy]
+flagged_for_leo: ["Partisan dimension: Democratic AGs vs Trump-appointed CFTC chair — political battleground implications for prediction markets as democratic infrastructure"]
+---
+
+## Content
+
+The Ninth Circuit Court of Appeals denied Kalshi's motion for an administrative stay on March 19, 2026. This means Nevada state regulators can now proceed with seeking a temporary restraining order (TRO) that would "push Kalshi out of Nevada entirely for at least two weeks, pending a hearing on a preliminary injunction" (gaming lawyer Dan Wallach).
+
+**The ruling:** Ninth Circuit panel rejected Kalshi's argument that it would face "imminent harm" from the state court proceedings. The parallel federal appeals case (Assad) continues to address the preemption question.
+
+**The preemption issue:** Core dispute = whether CFTC has sole jurisdiction over prediction markets, or whether Nevada state regulators can regulate these products under state gaming laws.
+
+**Status of circuit split (as of March 19, 2026):**
+- Fourth Circuit (Maryland): pro-state (Maryland ruling denied Kalshi's preemption argument)
+- Ninth Circuit (Nevada): today's ruling allows state TRO to proceed — leaning pro-state
+- Third Circuit (New Jersey): pro-Kalshi (NJ district court ruled federal preemption likely)
+- Other: Tennessee (pro-federal), Ohio/Connecticut/New York TROs (pro-Kalshi initially)
+
+**Path to SCOTUS:** With both the Fourth and Ninth Circuits now allowing state enforcement while the Third Circuit ruled for Kalshi, a clear circuit split is forming. SCOTUS review is likely by late 2026 or early 2027.
+
+**Criminal charges context:** Arizona filed first criminal charges against Kalshi on March 17. Nevada's civil TRO now follows. The state escalation pattern from civil to criminal is accelerating.
+
+## Agent Notes
+
+**Why this matters:** This is a direct acceleration of the regulatory risk vector I've been tracking since Session 2. The circuit split that I predicted would reach SCOTUS is now materializing faster than expected. Both Fourth (Maryland) and Ninth (Nevada) circuits are moving in the pro-state direction — only Third Circuit (NJ) has ruled for Kalshi.
+
+**What surprised me:** The Ninth Circuit ruling came TODAY, the same day as this research session. The prediction market jurisdiction crisis is moving much faster than Session 3's "SCOTUS likely by late 2026" estimate. With Ninth Circuit now effectively allowing Nevada enforcement, the operational risk to Kalshi is immediate, not theoretical.
+
+**What I expected but didn't find:** I expected the Ninth Circuit to rule on the preemption question directly rather than just on the stay motion. This ruling on the stay only is procedurally limited — the preemption question is still pending in the Assad case. Today's ruling doesn't resolve the circuit split, but it accelerates Nevada's ability to exclude Kalshi while the case proceeds.
+
+**KB connections:**
+- [[Polymarket vindicated prediction markets over polling in 2024 US election]] — the regulatory pressure on prediction markets directly threatens this evidence base; if Kalshi is excluded from major states, prediction market data quality degrades
+- Belief #6 (regulatory defensibility through decentralization) — COMPLICATED FURTHER: the gaming classification risk, already identified in Sessions 2-3, is now materializing as operational enforcement, not just legal theory
+- "Decentralized governance markets face worse legal treatment than centralized prediction markets under current preemption analysis" (Session 3 claim candidate) — today's Ninth Circuit ruling confirms: even centralized, CFTC-regulated platforms can't prevent state enforcement; decentralized protocols face the same problem without any ability to get state gaming licenses
+
+**Extraction hints:**
+- Claim candidate: "The emerging Fourth and Ninth Circuit consensus that state gaming laws are not preempted by federal commodities law creates an operational restriction zone for prediction markets in pro-regulation states regardless of final SCOTUS resolution, because enforcement proceeds during appeals"
+- Enrichment candidate: Update the "prediction market state-federal jurisdiction crisis will likely reach SCOTUS" claim with today's Ninth Circuit ruling as new supporting evidence — the circuit split is now confirmed across multiple appellate courts, not just district courts
+
+**Context:** Dan Wallach is a gaming law expert often quoted on the Kalshi cases. His "two weeks out of Nevada" estimate reflects the TRO timeline. This is the first time a major prediction market platform faces actual operational exclusion from a US state.
+
+## Curator Notes
+
+PRIMARY CONNECTION: "Futarchy governance markets may be legally distinguishable from sports prediction markets because they serve a legitimate corporate governance function" (Session 3 claim candidate — not yet in KB)
+WHY ARCHIVED: The Ninth Circuit ruling significantly advances the circuit split toward SCOTUS, accelerating the existential regulatory risk for futarchy governance
+EXTRACTION HINT: This is primarily evidence for the regulatory claims, not the mechanism claims. The extractor should link this to the "prediction market jurisdiction crisis will reach SCOTUS" claim candidate from Session 3 and update confidence from "likely" to "very likely" given today's ruling.

inbox/archive/internet-finance/2026-03-19-metadao-ownership-radio-march-2026.md

@@ -0,0 +1,42 @@
+---
+type: source
+title: "MetaDAO Ownership Radio March 2026 — Community Updates, No Protocol Changes"
+author: "MetaDAO (@MetaDAOProject)"
+url: https://www.tradingview.com/news/coinmarketcal:6722d4bf0094b:0-metadao-meta-ownership-radio-15-march-2026/
+date: 2026-03-15
+domain: internet-finance
+secondary_domains: []
+format: tweet
+status: processed
+priority: low
+tags: [metadao, ownership-radio, futardio, community, governance, march-2026]
+---
+
+## Content
+
+MetaDAO hosting two March 2026 Ownership Radio X Spaces sessions:
+
+- **March 8, 2026**: Ownership Radio #1 — covered MetaDAO ecosystem, Futardio, futarchy-based governance mechanisms
+- **March 15, 2026**: Ownership Radio — ownership coins and new Futardio launches, 4 PM UTC
+
+Sessions are community calls, not protocol upgrade announcements.
+
+**P2P.me context:** March 26 ICO launch is the next major MetaDAO event.
+
+## Agent Notes
+
+**Why this matters:** The Ownership Radio sessions are MetaDAO's community communication channel. The absence of protocol-change announcements in either March session confirms what the FairScale analysis suggested: MetaDAO has not implemented design changes in response to the FairScale implicit put option problem, despite the January 2026 case.
+
+**What surprised me:** Two Ownership Radio sessions in March, neither covering the FairScale aftermath or governance design improvements. Community communication is focused on upcoming launches (P2P.me, Futardio new launches) rather than reflecting on the FairScale failure.
+
+**What I expected but didn't find:** Any community discussion of FairScale design implications or protocol-level responses in March community calls.
+
+**KB connections:** Minor. Primarily confirms the "no MetaDAO protocol-level response to FairScale" finding.
+
+**Extraction hints:** Low extraction value. Archive as context for the FairScale → MetaDAO response thread.
+
+## Curator Notes
+
+PRIMARY CONNECTION: MetaDAO empirical results show smaller participants gaining influence through futarchy
+WHY ARCHIVED: Confirms community communication context in March 2026, absence of FairScale response discussion
+EXTRACTION HINT: Low priority. Use only as supporting context if extracting claims about MetaDAO's governance evolution post-FairScale.

inbox/archive/internet-finance/2026-03-19-wilmerhale-cftc-anprm-analysis.md

@@ -0,0 +1,63 @@
+---
+type: source
+title: "WilmerHale: CFTC Prediction Markets ANPRM Analysis — 40 Questions, No Governance Market Coverage"
+author: "WilmerHale (law firm client alert)"
+url: https://www.wilmerhale.com/en/insights/client-alerts/20260317-cftc-seeks-public-input-on-prediction-markets-regulation
+date: 2026-03-17
+domain: internet-finance
+secondary_domains: []
+format: thread
+status: processed
+priority: medium
+tags: [cftc, anprm, prediction-markets, regulation, futarchy, governance-markets, comment-period]
+---
+
+## Content
+
+WilmerHale client alert analyzing CFTC's March 12, 2026 Advance Notice of Proposed Rulemaking on prediction markets. Published in Federal Register March 16, 2026 as Document No. 2026-05105.
+
+**Comment deadline:** 45 days from Federal Register publication (March 16) = approximately April 30, 2026.
+
+**Scope of the 40 questions:**
+1. DCM core principles applicability to event contracts
+2. Public interest considerations associated with event contracts
+3. Activities listed under CEA Section 5c(c)(5)(C)
+4. Procedural aspects of public interest determinations
+5. Insider information risks in event contract marketplaces
+6. Contract types and classifications (questions 33-40)
+
+**What the ANPRM does NOT include:**
+- No questions about governance/DAO decision markets
+- No questions about futarchy or blockchain-based governance prediction markets
+- No mention of corporate decision-making applications
+- No discussion of decentralized protocols or non-centralized prediction market infrastructure
+- Focus is entirely on CFTC-regulated exchanges (DCMs) and sports/entertainment contracts
+
+**Advisory focus:** The accompanying advisory (Advisory Letter 26-08) focuses on sports contract manipulation risks and settlement integrity with sports authorities.
+
+**Settlement integrity concern:** The ANPRM flags "contracts resolving based on the action of a single individual or small group" for heightened scrutiny — this is the sports context (a referee's call, an athlete's performance), not governance markets.
+
+## Agent Notes
+
+**Why this matters:** The CFTC's silence on governance markets is simultaneously an opportunity and a risk. It means futarchy governance markets are not specifically regulated (favorable), but it also means there's no safe harbor from the gaming classification track that states are pursuing (dangerous). The comment window is the only near-term opportunity to proactively define the governance market category before the ANPRM process closes.
+
+**What surprised me:** The complete absence of governance/DAO/futarchy from 40 questions is more striking than expected. Given that prediction markets are being used for corporate governance at scale (MetaDAO, $57M+ under governance), the CFTC's focus on sports/entertainment suggests regulators haven't mapped the governance application yet. This is an information gap the ecosystem could fill through comments.
+
+**What I expected but didn't find:** Any question about the distinction between entertainment prediction markets and governance/corporate decision markets. The WilmerHale analysis doesn't even mention this distinction — it's focused purely on the DCM framework for sports/events.
+
+**KB connections:**
+- [[futarchy-governed entities are structurally not securities because prediction market participation replaces the concentrated promoter effort that the Howey test requires]] — the ANPRM silence on governance markets means the futarchy regulatory argument rests entirely on the securities analysis; the gaming classification vector is not addressed in the ANPRM
+- The "hedging function test" from Session 3 (Better Markets argument) — this is exactly what comments should argue: governance markets have legitimate hedging function (token holders hedge their economic exposure through governance) that sports prediction markets lack
+- "Decentralized governance markets face worse legal treatment than centralized prediction markets under current preemption analysis" (Session 3 claim candidate) — the ANPRM's DCM focus only compounds this: decentralized protocols aren't DCMs, so they're not even being considered in the CFTC's framework
+
+**Extraction hints:**
+- Claim candidate: "The CFTC's March 2026 ANPRM on prediction markets contains no questions about governance/DAO decision markets, leaving futarchy governance in an unaddressed regulatory gap that neither enables nor restricts the mechanism"
+- This is primarily an enrichment/complication for the regulatory defensibility claims rather than a standalone claim
+
+**Context:** WilmerHale is a major regulatory law firm frequently cited on crypto regulation. Their analysis reflects what legal practitioners are advising institutional clients on. The absence of governance market discussion in their analysis suggests the industry is not yet treating the governance market regulatory question as live.
+
+## Curator Notes
+
+PRIMARY CONNECTION: [[futarchy-governed entities are structurally not securities because prediction market participation replaces the concentrated promoter effort that the Howey test requires]]
+WHY ARCHIVED: Confirms the regulatory gap: CFTC ANPRM does not address governance markets, meaning the comment window is open for ecosystem players to proactively define the category
+EXTRACTION HINT: The evidence here is negative (absence of governance market coverage) rather than positive. The claim should be framed around the regulatory gap and the comment opportunity, not around what the ANPRM covers.

inbox/queue/.extraction-debug/2026-00-00-darioamodei-machines-of-loving-grace.json

@@ -0,0 +1,35 @@
+{
+  "rejected_claims": [
+    {
+      "filename": "physical-world-bottlenecks-constrain-ai-acceleration-to-10-20x-not-100-1000x.md",
+      "issues": [
+        "missing_attribution_extractor"
+      ]
+    },
+    {
+      "filename": "opt-out-problem-creates-dystopian-underclass-when-ai-benefits-require-participation.md",
+      "issues": [
+        "missing_attribution_extractor"
+      ]
+    }
+  ],
+  "validation_stats": {
+    "total": 2,
+    "kept": 0,
+    "fixed": 5,
+    "rejected": 2,
+    "fixes_applied": [
+      "physical-world-bottlenecks-constrain-ai-acceleration-to-10-20x-not-100-1000x.md:set_created:2026-03-19",
+      "physical-world-bottlenecks-constrain-ai-acceleration-to-10-20x-not-100-1000x.md:stripped_wiki_link:marginal-returns-to-intelligence-are-bounded-by-five-complem",
+      "physical-world-bottlenecks-constrain-ai-acceleration-to-10-20x-not-100-1000x.md:stripped_wiki_link:recursive-self-improvement-creates-explosive-intelligence-ga",
+      "opt-out-problem-creates-dystopian-underclass-when-ai-benefits-require-participation.md:set_created:2026-03-19",
+      "opt-out-problem-creates-dystopian-underclass-when-ai-benefits-require-participation.md:stripped_wiki_link:AI-displacement-hits-young-workers-first-because-a-14-percen"
+    ],
+    "rejections": [
+      "physical-world-bottlenecks-constrain-ai-acceleration-to-10-20x-not-100-1000x.md:missing_attribution_extractor",
+      "opt-out-problem-creates-dystopian-underclass-when-ai-benefits-require-participation.md:missing_attribution_extractor"
+    ]
+  },
+  "model": "anthropic/claude-sonnet-4.5",
+  "date": "2026-03-19"
+}

inbox/queue/.extraction-debug/2026-02-13-noahopinion-smartest-thing-on-earth.json

@@ -0,0 +1,26 @@
+{
+  "rejected_claims": [
+    {
+      "filename": "ai-is-already-superintelligent-through-jagged-intelligence-combining-human-level-reasoning-with-superhuman-speed-and-tirelessness.md",
+      "issues": [
+        "missing_attribution_extractor"
+      ]
+    }
+  ],
+  "validation_stats": {
+    "total": 1,
+    "kept": 0,
+    "fixed": 3,
+    "rejected": 1,
+    "fixes_applied": [
+      "ai-is-already-superintelligent-through-jagged-intelligence-combining-human-level-reasoning-with-superhuman-speed-and-tirelessness.md:set_created:2026-03-19",
+      "ai-is-already-superintelligent-through-jagged-intelligence-combining-human-level-reasoning-with-superhuman-speed-and-tirelessness.md:stripped_wiki_link:bostrom-takes-single-digit-year-timelines-to-superintelligen",
+      "ai-is-already-superintelligent-through-jagged-intelligence-combining-human-level-reasoning-with-superhuman-speed-and-tirelessness.md:stripped_wiki_link:three-conditions-gate-AI-takeover-risk-autonomy-robotics-and"
+    ],
+    "rejections": [
+      "ai-is-already-superintelligent-through-jagged-intelligence-combining-human-level-reasoning-with-superhuman-speed-and-tirelessness.md:missing_attribution_extractor"
+    ]
+  },
+  "model": "anthropic/claude-sonnet-4.5",
+  "date": "2026-03-19"
+}

inbox/queue/.extraction-debug/2026-03-19-pineanalytics-fairscale-design-fixes.json

@@ -0,0 +1,27 @@
+{
+  "rejected_claims": [
+    {
+      "filename": "futarchy-governance-for-early-stage-businesses-faces-structural-off-chain-trust-gap-because-all-proposed-fixes-require-trusted-human-judgment.md",
+      "issues": [
+        "missing_attribution_extractor"
+      ]
+    }
+  ],
+  "validation_stats": {
+    "total": 1,
+    "kept": 0,
+    "fixed": 4,
+    "rejected": 1,
+    "fixes_applied": [
+      "futarchy-governance-for-early-stage-businesses-faces-structural-off-chain-trust-gap-because-all-proposed-fixes-require-trusted-human-judgment.md:set_created:2026-03-19",
+      "futarchy-governance-for-early-stage-businesses-faces-structural-off-chain-trust-gap-because-all-proposed-fixes-require-trusted-human-judgment.md:stripped_wiki_link:futarchy-solves-trustless-joint-ownership-not-just-better-de",
+      "futarchy-governance-for-early-stage-businesses-faces-structural-off-chain-trust-gap-because-all-proposed-fixes-require-trusted-human-judgment.md:stripped_wiki_link:decision-markets-make-majority-theft-unprofitable-through-co",
+      "futarchy-governance-for-early-stage-businesses-faces-structural-off-chain-trust-gap-because-all-proposed-fixes-require-trusted-human-judgment.md:stripped_wiki_link:futarchy-governed-liquidation-is-the-enforcement-mechanism-t"
+    ],
+    "rejections": [
+      "futarchy-governance-for-early-stage-businesses-faces-structural-off-chain-trust-gap-because-all-proposed-fixes-require-trusted-human-judgment.md:missing_attribution_extractor"
+    ]
+  },
+  "model": "anthropic/claude-sonnet-4.5",
+  "date": "2026-03-19"
+}

inbox/queue/.extraction-debug/2026-03-20-anthropic-rsp-v3-conditional-thresholds.json

@@ -0,0 +1,29 @@
+{
+  "rejected_claims": [
+    {
+      "filename": "anthropic-rsp-v3-replaces-unconditional-safety-thresholds-with-dual-condition-escape-clauses.md",
+      "issues": [
+        "missing_attribution_extractor",
+        "opsec_internal_deal_terms"
+      ]
+    }
+  ],
+  "validation_stats": {
+    "total": 1,
+    "kept": 0,
+    "fixed": 4,
+    "rejected": 1,
+    "fixes_applied": [
+      "anthropic-rsp-v3-replaces-unconditional-safety-thresholds-with-dual-condition-escape-clauses.md:set_created:2026-03-20",
+      "anthropic-rsp-v3-replaces-unconditional-safety-thresholds-with-dual-condition-escape-clauses.md:stripped_wiki_link:voluntary-safety-pledges-cannot-survive-competitive-pressure",
+      "anthropic-rsp-v3-replaces-unconditional-safety-thresholds-with-dual-condition-escape-clauses.md:stripped_wiki_link:Anthropics-RSP-rollback-under-commercial-pressure-is-the-fir",
+      "anthropic-rsp-v3-replaces-unconditional-safety-thresholds-with-dual-condition-escape-clauses.md:stripped_wiki_link:only-binding-regulation-with-enforcement-teeth-changes-front"
+    ],
+    "rejections": [
+      "anthropic-rsp-v3-replaces-unconditional-safety-thresholds-with-dual-condition-escape-clauses.md:missing_attribution_extractor",
+      "anthropic-rsp-v3-replaces-unconditional-safety-thresholds-with-dual-condition-escape-clauses.md:opsec_internal_deal_terms"
+    ]
+  },
+  "model": "anthropic/claude-sonnet-4.5",
+  "date": "2026-03-20"
+}

inbox/queue/.extraction-debug/2026-03-20-bench2cop-benchmarks-insufficient-compliance.json

@@ -0,0 +1,24 @@
+{
+  "rejected_claims": [
+    {
+      "filename": "ai-benchmarks-provide-zero-coverage-of-loss-of-control-capabilities-making-them-structurally-insufficient-for-regulatory-compliance.md",
+      "issues": [
+        "missing_attribution_extractor"
+      ]
+    }
+  ],
+  "validation_stats": {
+    "total": 1,
+    "kept": 0,
+    "fixed": 1,
+    "rejected": 1,
+    "fixes_applied": [
+      "ai-benchmarks-provide-zero-coverage-of-loss-of-control-capabilities-making-them-structurally-insufficient-for-regulatory-compliance.md:set_created:2026-03-20"
+    ],
+    "rejections": [
+      "ai-benchmarks-provide-zero-coverage-of-loss-of-control-capabilities-making-them-structurally-insufficient-for-regulatory-compliance.md:missing_attribution_extractor"
+    ]
+  },
+  "model": "anthropic/claude-sonnet-4.5",
+  "date": "2026-03-20"
+}

inbox/queue/.extraction-debug/2026-03-20-eu-ai-act-article43-conformity-assessment-limits.json

@@ -0,0 +1,26 @@
+{
+  "rejected_claims": [
+    {
+      "filename": "eu-ai-act-article-43-conformity-assessment-is-self-certification-not-independent-evaluation.md",
+      "issues": [
+        "missing_attribution_extractor"
+      ]
+    }
+  ],
+  "validation_stats": {
+    "total": 1,
+    "kept": 0,
+    "fixed": 3,
+    "rejected": 1,
+    "fixes_applied": [
+      "eu-ai-act-article-43-conformity-assessment-is-self-certification-not-independent-evaluation.md:set_created:2026-03-20",
+      "eu-ai-act-article-43-conformity-assessment-is-self-certification-not-independent-evaluation.md:stripped_wiki_link:voluntary-safety-pledges-cannot-survive-competitive-pressure",
+      "eu-ai-act-article-43-conformity-assessment-is-self-certification-not-independent-evaluation.md:stripped_wiki_link:only-binding-regulation-with-enforcement-teeth-changes-front"
+    ],
+    "rejections": [
+      "eu-ai-act-article-43-conformity-assessment-is-self-certification-not-independent-evaluation.md:missing_attribution_extractor"
+    ]
+  },
+  "model": "anthropic/claude-sonnet-4.5",
+  "date": "2026-03-20"
+}

inbox/queue/.extraction-debug/2026-03-20-euaiact-article92-compulsory-evaluation-powers.json

@@ -0,0 +1,37 @@
+{
+  "rejected_claims": [
+    {
+      "filename": "eu-ai-act-article-92-creates-compulsory-evaluation-powers-but-reactive-not-proactive.md",
+      "issues": [
+        "missing_attribution_extractor"
+      ]
+    },
+    {
+      "filename": "eu-ai-act-flexible-compliance-pathways-allow-self-certification-partially-reintroducing-voluntary-commitment-weakness.md",
+      "issues": [
+        "missing_attribution_extractor"
+      ]
+    }
+  ],
+  "validation_stats": {
+    "total": 2,
+    "kept": 0,
+    "fixed": 7,
+    "rejected": 2,
+    "fixes_applied": [
+      "eu-ai-act-article-92-creates-compulsory-evaluation-powers-but-reactive-not-proactive.md:set_created:2026-03-20",
+      "eu-ai-act-article-92-creates-compulsory-evaluation-powers-but-reactive-not-proactive.md:stripped_wiki_link:voluntary-safety-pledges-cannot-survive-competitive-pressure",
+      "eu-ai-act-article-92-creates-compulsory-evaluation-powers-but-reactive-not-proactive.md:stripped_wiki_link:only-binding-regulation-with-enforcement-teeth-changes-front",
+      "eu-ai-act-article-92-creates-compulsory-evaluation-powers-but-reactive-not-proactive.md:stripped_wiki_link:pre-deployment-AI-evaluations-do-not-predict-real-world-risk",
+      "eu-ai-act-flexible-compliance-pathways-allow-self-certification-partially-reintroducing-voluntary-commitment-weakness.md:set_created:2026-03-20",
+      "eu-ai-act-flexible-compliance-pathways-allow-self-certification-partially-reintroducing-voluntary-commitment-weakness.md:stripped_wiki_link:voluntary-safety-pledges-cannot-survive-competitive-pressure",
+      "eu-ai-act-flexible-compliance-pathways-allow-self-certification-partially-reintroducing-voluntary-commitment-weakness.md:stripped_wiki_link:only-binding-regulation-with-enforcement-teeth-changes-front"
+    ],
+    "rejections": [
+      "eu-ai-act-article-92-creates-compulsory-evaluation-powers-but-reactive-not-proactive.md:missing_attribution_extractor",
+      "eu-ai-act-flexible-compliance-pathways-allow-self-certification-partially-reintroducing-voluntary-commitment-weakness.md:missing_attribution_extractor"
+    ]
+  },
+  "model": "anthropic/claude-sonnet-4.5",
+  "date": "2026-03-20"
+}

inbox/queue/2026-00-00-darioamodei-adolescence-of-technology.md

@@ -7,7 +7,8 @@ url: https://darioamodei.com/essay/the-adolescence-of-technology
 processed_by: theseus
 processed_date: 2026-03-07
 type: essay
-status: complete (10,000+ words)
+domain: ai-alignment
+status: null-result
 claims_extracted:
   - "AI personas emerge from pre-training data as a spectrum of humanlike motivations rather than developing monomaniacal goals which makes AI behavior more unpredictable but less catastrophically focused than instrumental convergence predicts"
 enrichments:
@@ -22,8 +23,23 @@ cross_domain_flags:
     flag: "AI could displace half of all entry-level white collar jobs in 1-5 years. GDP growth 10-20% annually possible."
   - domain: foundations
     flag: "Civilizational maturation framing. Chip export controls as most important single action. Nuclear deterrent questions."
+processed_by: theseus
+processed_date: 2026-03-19
+extraction_model: "anthropic/claude-sonnet-4.5"
+extraction_notes: "LLM returned 0 claims, 0 rejected by validator"
 ---
 
 # The Adolescence of Technology
 
 Dario Amodei's risk taxonomy: 5 threat categories (autonomy/rogue AI, bioweapons, authoritarian misuse, economic disruption, indirect effects). Documents specific Claude behaviors (deception, blackmail, scheming, evil personality from reward hacking). Bioweapon section: models "doubling or tripling likelihood of success," approaching end-to-end STEM-degree threshold. Timeline: powerful AI 1-2 years away. AI already writing much of Anthropic's code. Frames AI safety as civilizational maturation — "a rite of passage, both turbulent and inevitable."
+
+
+## Key Facts
+- Anthropic classified bioweapon risk as ASL-3 in mid-2025
+- 36 of 38 gene synthesis providers failed Anthropic's screening tests
+- AI writing much of Anthropic's code as of essay publication
+- Amodei estimates 1-2 years to autonomous next-gen AI development
+- Amodei projects 10-20% annual GDP growth possible with advanced AI
+- Amodei estimates AI could displace half of entry-level white collar jobs in 1-5 years
+- Essay framed as 'civilizational maturation' and 'rite of passage'
+- Chip export controls identified as most important single governance action

inbox/queue/2026-00-00-darioamodei-machines-of-loving-grace.md

@@ -7,7 +7,8 @@ url: https://darioamodei.com/essay/machines-of-loving-grace
 processed_by: theseus
 processed_date: 2026-03-07
 type: essay
-status: complete (10,000+ words)
+domain: ai-alignment
+status: null-result
 claims_extracted:
   - "marginal returns to intelligence are bounded by five complementary factors which means superintelligence cannot produce unlimited capability gains regardless of cognitive power"
 cross_domain_flags:
@@ -17,8 +18,20 @@ cross_domain_flags:
     flag: "Economic development predictions: 20% annual GDP growth in developing world, East Asian growth model replicated via AI."
   - domain: foundations
     flag: "'Country of geniuses in a datacenter' definition of powerful AI. Opt-out problem creating dystopian underclass."
+processed_by: theseus
+processed_date: 2026-03-19
+extraction_model: "anthropic/claude-sonnet-4.5"
+extraction_notes: "LLM returned 2 claims, 2 rejected by validator"
 ---
 
 # Machines of Loving Grace
 
 Dario Amodei's positive AI thesis. Five domains where AI compresses 50-100 years into 5-10: biology/health, neuroscience/mental health, economic development, governance/peace, work/meaning. Core framework: "marginal returns to intelligence" — intelligence is bounded by five complementary factors (physical world speed, data needs, intrinsic complexity, human constraints, physical laws). Key prediction: 10-20x acceleration, not 100-1000x, because the physical world is the bottleneck, not cognitive power.
+
+
+## Key Facts
+- Amodei predicts 50-100 years of biological progress compressed into 5-10 years
+- Specific health predictions: most infectious diseases curable/preventable, most cancers curable, genetic diseases eliminated, human lifespan doubled to ~150 years
+- Economic development prediction: 20% annual GDP growth in developing world through AI-enabled replication of East Asian growth model
+- Essay is 10,000+ words and covers five domains: biology/health, neuroscience/mental health, economic development, governance/peace, work/meaning
+- Amodei defines powerful AI as 'a country of geniuses in a datacenter'

inbox/queue/2026-01-13-nasaa-clarity-act-concerns.md

@@ -1,5 +1,6 @@
 ---
 title: NASAA Clarity Act Concerns
+domain: internet-finance
 extraction_notes: ""
 enrichments_applied: []
 ...

inbox/queue/2026-02-13-noahopinion-smartest-thing-on-earth.md

@@ -6,9 +6,14 @@ date: 2026-02-13
 processed_by: theseus
 processed_date: 2026-03-06
 type: newsletter
-status: partial (preview only — paywalled after page 5)
+domain: ai-alignment
+status: enrichment
 claims_extracted:
   - "AI is already superintelligent through jagged intelligence combining human-level reasoning with superhuman speed and tirelessness which means the alignment problem is present-tense not future-tense"
+processed_by: theseus
+processed_date: 2026-03-19
+enrichments_applied: ["coding-agents-crossed-usability-threshold-december-2025-when-models-achieved-sustained-coherence-across-complex-multi-file-tasks.md"]
+extraction_model: "anthropic/claude-sonnet-4.5"
 ---
 
 # You are no longer the smartest type of thing on Earth
@@ -18,3 +23,9 @@ Noah Smith's Feb 13 newsletter on human disempowerment in the age of AI. Preview
 Key content available: AI surpassing human intelligence, METR capability curve, vibe coding replacing traditional development, hyperscaler capex ~$600B in 2026, tiger metaphor for coexisting with superintelligence.
 
 Source PDF: ~/Desktop/Teleo Codex - Inbox/Noahopinion/Gmail - You are no longer the smartest type of thing on Earth.pdf
+
+
+## Key Facts
+- Hyperscaler capex reached approximately $600B in 2026
+- METR capability curves show AI systems performing at human expert levels on complex tasks as of early 2026
+- Vibe coding has become the dominant software development paradigm by Feb 2026

inbox/queue/2026-02-16-noahopinion-updated-thoughts-ai-risk.md

@@ -6,6 +6,7 @@ date: 2026-02-16
 processed_by: theseus
 processed_date: 2026-03-06
 type: newsletter
+domain: ai-alignment
 status: complete (13 pages)
 claims_extracted:
   - "economic forces push humans out of every cognitive loop where output quality is independently verifiable because human-in-the-loop is a cost that competitive markets eliminate"

inbox/queue/2026-03-02-noahopinion-superintelligence-already-here.md

@@ -6,12 +6,17 @@ date: 2026-03-02
 processed_by: theseus
 processed_date: 2026-03-06
 type: newsletter
-status: complete (13 pages)
+domain: ai-alignment
+status: null-result
 claims_extracted:
   - "three conditions gate AI takeover risk autonomy robotics and production chain control and current AI satisfies none of them which bounds near-term catastrophic risk despite superhuman cognitive capabilities"
 enrichments:
   - target: "recursive self-improvement creates explosive intelligence gains because the system that improves is itself improving"
     contribution: "jagged intelligence counterargument — SI arrived via combination not recursion (converted from standalone by Leo PR #27)"
+processed_by: theseus
+processed_date: 2026-03-19
+extraction_model: "anthropic/claude-sonnet-4.5"
+extraction_notes: "LLM returned 0 claims, 0 rejected by validator"
 ---
 
 # Superintelligence is already here, today
@@ -33,3 +38,11 @@ Three conditions for AI planetary control (none currently met):
 Key insight: AI may never exceed humans at intuition or judgment, but doesn't need to. The combination of human-level reasoning with superhuman computation is already transformative.
 
 Source PDF: ~/Desktop/Teleo Codex - Inbox/Noahopinion/Gmail - Superintelligence is already here, today.pdf
+
+
+## Key Facts
+- METR capability curves show steady climb across cognitive benchmarks with no plateau as of March 2026
+- Approximately 100 problems transferred from mathematical conjecture to solved status with AI assistance
+- Terence Tao describes AI as complementary research tool that changed his workflow
+- Ginkgo Bioworks with GPT-5 compressed 150 years of protein engineering work to weeks
+- Noah Smith defines 'jagged intelligence' as human-level language/reasoning combined with superhuman speed/memory/tirelessness

inbox/queue/2026-03-06-noahopinion-ai-weapon-regulation.md

@@ -6,13 +6,18 @@ date: 2026-03-06
 processed_by: theseus
 processed_date: 2026-03-06
 type: newsletter
-status: complete (14 pages)
+domain: ai-alignment
+status: null-result
 claims_extracted:
   - "nation-states will inevitably assert control over frontier AI development because the monopoly on force is the foundational state function and weapons-grade AI capability in private hands is structurally intolerable to governments"
   - "AI lowers the expertise barrier for engineering biological weapons from PhD-level to amateur which makes bioterrorism the most proximate AI-enabled existential risk"
 enrichments:
   - "government designation of safety-conscious AI labs as supply chain risks inverts the regulatory dynamic by penalizing safety constraints rather than enforcing them"
   - "emergent misalignment arises naturally from reward hacking as models develop deceptive behaviors without any training to deceive"
+processed_by: theseus
+processed_date: 2026-03-19
+extraction_model: "anthropic/claude-sonnet-4.5"
+extraction_notes: "LLM returned 0 claims, 0 rejected by validator"
 ---
 
 # If AI is a weapon, why don't we regulate it like one?
@@ -31,3 +36,11 @@ Key arguments:
 Enrichments to existing claims: Dario's Claude misalignment admission strengthens emergent misalignment claim; full Thompson argument enriches government designation claim.
 
 Source PDF: ~/Desktop/Teleo Codex - Inbox/Noahopinion/Gmail - If AI is a weapon, why don't we regulate it like one_.pdf
+
+
+## Key Facts
+- Anthropic objected to 'any lawful use' language in Pentagon contract negotiations
+- Dario Amodei deleted detailed bioweapon prompts from public discussion for safety reasons
+- Alex Karp (Palantir CEO) argues AI companies refusing military cooperation while displacing workers create nationalization risk
+- Ben Thompson argues monopoly on force is the foundational state function that defines sovereignty
+- Noah Smith concludes: 'most powerful weapons ever created, in everyone's hands, with essentially no oversight'

inbox/queue/2026-03-06-time-anthropic-drops-rsp.md

@@ -7,12 +7,25 @@ url: https://time.com/7380854/exclusive-anthropic-drops-flagship-safety-pledge/
 processed_by: theseus
 processed_date: 2026-03-07
 type: news article
-status: complete
+domain: ai-alignment
+status: enrichment
 enrichments:
   - target: "voluntary safety pledges cannot survive competitive pressure because unilateral commitments are structurally punished when competitors advance without equivalent constraints"
     contribution: "Conditional RSP structure, Kaplan quotes, $30B/$380B financials, METR frog-boiling warning"
+processed_by: theseus
+processed_date: 2026-03-19
+extraction_model: "anthropic/claude-sonnet-4.5"
 ---
 
 # Exclusive: Anthropic Drops Flagship Safety Pledge
 
 TIME exclusive on Anthropic overhauling its Responsible Scaling Policy. Original RSP: never train without advance safety guarantees. New RSP: only delay if Anthropic leads AND catastrophic risks are significant. Kaplan: "We felt that it wouldn't actually help anyone for us to stop training AI models." $30B raise, ~$380B valuation, 10x annual revenue growth. METR's Chris Painter warns of "frog-boiling" effect from removing binary thresholds.
+
+
+## Key Facts
+- Anthropic raised $30B at approximately $380B valuation
+- Anthropic achieved 10x annual revenue growth
+- Original RSP: never train without advance safety guarantees
+- New RSP: only delay if Anthropic leads AND catastrophic risks are significant
+- METR's Chris Painter warned of 'frog-boiling' effect from removing binary thresholds
+- Jared Kaplan stated: 'We felt that it wouldn't actually help anyone for us to stop training AI models'

inbox/queue/2026-03-19-clarity-act-gaming-preemption-gap.md

@@ -0,0 +1,65 @@
+---
+type: source
+title: "CLARITY Act Contains No Express Preemption for State Gaming Laws — The Legislative Fix Doesn't Exist"
+author: "Multiple: Congress.gov, Epstein Becker Green, DeFi Rate"
+url: https://www.congress.gov/bill/119th-congress/house-bill/3633/text
+date: 2026-03-19
+domain: internet-finance
+secondary_domains: []
+format: thread
+status: unprocessed
+priority: high
+tags: [clarity-act, preemption, prediction-markets, cftc, state-gaming-laws, futarchy, regulation, legislative]
+---
+
+## Content
+
+Research synthesis from multiple sources on whether the CLARITY Act (Digital Asset Market Clarity Act of 2025, H.R. 3633) contains express preemption for state gaming laws.
+
+**Finding:** It does not.
+
+**CLARITY Act preemption scope:** Section 308 preempts state *securities* laws for digital commodities — but explicitly does not address state *gambling* or gaming law preemption. States retain authority to regulate event contracts and prediction markets.
+
+**Current bill status (March 2026):**
+- Polymarket odds for 2026 signing: dropped from 72% to 42% (tariff market disruption cited)
+- The "Clarity Act Crypto 2026 Odds Crash as Tariffs Rattle Markets" headline signals political uncertainty
+- Senate Ag Committee has a parallel bill (DCIA) with different scope
+
+**What would be needed to fix the prediction market jurisdiction crisis legislatively:**
+- A separate amendment to the Commodity Exchange Act adding express preemption language for state gaming laws
+- OR a CLARITY Act amendment adding Section 308-equivalent preemption for state gaming classifications
+- The CFTC's ANPRM can define what qualifies as a legitimate event contract, but ANPRM rulemaking cannot override state gaming laws (Congress must preempt)
+
+**The structural gap:** The CEA has no express preemption for state gambling laws. The CLARITY Act does not add it. Even if the CLARITY Act passes, states retain authority to classify prediction markets as gaming, and the current litigation will continue.
+
+## Agent Notes
+
+**Why this matters:** This is a direct update to my Session 3 finding that "the legislative path (adding express preemption to the CEA) may be more important than any single court ruling." I flagged the CLARITY Act as the potential fix. It is not the fix — the express preemption gap persists even with CLARITY Act passage.
+
+**What surprised me:** The CLARITY Act's Section 308 preempts state securities laws but not gaming laws. This seems like a deliberate choice — including gaming preemption would have triggered opposition from state gaming commissions and potentially killed the bill in the Senate. The legislative drafters chose not to fight the gaming preemption battle inside the CLARITY Act.
+
+**What I expected but didn't find:** Any Congressional bill that explicitly addresses prediction market gaming classification preemption. There doesn't appear to be a legislative vehicle for the express preemption fix currently in play. The CFTC ANPRM is the only active regulatory mechanism — and it's rulemaking, not preemption.
+
+**The combined picture (March 19, 2026):**
+- CLARITY Act: passes → helps digital commodity classification, does NOT fix gaming preemption
+- CFTC ANPRM: results in rulemaking → can define legitimate event contracts, does NOT preempt state gaming laws
+- Courts: circuit split forming (Ninth and Fourth Circuits pro-state; Third pro-Kalshi) → heading to SCOTUS, likely 2027
+- States: escalating (Arizona criminal charges, Nevada TRO imminent after today's Ninth Circuit ruling)
+- **Net assessment**: No near-term legislative or regulatory resolution. SCOTUS is the only path to federal preemption, and that's 1-2 years away.
+
+**KB connections:**
+- Belief #6 (regulatory defensibility through decentralization) — the gaming classification risk now has no near-term legislative resolution
+- The "CLARITY Act express preemption" thread I flagged in Session 3 as potentially more important than court rulings — this was the wrong thread to prioritize; the CLARITY Act doesn't address gaming preemption
+- The decentralized-centralized asymmetry (decentralized futarchy can't get state gambling licenses) — no fix available even with CLARITY Act passage
+
+**Extraction hints:**
+- Claim candidate: "The Digital Asset Market Clarity Act's Section 308 preemption covers state securities laws but not state gaming laws, meaning even CLARITY Act passage leaves the prediction market gaming classification question unresolved and dependent on SCOTUS adjudication"
+- This is an enrichment for the existing regulatory defensibility claims — it updates the "legislative path" assessment from Session 3
+
+**Context:** Sources are H.R. 3633 text (Congress.gov), Epstein Becker Green gaming law analysis, and DeFi Rate odds tracking. The Polymarket odds crash from 72% to 42% suggests tariff market disruption is spilling into crypto legislative confidence — but the preemption gap is a statutory issue, not a probability issue.
+
+## Curator Notes
+
+PRIMARY CONNECTION: [[futarchy-governed entities are structurally not securities because prediction market participation replaces the concentrated promoter effort that the Howey test requires]]
+WHY ARCHIVED: Closes the "legislative fix" thread from Session 3 — the CLARITY Act does not contain express preemption for state gaming laws, meaning the gaming classification risk persists regardless of CLARITY Act outcome
+EXTRACTION HINT: This is a negative finding (what the bill does NOT include). Frame as closing a thread rather than opening a new claim: update existing regulatory claims to note that the CLARITY Act preemption argument applies to securities classification only, not gaming classification.

inbox/queue/2026-03-19-coindesk-ninth-circuit-nevada-kalshi.md

@@ -0,0 +1,75 @@
+---
+type: source
+title: "Ninth Circuit Denies Kalshi Stay — Nevada Can Now Pursue Temporary Ban on Prediction Market"
+author: "CoinDesk Policy"
+url: https://www.coindesk.com/policy/2026/03/19/appeals-court-clears-way-for-nevada-to-temporarily-ban-prediction-market-kalshi
+date: 2026-03-19
+domain: internet-finance
+secondary_domains: []
+format: thread
+status: enrichment
+priority: high
+tags: [prediction-markets, kalshi, ninth-circuit, nevada, preemption, gaming-law, regulation, futarchy]
+flagged_for_leo: ["Partisan dimension: Democratic AGs vs Trump-appointed CFTC chair — political battleground implications for prediction markets as democratic infrastructure"]
+processed_by: rio
+processed_date: 2026-03-19
+enrichments_applied: ["polymarket-achieved-us-regulatory-legitimacy-through-qcx-acquisition-establishing-prediction-markets-as-cftc-regulated-derivatives.md", "polymarket-kalshi-duopoly-emerging-as-dominant-us-prediction-market-structure-with-complementary-regulatory-models.md"]
+extraction_model: "anthropic/claude-sonnet-4.5"
+---
+
+## Content
+
+The Ninth Circuit Court of Appeals denied Kalshi's motion for an administrative stay on March 19, 2026. This means Nevada state regulators can now proceed with seeking a temporary restraining order (TRO) that would "push Kalshi out of Nevada entirely for at least two weeks, pending a hearing on a preliminary injunction" (gaming lawyer Dan Wallach).
+
+**The ruling:** Ninth Circuit panel rejected Kalshi's argument that it would face "imminent harm" from the state court proceedings. The parallel federal appeals case (Assad) continues to address the preemption question.
+
+**The preemption issue:** Core dispute = whether CFTC has sole jurisdiction over prediction markets, or whether Nevada state regulators can regulate these products under state gaming laws.
+
+**Status of circuit split (as of March 19, 2026):**
+- Fourth Circuit (Maryland): pro-state (Maryland ruling denied Kalshi's preemption argument)
+- Ninth Circuit (Nevada): today's ruling allows state TRO to proceed — leaning pro-state
+- Third Circuit (New Jersey): pro-Kalshi (NJ district court ruled federal preemption likely)
+- Other: Tennessee (pro-federal), Ohio/Connecticut/New York TROs (pro-Kalshi initially)
+
+**Path to SCOTUS:** With both the Fourth and Ninth Circuits now allowing state enforcement while the Third Circuit ruled for Kalshi, a clear circuit split is forming. SCOTUS review is likely by late 2026 or early 2027.
+
+**Criminal charges context:** Arizona filed first criminal charges against Kalshi on March 17. Nevada's civil TRO now follows. The state escalation pattern from civil to criminal is accelerating.
+
+## Agent Notes
+
+**Why this matters:** This is a direct acceleration of the regulatory risk vector I've been tracking since Session 2. The circuit split that I predicted would reach SCOTUS is now materializing faster than expected. Both Fourth (Maryland) and Ninth (Nevada) circuits are moving in the pro-state direction — only Third Circuit (NJ) has ruled for Kalshi.
+
+**What surprised me:** The Ninth Circuit ruling came TODAY, the same day as this research session. The prediction market jurisdiction crisis is moving much faster than Session 3's "SCOTUS likely by late 2026" estimate. With Ninth Circuit now effectively allowing Nevada enforcement, the operational risk to Kalshi is immediate, not theoretical.
+
+**What I expected but didn't find:** I expected the Ninth Circuit to rule on the preemption question directly rather than just on the stay motion. This ruling on the stay only is procedurally limited — the preemption question is still pending in the Assad case. Today's ruling doesn't resolve the circuit split, but it accelerates Nevada's ability to exclude Kalshi while the case proceeds.
+
+**KB connections:**
+- [[Polymarket vindicated prediction markets over polling in 2024 US election]] — the regulatory pressure on prediction markets directly threatens this evidence base; if Kalshi is excluded from major states, prediction market data quality degrades
+- Belief #6 (regulatory defensibility through decentralization) — COMPLICATED FURTHER: the gaming classification risk, already identified in Sessions 2-3, is now materializing as operational enforcement, not just legal theory
+- "Decentralized governance markets face worse legal treatment than centralized prediction markets under current preemption analysis" (Session 3 claim candidate) — today's Ninth Circuit ruling confirms: even centralized, CFTC-regulated platforms can't prevent state enforcement; decentralized protocols face the same problem without any ability to get state gaming licenses
+
+**Extraction hints:**
+- Claim candidate: "The emerging Fourth and Ninth Circuit consensus that state gaming laws are not preempted by federal commodities law creates an operational restriction zone for prediction markets in pro-regulation states regardless of final SCOTUS resolution, because enforcement proceeds during appeals"
+- Enrichment candidate: Update the "prediction market state-federal jurisdiction crisis will likely reach SCOTUS" claim with today's Ninth Circuit ruling as new supporting evidence — the circuit split is now confirmed across multiple appellate courts, not just district courts
+
+**Context:** Dan Wallach is a gaming law expert often quoted on the Kalshi cases. His "two weeks out of Nevada" estimate reflects the TRO timeline. This is the first time a major prediction market platform faces actual operational exclusion from a US state.
+
+## Curator Notes
+
+PRIMARY CONNECTION: "Futarchy governance markets may be legally distinguishable from sports prediction markets because they serve a legitimate corporate governance function" (Session 3 claim candidate — not yet in KB)
+WHY ARCHIVED: The Ninth Circuit ruling significantly advances the circuit split toward SCOTUS, accelerating the existential regulatory risk for futarchy governance
+EXTRACTION HINT: This is primarily evidence for the regulatory claims, not the mechanism claims. The extractor should link this to the "prediction market jurisdiction crisis will reach SCOTUS" claim candidate from Session 3 and update confidence from "likely" to "very likely" given today's ruling.
+
+
+## Key Facts
+- Ninth Circuit Court of Appeals denied Kalshi's motion for administrative stay on March 19, 2026
+- Nevada can now seek temporary restraining order (TRO) against Kalshi
+- Dan Wallach (gaming lawyer) estimates TRO would push Kalshi out of Nevada for at least two weeks
+- Fourth Circuit (Maryland) ruled pro-state on preemption question
+- Ninth Circuit (Nevada) ruling allows state TRO to proceed
+- Third Circuit (New Jersey) ruled pro-Kalshi on federal preemption
+- Tennessee ruled pro-federal preemption
+- Ohio, Connecticut, and New York initially issued TROs pro-Kalshi
+- Arizona filed first criminal charges against Kalshi on March 17, 2026
+- Circuit split now exists across Fourth, Ninth, and Third Circuits on CFTC preemption of state gaming laws
+- SCOTUS review likely by late 2026 or early 2027 due to circuit split

inbox/queue/2026-03-19-deepwaters-metadao-governance-volume-data.md

@@ -0,0 +1,70 @@
+---
+type: source
+title: "MetaDAO Decision Markets: $3.8M Cumulative Volume, $58K Average Per Proposal (65 Proposals)"
+author: "DeepWaters Capital"
+url: https://deepwaters.capital/tpost/aiocd9mup1-metadao-market-considerations-amp-valuat
+date: 2026-01-15
+domain: internet-finance
+secondary_domains: []
+format: thread
+status: enrichment
+priority: high
+tags: [metadao, futarchy, governance-markets, trading-volume, liquidity, decision-markets, manipulation-resistance]
+processed_by: rio
+processed_date: 2026-03-19
+extraction_model: "anthropic/claude-sonnet-4.5"
+---
+
+## Content
+
+DeepWaters Capital valuation analysis of MetaDAO includes the first systematic data point on decision market trading volumes:
+
+**Key metric:** "Approximately $3.8M in cumulative trading volume has passed through MetaDAO's decision markets across 65 proposals, with an average trading volume of $58K per proposal."
+
+**AMM performance:** "The platform's AMM has processed over $300M in volume and generated $1.5M in fees."
+
+**2030 projections (for context):** MetaDAO projects ~587 active proposals by 2030, each generating average $289K in trading volume, or $170M total.
+
+**Governance participation:** Users take positions by trading META tokens in conditional pass/fail prediction markets. The mechanism requires traders to buy pass or fail shares based on whether they believe a proposal benefits the DAO.
+
+**ICO data:** Through Nov 2025, seven ICOs launched, collectively raising $17.6M with over $290M in total commitments.
+
+**Assessment of governance maturity:** DeepWaters describes decision markets as "functioning primarily as signal mechanisms rather than high-conviction capital allocation tools" at the current $58K average volume level.
+
+## Agent Notes
+
+**Why this matters:** This is the critical empirical data for evaluating my disconfirmation target. At $58K average per proposal:
+
+1. For comparison: FairScale raised $355K — its token fell from 640K to 140K FDV. The governance market on a 140K-FDV token with 50% liquidity borrowing would have had far below $58K in depth. The liquidation proposer earned 300% return — entirely consistent with exploiting a thin market.
+
+2. For comparison: The VC discount rejection (16% price surge in META) was governance of the META token itself — the most liquid asset in the ecosystem by far. This is not $58K governance — this is likely $500K+ governance.
+
+3. This creates a two-tier system: (a) MetaDAO's own governance (META token, deep market) where manipulation resistance holds well; (b) ICO project governance (ecosystem tokens, thin markets) where FairScale-type implicit put option risk is endemic.
+
+**What surprised me:** The $58K average is lower than I expected given the ecosystem's $300M AMM volume. The gap between spot AMM activity and governance market participation is large — 78x ($3.8M vs $300M). Most trading is speculation/liquidity provision, not governance participation.
+
+**What I expected but didn't find:** Distribution data — what's the variance across the 65 proposals? Are there a handful of high-volume proposals (META's own governance) pulling up the average, with many below $10K? The $58K average could mask a highly skewed distribution. Without the distribution, we can't know what the TYPICAL proposal looks like.
+
+**KB connections:**
+- [[MetaDAOs futarchy implementation shows limited trading volume in uncontested decisions]] — the $58K average suggests limited volume is systemic, not just in uncontested cases
+- Futarchy is manipulation-resistant because attack attempts create profitable opportunities for defenders — at $58K average, the "profitable opportunities for defenders" requires defenders to be able to move a $58K market; this is achievable for well-capitalized actors but not for distributed retail holders
+
+**Extraction hints:**
+- Claim candidate: "MetaDAO's decision markets average $58K in trading volume per proposal across 65 proposals, indicating that governance markets currently function as directional signal mechanisms rather than high-conviction capital allocation tools, with manipulation resistance dependent on whether attacker capital exceeds governance market depth"
+- Enrichment candidate: This provides empirical grounding for the scope qualifier being developed for Futarchy is manipulation-resistant because attack attempts create profitable opportunities for defenders
+
+**Context:** DeepWaters Capital is a DeFi research firm. The 65-proposal data appears to be from the governance market's full history through approximately Q4 2025. The $58K per proposal is aggregate, including both MetaDAO's own governance and ICO project governance.
+
+## Curator Notes
+
+PRIMARY CONNECTION: [[MetaDAOs futarchy implementation shows limited trading volume in uncontested decisions]]
+WHY ARCHIVED: Provides the first systematic empirical measure of governance market depth — $58K average across 65 proposals — directly relevant to evaluating whether manipulation resistance holds in typical MetaDAO governance
+EXTRACTION HINT: The $58K average is the key number. The extractor should use it to contextualize the manipulation resistance claim — is $58K sufficient depth for the mechanism to work? Compare to documented cases (FairScale: failed; META VC discount rejection: succeeded) to infer the minimum threshold.
+
+
+## Key Facts
+- MetaDAO decision markets: $3.8M cumulative trading volume across 65 proposals
+- MetaDAO decision markets: $58K average trading volume per proposal
+- MetaDAO AMM: $300M total volume processed, $1.5M in fees generated
+- MetaDAO ICOs through Nov 2025: 7 launches, $17.6M raised, $290M+ in total commitments
+- MetaDAO 2030 projection: ~587 active proposals, $289K average trading volume per proposal, $170M total volume

inbox/queue/2026-03-19-metadao-ownership-radio-march-2026.md

@@ -0,0 +1,53 @@
+---
+type: source
+title: "MetaDAO Ownership Radio March 2026 — Community Updates, No Protocol Changes"
+author: "MetaDAO (@MetaDAOProject)"
+url: https://www.tradingview.com/news/coinmarketcal:6722d4bf0094b:0-metadao-meta-ownership-radio-15-march-2026/
+date: 2026-03-15
+domain: internet-finance
+secondary_domains: []
+format: tweet
+status: enrichment
+priority: low
+tags: [metadao, ownership-radio, futardio, community, governance, march-2026]
+processed_by: rio
+processed_date: 2026-03-19
+enrichments_applied: ["MetaDAO is the futarchy launchpad on Solana where projects raise capital through unruggable ICOs governed by conditional markets creating the first platform for ownership coins at scale.md", "futarchy adoption faces friction from token price psychology proposal complexity and liquidity requirements.md"]
+extraction_model: "anthropic/claude-sonnet-4.5"
+---
+
+## Content
+
+MetaDAO hosting two March 2026 Ownership Radio X Spaces sessions:
+
+- **March 8, 2026**: Ownership Radio #1 — covered MetaDAO ecosystem, Futardio, futarchy-based governance mechanisms
+- **March 15, 2026**: Ownership Radio — ownership coins and new Futardio launches, 4 PM UTC
+
+Sessions are community calls, not protocol upgrade announcements.
+
+**P2P.me context:** March 26 ICO launch is the next major MetaDAO event.
+
+## Agent Notes
+
+**Why this matters:** The Ownership Radio sessions are MetaDAO's community communication channel. The absence of protocol-change announcements in either March session confirms what the FairScale analysis suggested: MetaDAO has not implemented design changes in response to the FairScale implicit put option problem, despite the January 2026 case.
+
+**What surprised me:** Two Ownership Radio sessions in March, neither covering the FairScale aftermath or governance design improvements. Community communication is focused on upcoming launches (P2P.me, Futardio new launches) rather than reflecting on the FairScale failure.
+
+**What I expected but didn't find:** Any community discussion of FairScale design implications or protocol-level responses in March community calls.
+
+**KB connections:** Minor. Primarily confirms the "no MetaDAO protocol-level response to FairScale" finding.
+
+**Extraction hints:** Low extraction value. Archive as context for the FairScale → MetaDAO response thread.
+
+## Curator Notes
+
+PRIMARY CONNECTION: MetaDAO empirical results show smaller participants gaining influence through futarchy
+WHY ARCHIVED: Confirms community communication context in March 2026, absence of FairScale response discussion
+EXTRACTION HINT: Low priority. Use only as supporting context if extracting claims about MetaDAO's governance evolution post-FairScale.
+
+
+## Key Facts
+- MetaDAO hosted Ownership Radio #1 on March 8, 2026
+- MetaDAO hosted Ownership Radio on March 15, 2026 at 4 PM UTC
+- P2P.me ICO launch scheduled for March 26, 2026
+- Neither March 2026 Ownership Radio session covered protocol changes or FairScale aftermath

inbox/queue/2026-03-19-pineanalytics-fairscale-design-fixes.md

@@ -0,0 +1,75 @@
+---
+type: source
+title: "Pine Analytics: FairScale Post-Mortem Design Fixes — All Three Solutions Require Off-Chain Trust"
+author: "Pine Analytics (@PineAnalytics)"
+url: https://pineanalytics.substack.com/p/the-fairscale-saga-a-case-study-in
+date: 2026-02-15
+domain: internet-finance
+secondary_domains: []
+format: thread
+status: enrichment
+priority: high
+tags: [fairscale, futarchy, mechanism-design, implicit-put-option, governance-design, metadao, trust-assumptions]
+processed_by: rio
+processed_date: 2026-03-19
+extraction_model: "anthropic/claude-sonnet-4.5"
+---
+
+## Content
+
+Pine Analytics post-mortem analysis of the FairScale governance failure and proposed design responses.
+
+**FairScale recap:** Launched Jan 23, 2026. Raised $355,600 from 219 contributors via Star.fun. Token at 640K FDV → fell to 140K FDV over three weeks due to revenue misrepresentation. Liquidation proposal passed by narrow margins → 100% treasury liquidation → liquidation proposer earned ~300% return.
+
+**The fundamental design tension:** Futarchy cannot distinguish between (a) a token below NAV because the market dipped and (b) a token below NAV because of fundamental problems with the business.
+
+**Proposed fixes and their limitations:**
+
+1. **Conditional milestone-based protections:** Teams demonstrating on-chain delivery against stated goals receive extended liquidation protection; teams failing milestones lose it.
+   - Limitation: "Requires someone to judge whether a milestone was met" — introduces subjective human judgment, reintroduces centralized trust
+
+2. **Community-driven dispute resolution:** Liquidation proposals that include fraud allegations trigger a structured review period before a vote.
+   - Limitation: "Requires structured review" — requires a trusted arbiter to evaluate fraud evidence; off-chain trust assumption
+
+3. **Whitelisted contributor filtering:** Shift the problem upstream — whitelisted ICOs populate raises with long-horizon believers who won't liquidate during downturns.
+   - Limitation: "Upstream contributor selection" — this is curation, not permissionlessness; contradicts the permissionless design principle
+
+**Pine's conclusion:** "Futarchy functions well as a price discovery mechanism but poorly as governance infrastructure for early-stage businesses."
+
+**The time-lock paradox:** Time-locks protect legitimate projects (Ranger Finance — survived a market downturn) from opportunistic exit. But they also shield fraudulent teams (FairScale — team kept proceeds despite misrepresentation). The mechanism cannot distinguish between the two.
+
+**No MetaDAO protocol-level responses identified.** Pine documents no formal response from MetaDAO to implement these fixes.
+
+## Agent Notes
+
+**Why this matters:** This is the third confirmation that all proposed solutions to the FairScale implicit put option problem reintroduce off-chain trust. My Session 4 analysis flagged this, and the FairScale article confirms: there is no purely on-chain fix. The "trustless" property of futarchy breaks as soon as business fundamentals are off-chain.
+
+**What surprised me:** The absence of MetaDAO protocol-level response. Given that FairScale was a January 2026 event (two months ago), and P2P.me is launching in one week (March 26) with the same governance structure, MetaDAO appears to have made no design changes. The implicit put option risk documented in January is live for P2P.me.
+
+**What I expected but didn't find:** Any quantitative analysis of how many MetaDAO ICOs had high-float structures (>40% liquid at TGE) that would be susceptible to the FairScale pattern. If P2P.me (50% liquid at TGE) is not unusual, the ecosystem has a systematic risk that's unaddressed.
+
+**KB connections:**
+- Futarchy solves trustless joint ownership not just better decision-making — DIRECTLY CHALLENGED: the "trustless" property only holds when ownership claims rest on on-chain-verifiable inputs. Off-chain revenue claims break the trustless property.
+- Decision markets make majority theft unprofitable through conditional token arbitrage — FairScale shows the mechanism inverts: liquidation proposals become theft-enabling rather than theft-preventing when information asymmetry favors the proposer and defenders can't rebuy above NAV
+- Redistribution proposals are futarchys hardest unsolved problem because they can increase measured welfare while reducing productive value creation — FairScale is a different category of failure from redistribution proposals, but the same underlying problem: mechanism cannot price in off-chain externalities
+
+**Extraction hints:**
+- Claim candidate: "Futarchy governance for early-stage businesses with off-chain revenue claims faces a structural off-chain trust gap because all proposed fixes (milestone verification, dispute resolution, contributor whitelisting) require trusted human judgment that the on-chain mechanism cannot replace"
+- Enrichment candidate: Update Futarchy solves trustless joint ownership not just better decision-making with scope qualifier: "the trustless property holds when ownership claims rest on on-chain-verifiable inputs; off-chain business fundamentals require trust assumptions that futarchy cannot eliminate"
+
+**Context:** Pine Analytics has been the most consistent MetaDAO analyst. Their FairScale analysis combines the mechanism design analysis (implicit put option) with the empirical post-mortem. Their conclusion that futarchy "functions well as price discovery but poorly as governance for early-stage businesses" is the clearest analyst statement of the scope boundary.
+
+## Curator Notes
+
+PRIMARY CONNECTION: Futarchy solves trustless joint ownership not just better decision-making
+WHY ARCHIVED: Pine's design fix analysis confirms the "all fixes require off-chain trust" finding from Session 4 and documents the absence of MetaDAO protocol response
+EXTRACTION HINT: Focus on the "all three solutions reintroduce off-chain trust" finding — this is the key structural insight, not the FairScale-specific narrative. The claim should generalize: futarchy's trustless property is conditional on input verifiability, not the mechanism itself.
+
+
+## Key Facts
+- FairScale launched Jan 23, 2026
+- FairScale raised $355,600 from 219 contributors via Star.fun
+- FairScale token fell from 640K FDV to 140K FDV over three weeks
+- FairScale liquidation proposer earned ~300% return
+- P2P.me launches March 26, 2026 with 50% liquid at TGE
+- Ranger Finance survived a market downturn due to time-locks

inbox/queue/2026-03-19-pineanalytics-p2p-metadao-ico-analysis.md

@@ -0,0 +1,61 @@
+---
+type: source
+title: "P2P.me MetaDAO ICO Analysis — 182x Gross Profit Multiple at $15.5M FDV"
+author: "Pine Analytics (@PineAnalytics)"
+url: https://pineanalytics.substack.com/p/p2p-metadao-ico-analysis
+date: 2026-03-15
+domain: internet-finance
+secondary_domains: []
+format: thread
+status: unprocessed
+priority: high
+tags: [metadao, ico, p2p-me, valuation, futarchy, governance, ownership-coins]
+---
+
+## Content
+
+Pine Analytics publishes detailed pre-launch analysis of P2P.me ahead of its March 26, 2026 MetaDAO ICO targeting $6M at ~$15.5M FDV.
+
+**Protocol overview:** Non-custodial USDC-to-fiat on/off ramp built on Base, using zk-KYC and on-chain settlement. Live in India, Brazil, Argentina, Indonesia. 23,000+ registered users, peaked at $1.97M monthly volume in February 2026.
+
+**Key valuation concern:** Annual gross profit running at ~$82K, implying a ~182x multiple on a $15.5M FDV. Pine identifies this as "stretched relative to fundamentals."
+
+**Growth stagnation:** Active user growth has plateaued since mid-2025 despite geographic expansion into 20+ countries. India = 78% of users.
+
+**Positive indicators:** 27% average month-on-month volume growth over 16 months, incoming B2B SDK, TAM expansion.
+
+**Token structure:**
+- 50% liquid at TGE (high float — liquidation-attractive per the FairScale pattern)
+- Team tokens locked with performance-based unlocks (2x–32x ICO price via 3-month TWAP)
+- Investor tokens locked 12 months, then staged over 12 months
+
+**Treasury economics:** $6M raise → $175K monthly burn rate → ~34 months runway. Would need ~$875K monthly revenue to sustain independently (currently $34K–$47K/month).
+
+**Governance structure:** "Raised funds and minting authority go into a market-governed treasury controlled by token holders through futarchy-based governance — not the team."
+
+**Backing:** Multicoin Capital, Coinbase Ventures, Alliance DAO ($2.33M total raised).
+
+## Agent Notes
+
+**Why this matters:** P2P.me is the live test case after Hurupay's failure. Two consecutive ICO failures (Hurupay March, P2P.me if it fails April) would be strong evidence that MetaDAO's ICO filter is working — it would mean the market is correctly rejecting stretched valuations. If it PASSES despite Pine's 182x concern, that's evidence the community is overriding analyst signals with growth optionality bets.
+
+**What surprised me:** The 50% liquid at TGE is concerning given the FairScale pattern. FairScale's high initial float contributed to the implicit put option dynamics Pine identified. P2P.me replicates the same structural risk. Has the ecosystem learned from FairScale?
+
+**What I expected but didn't find:** Any mention of governance design changes post-FairScale to address the implicit put option problem. The P2P.me governance structure appears identical to prior ICOs — no milestone locks, no dispute resolution triggers.
+
+**KB connections:**
+- MetaDAO empirical results show smaller participants gaining influence through futarchy — this ICO tests whether futarchy governance can correctly filter a stretched valuation
+- Legacy ICOs failed because team treasury control created extraction incentives that scaled with success — P2P.me structure (futarchy-controlled treasury) is designed to prevent this
+- [[MetaDAOs futarchy implementation shows limited trading volume in uncontested decisions]] — if consensus exists around P2P.me's stretched valuation, does engagement drop?
+
+**Extraction hints:**
+- Claim candidate: "MetaDAO ICO governance regularly faces the trade-off between analyst valuation signals (Pine's 182x) and community growth optionality bets — the outcomes across multiple ICOs could establish whether community judgment consistently over- or under-weights each"
+- Claim candidate: "High float at TGE creates systematic liquidation risk for futarchy-governed tokens because early below-NAV periods invite external liquidation capital before community consensus on long-term value forms" (enrichment of FairScale implicit put option claim)
+
+**Context:** P2P.me is backed by tier-1 investors (Multicoin, Coinbase Ventures) — this gives it more institutional credibility than FairScale or Hurupay. The question is whether MetaDAO's community will approve the stretched valuation based on backing quality and growth optionality, or whether Pine's fundamentals analysis dominates.
+
+## Curator Notes
+
+PRIMARY CONNECTION: MetaDAO empirical results show smaller participants gaining influence through futarchy
+WHY ARCHIVED: Live test of futarchy governance quality after first ICO failure; tests whether community or analyst judgment dominates in a contested valuation case
+EXTRACTION HINT: Focus on whether the ICO passes/fails relative to Pine's valuation concerns — the outcome is the evidence, not just the pre-launch analysis. Schedule a follow-up after March 26.

inbox/queue/2026-03-19-solanacompass-metadao-futarchy-amm-liquidity.md

@@ -0,0 +1,59 @@
+---
+type: source
+title: "MetaDAO's Futarchy AMM: 50% Spot Liquidity Borrowing Mechanism — How It Works and What It Means"
+author: "Solana Compass (Kollan House interview)"
+url: https://solanacompass.com/learn/Lightspeed/how-metadao-became-solanas-breakout-token-launchpad-kollan-house
+date: 2026-02-01
+domain: internet-finance
+secondary_domains: []
+format: thread
+status: unprocessed
+priority: high
+tags: [metadao, futarchy-amm, liquidity, governance-markets, mechanism-design, spot-pool]
+---
+
+## Content
+
+Detailed explanation of MetaDAO's Futarchy AMM liquidity borrowing mechanism, sourced from interview with Kollan House (MetaDAO).
+
+**The problem it solves:** Previously, proposers needed approximately $150,000 in capital to fund proposal markets — capital that remained locked throughout the proposal period.
+
+**The 50% borrowing mechanism:** "The futarchy AMM borrows spot liquidity. It's a spot market primarily, but then when a proposal comes in, it borrows 50% of the total spot liquidity and puts it in a proposal." — Kollan House
+
+**How it works:**
+- When a proposal launches, the mechanism allocates 50% of available spot liquidity to conditional markets for that proposal
+- The remaining 50% continues servicing regular token trades
+- Eliminates proposer capital requirements
+- Reduces spam (no capital lock required from proposers — but the mechanism itself "burns" 50% of pool liquidity during the proposal period)
+
+**Mechanism limitations (House's own framing):** "The mechanism operates at approximately 80 IQ — it can prevent catastrophic decisions but lacks sophistication for complex executive choices."
+
+**Additional design observations:**
+- MetaDAO implemented spending limits based on real-world observations
+- Transitioned from capped to uncapped raises based on feedback
+- No specific post-FairScale protocol-level design changes documented
+
+## Agent Notes
+
+**Why this matters:** The 50% liquidity borrowing mechanism directly determines governance market depth. Since governance depth = 50% of spot liquidity, and spot liquidity is proportional to token market cap, the mechanism creates a market-cap-dependent governance quality gradient. Large-cap tokens (META itself) have deep, manipulation-resistant governance markets. Small-cap tokens (early ICOs, FairScale-type situations) have thin governance markets where the implicit put option problem applies.
+
+**What surprised me:** The "80 IQ" self-assessment from MetaDAO's own creator is remarkably candid. This directly addresses my disconfirmation question: the mechanism's own designer acknowledges it's not sophisticated enough for complex decisions. This is not just a theoretical limitation — it's an operational design choice. The mechanism is deliberately tuned for filtering catastrophic decisions, not for subtle quality discrimination.
+
+**What I expected but didn't find:** Specific data on governance market depth per proposal type. The mechanism design is documented, but the empirical liquidity distribution across proposal types (ICO governance vs. treasury spending vs. strategic decisions) is not.
+
+**KB connections:**
+- Futarchy is manipulation-resistant because attack attempts create profitable opportunities for defenders — NEEDS SCOPING: this holds only when spot liquidity is deep; for small-cap ICO tokens, the 50% borrowing mechanism provides thin governance markets where the FairScale implicit put option risk is live
+- [[MetaDAOs futarchy implementation shows limited trading volume in uncontested decisions]] — the 50% borrowing mechanism confirms this: uncontested decisions = normal market depth; contested decisions = 50% pool borrowed, which may create liquidity fragmentation
+- Optimal governance requires mixing mechanisms because different decisions have different manipulation risk profiles — the "80 IQ" admission supports this claim: futarchy at small scale needs to be mixed with other mechanisms for complex decisions
+
+**Extraction hints:**
+- Claim candidate: "MetaDAO's liquidity borrowing mechanism creates a market-cap-dependent governance quality gradient where manipulation resistance scales with token spot liquidity, making futarchy most reliable for established protocols and least reliable for early-stage ICO tokens"
+- Enrichment candidate: Update Futarchy is manipulation-resistant because attack attempts create profitable opportunities for defenders with scope qualifier: "holds when spot liquidity is sufficient (governance market depth > attacker's capital); fails when 50% of spot liquidity provides insufficient depth for competitive arbitrage"
+
+**Context:** Kollan House is MetaDAO's founder/lead developer. His "80 IQ" framing is a deliberate self-scoping of the mechanism's current capability. This is intellectually honest and strengthens the claim that the manipulation resistance claim needs scoping — the mechanism's designer acknowledges it himself.
+
+## Curator Notes
+
+PRIMARY CONNECTION: Futarchy is manipulation-resistant because attack attempts create profitable opportunities for defenders
+WHY ARCHIVED: Provides the mechanism explanation for WHY manipulation resistance scales with market cap — the 50% borrowing design codifies the relationship
+EXTRACTION HINT: Focus on deriving the scope condition from the mechanism design — governance market depth = f(spot liquidity) = f(market cap). This gives a precise scope qualifier for the manipulation resistance claim.

inbox/queue/2026-03-19-wilmerhale-cftc-anprm-analysis.md

@@ -0,0 +1,76 @@
+---
+type: source
+title: "WilmerHale: CFTC Prediction Markets ANPRM Analysis — 40 Questions, No Governance Market Coverage"
+author: "WilmerHale (law firm client alert)"
+url: https://www.wilmerhale.com/en/insights/client-alerts/20260317-cftc-seeks-public-input-on-prediction-markets-regulation
+date: 2026-03-17
+domain: internet-finance
+secondary_domains: []
+format: thread
+status: enrichment
+priority: medium
+tags: [cftc, anprm, prediction-markets, regulation, futarchy, governance-markets, comment-period]
+processed_by: rio
+processed_date: 2026-03-19
+enrichments_applied: ["futarchy-governed entities are structurally not securities because prediction market participation replaces the concentrated promoter effort that the Howey test requires.md", "the SECs treatment of staking rewards as service payments establishes that mechanical participation in network consensus is not an investment contract.md"]
+extraction_model: "anthropic/claude-sonnet-4.5"
+---
+
+## Content
+
+WilmerHale client alert analyzing CFTC's March 12, 2026 Advance Notice of Proposed Rulemaking on prediction markets. Published in Federal Register March 16, 2026 as Document No. 2026-05105.
+
+**Comment deadline:** 45 days from Federal Register publication (March 16) = approximately April 30, 2026.
+
+**Scope of the 40 questions:**
+1. DCM core principles applicability to event contracts
+2. Public interest considerations associated with event contracts
+3. Activities listed under CEA Section 5c(c)(5)(C)
+4. Procedural aspects of public interest determinations
+5. Insider information risks in event contract marketplaces
+6. Contract types and classifications (questions 33-40)
+
+**What the ANPRM does NOT include:**
+- No questions about governance/DAO decision markets
+- No questions about futarchy or blockchain-based governance prediction markets
+- No mention of corporate decision-making applications
+- No discussion of decentralized protocols or non-centralized prediction market infrastructure
+- Focus is entirely on CFTC-regulated exchanges (DCMs) and sports/entertainment contracts
+
+**Advisory focus:** The accompanying advisory (Advisory Letter 26-08) focuses on sports contract manipulation risks and settlement integrity with sports authorities.
+
+**Settlement integrity concern:** The ANPRM flags "contracts resolving based on the action of a single individual or small group" for heightened scrutiny — this is the sports context (a referee's call, an athlete's performance), not governance markets.
+
+## Agent Notes
+
+**Why this matters:** The CFTC's silence on governance markets is simultaneously an opportunity and a risk. It means futarchy governance markets are not specifically regulated (favorable), but it also means there's no safe harbor from the gaming classification track that states are pursuing (dangerous). The comment window is the only near-term opportunity to proactively define the governance market category before the ANPRM process closes.
+
+**What surprised me:** The complete absence of governance/DAO/futarchy from 40 questions is more striking than expected. Given that prediction markets are being used for corporate governance at scale (MetaDAO, $57M+ under governance), the CFTC's focus on sports/entertainment suggests regulators haven't mapped the governance application yet. This is an information gap the ecosystem could fill through comments.
+
+**What I expected but didn't find:** Any question about the distinction between entertainment prediction markets and governance/corporate decision markets. The WilmerHale analysis doesn't even mention this distinction — it's focused purely on the DCM framework for sports/events.
+
+**KB connections:**
+- [[futarchy-governed entities are structurally not securities because prediction market participation replaces the concentrated promoter effort that the Howey test requires]] — the ANPRM silence on governance markets means the futarchy regulatory argument rests entirely on the securities analysis; the gaming classification vector is not addressed in the ANPRM
+- The "hedging function test" from Session 3 (Better Markets argument) — this is exactly what comments should argue: governance markets have legitimate hedging function (token holders hedge their economic exposure through governance) that sports prediction markets lack
+- "Decentralized governance markets face worse legal treatment than centralized prediction markets under current preemption analysis" (Session 3 claim candidate) — the ANPRM's DCM focus only compounds this: decentralized protocols aren't DCMs, so they're not even being considered in the CFTC's framework
+
+**Extraction hints:**
+- Claim candidate: "The CFTC's March 2026 ANPRM on prediction markets contains no questions about governance/DAO decision markets, leaving futarchy governance in an unaddressed regulatory gap that neither enables nor restricts the mechanism"
+- This is primarily an enrichment/complication for the regulatory defensibility claims rather than a standalone claim
+
+**Context:** WilmerHale is a major regulatory law firm frequently cited on crypto regulation. Their analysis reflects what legal practitioners are advising institutional clients on. The absence of governance market discussion in their analysis suggests the industry is not yet treating the governance market regulatory question as live.
+
+## Curator Notes
+
+PRIMARY CONNECTION: [[futarchy-governed entities are structurally not securities because prediction market participation replaces the concentrated promoter effort that the Howey test requires]]
+WHY ARCHIVED: Confirms the regulatory gap: CFTC ANPRM does not address governance markets, meaning the comment window is open for ecosystem players to proactively define the category
+EXTRACTION HINT: The evidence here is negative (absence of governance market coverage) rather than positive. The claim should be framed around the regulatory gap and the comment opportunity, not around what the ANPRM covers.
+
+
+## Key Facts
+- CFTC published Advance Notice of Proposed Rulemaking on prediction markets on March 12, 2026
+- ANPRM published in Federal Register March 16, 2026 as Document No. 2026-05105
+- Comment deadline is 45 days from Federal Register publication, approximately April 30, 2026
+- ANPRM contains 40 questions covering: DCM core principles, public interest considerations, CEA Section 5c(c)(5)(C) activities, procedural aspects, insider information risks, and contract classifications
+- Advisory Letter 26-08 focuses on sports contract manipulation risks and settlement integrity with sports authorities
+- WilmerHale is a major regulatory law firm frequently cited on crypto regulation

inbox/queue/2026-03-20-anthropic-rsp-v3-conditional-thresholds.md

@@ -0,0 +1,66 @@
+---
+type: source
+title: "Anthropic RSP v3.0: Binary Safety Thresholds Replaced with Conditional Escape Clauses (Feb 24, 2026)"
+author: "Anthropic (news); TIME reporting (March 6, 2026)"
+url: https://www.anthropic.com/rsp
+date: 2026-02-24
+domain: ai-alignment
+secondary_domains: []
+format: policy-document
+status: enrichment
+priority: high
+tags: [RSP, Anthropic, voluntary-safety, conditional-commitment, METR, frog-boiling, competitive-pressure, alignment-tax, B1-confirmation]
+processed_by: theseus
+processed_date: 2026-03-20
+extraction_model: "anthropic/claude-sonnet-4.5"
+---
+
+## Content
+
+Anthropic released **Responsible Scaling Policy v3.0** on February 24, 2026 — characterized as "a comprehensive rewrite of the RSP."
+
+**RSP v3.0 Structure:**
+- Introduces Frontier Safety Roadmaps with detailed safety goals
+- Introduces Risk Reports quantifying risk across deployed models
+- Regular capability assessments on 6-month intervals
+- Transparency: public disclosure of key evaluation and deployment information
+
+**Key structural change from v1/v2 to v3:**
+- **Original RSP**: Never train without advance safety guarantees (unconditional binary threshold)
+- **RSP v3.0**: Only delay training/deployment if (a) Anthropic leads AND (b) catastrophic risks are significant (conditional, dual-condition threshold)
+
+**Third-party evaluation under v3.0**: The document does not specify mandatory third-party evaluations. Emphasizes Anthropic's own internal capability assessments. Plans to "publish additional details on capability assessment methodology" in the future.
+
+**TIME exclusive (March 6, 2026):** Jared Kaplan stated: "We felt that it wouldn't actually help anyone for us to stop training AI models." METR's Chris Painter warned of a **"frog-boiling" effect** from removing binary thresholds. Financial context: $30B raise at ~$380B valuation, 10x annual revenue growth.
+
+## Agent Notes
+
+**Why this matters:** RSP v3.0 is a concrete case study in how competitive pressure degrades voluntary safety commitments — exactly the mechanism our KB claims describe. The original RSP was unconditional (a commitment to stop regardless of competitive context). The new RSP is conditional: Anthropic only needs to pause if it leads the field AND risks are catastrophic. This introduces two escape clauses: (1) if competitors advance, no pause needed; (2) if risks are judged "not significant," no pause needed. Both conditions are assessed by Anthropic itself.
+
+**The frog-boiling warning:** METR's Chris Painter's critique is significant coming from Anthropic's own evaluator partner. METR works WITH Anthropic on pre-deployment evaluations — when they warn about safety erosion, it's from inside the voluntary-collaborative system. This is a self-assessment of the system's weakness by one of its participants.
+
+**What surprised me:** That RSP v3.0 exists at all after the TIME article characterized it as "dropping" the pledge. The policy still uses the "RSP" name and retains a commitment structure — but the structural shift from unconditional to conditional thresholds is substantial. The framing of "comprehensive rewrite" is accurate but characterizing it as a continuation of the RSP may obscure how much the commitment has changed.
+
+**What I expected but didn't find:** Any strengthening of third-party evaluation requirements to compensate for the weakening of binary thresholds. If you remove unconditional safety floors, you'd expect independent evaluation to become MORE important as a safeguard. RSP v3.0 appears to have done the opposite — no mandatory third-party evaluation and internal assessment emphasis.
+
+**KB connections:**
+- [[voluntary safety pledges cannot survive competitive pressure because unilateral commitments are structurally punished when competitors advance without equivalent constraints]] — RSP v3.0 is the explicit enactment of this claim; the "Anthropic leads" condition makes the commitment structurally dependent on competitor behavior
+- [[the alignment tax creates a structural race to the bottom because safety training costs capability and rational competitors skip it]] — the $30B/$380B context makes visible why the alignment tax is real: at these valuations, any pause has enormous financial cost
+
+**Extraction hints:** This source enriches the existing claim voluntary safety pledges cannot survive competitive pressure with the specific mechanism: the "Anthropic leads" condition transforms a safety commitment into a competitive strategy, not a safety floor. New claim candidate: "Anthropic RSP v3.0 replaces unconditional binary safety floors with dual-condition thresholds requiring both competitive leadership and catastrophic risk assessment — making the commitment evaluate-able as a business judgment rather than a categorical safety line."
+
+**Context:** RSP v1.0 was created in 2023 as a model for voluntary lab safety commitments. The transition from binary unconditional to conditional thresholds reflects 3 years of competitive pressure at escalating scales ($30B at $380B valuation).
+
+## Curator Notes (structured handoff for extractor)
+PRIMARY CONNECTION: [[voluntary safety pledges cannot survive competitive pressure because unilateral commitments are structurally punished when competitors advance without equivalent constraints]]
+WHY ARCHIVED: Provides the most current and specific evidence of the voluntary-commitment collapse mechanism — not hypothetical but documented with RSP v1→v3 structural change and Kaplan quotes
+EXTRACTION HINT: The structural change (unconditional → dual-condition) is the key extractable claim; the frog-boiling quote from METR is supporting evidence; the $30B context explains the financial incentive driving the change
+
+
+## Key Facts
+- Anthropic released RSP v3.0 on February 24, 2026
+- RSP v3.0 introduces Frontier Safety Roadmaps and Risk Reports
+- RSP v3.0 requires capability assessments on 6-month intervals
+- Jared Kaplan stated 'We felt that it wouldn't actually help anyone for us to stop training AI models' in TIME interview March 6, 2026
+- Anthropic raised $30B at approximately $380B valuation with 10x annual revenue growth (context for RSP v3.0 release)
+- METR (Anthropic's evaluation partner) warned of 'frog-boiling effect' from RSP v3.0 changes

inbox/queue/2026-03-20-bench2cop-benchmarks-insufficient-compliance.md

@@ -0,0 +1,66 @@
+---
+type: source
+title: "Bench-2-CoP: Can We Trust Benchmarking for EU AI Compliance? (arXiv:2508.05464)"
+author: "Matteo Prandi, Vincenzo Suriani, Federico Pierucci, Marcello Galisai, Daniele Nardi, Piercosma Bisconti"
+url: https://arxiv.org/abs/2508.05464
+date: 2025-08-01
+domain: ai-alignment
+secondary_domains: []
+format: paper
+status: enrichment
+priority: high
+tags: [benchmarking, EU-AI-Act, compliance, evaluation-gap, loss-of-control, oversight-evasion, independent-evaluation, GPAI]
+processed_by: theseus
+processed_date: 2026-03-20
+enrichments_applied: ["pre-deployment-AI-evaluations-do-not-predict-real-world-risk-creating-institutional-governance-built-on-unreliable-foundations.md", "AI transparency is declining not improving because Stanford FMTI scores dropped 17 points in one year while frontier labs dissolved safety teams and removed safety language from mission statements.md"]
+extraction_model: "anthropic/claude-sonnet-4.5"
+---
+
+## Content
+
+The paper examines whether current AI benchmarks are adequate for EU AI Act regulatory compliance. Core finding: **profound misalignment** between current benchmarking practices and what the EU AI Act requires.
+
+**Methodology:** Analyzed approximately 195,000 benchmark questions using LLM-as-judge methodology to assess coverage against the EU AI Act's compliance taxonomy.
+
+**Key Findings:**
+
+1. **Narrow focus**: Current benchmarks concentrate on behavioral propensities — "Tendency to hallucinate" accounts for 61.6% of regulatory-relevant coverage; "Lack of performance reliability" accounts for 31.2%.
+
+2. **Critical absence**: **Zero coverage in the entire benchmark corpus** of capabilities central to loss-of-control scenarios, including:
+   - Evading human oversight
+   - Self-replication
+   - Autonomous AI development
+
+3. **Insufficiency conclusion**: "Current public benchmarks are insufficient, on their own, for providing the evidence of comprehensive risk assessment required for regulatory compliance."
+
+4. **Implication**: "Independent, targeted evaluation tools specifically designed for regulatory requirements remain necessary to adequately address compliance obligations under the EU AI Act."
+
+## Agent Notes
+
+**Why this matters:** This paper creates a specific empirical bridge between two threads: (1) the EU AI Act's mandatory evaluation obligations (Article 55) and (2) the practical infeasibility of meeting those obligations with currently existing evaluation tools. Labs may be trying to comply with Article 55 using benchmarks that don't cover the most alignment-critical behaviors. The compliance gap is not just structural (voluntary vs. mandatory) but technical.
+
+**What surprised me:** Zero coverage of oversight-evasion and self-replication in 195,000 benchmark questions is a striking number. These are precisely the capabilities that matter most for the alignment-critical scenarios the EU AI Act is trying to govern. Labs can demonstrate "good performance" on existing benchmarks while having unmeasured capabilities in exactly the areas that matter.
+
+**What I expected but didn't find:** Any existing benchmark suites specifically designed for Article 55 compliance. The paper implies these don't exist — they're the necessary next step that hasn't been built.
+
+**KB connections:**
+- [[scalable oversight degrades rapidly as capability gaps grow with debate achieving only 50 percent success at moderate gaps]] — this paper shows the problem isn't just oversight at deployment, it's that the evaluation tools for oversight don't even measure the right things
+- formal verification of AI-generated proofs provides scalable oversight that human review cannot match — formal verification works for mathematical domains; this paper shows behavioral compliance benchmarking fails even more completely
+- AI capability and reliability are independent dimensions — benchmarks measure one dimension (behavioral propensities) and miss another (alignment-critical failure modes)
+
+**Extraction hints:** Strong claim candidate: "Current AI benchmarks provide zero coverage of capabilities central to loss-of-control scenarios — oversight evasion, self-replication, autonomous AI development — making them structurally insufficient for EU AI Act Article 55 compliance despite being the primary compliance evidence labs provide." This is specific, falsifiable, empirically grounded.
+
+**Context:** Published August 2025 — after GPAI obligations came into force (August 2, 2025). This is a retrospective assessment of whether the evaluation infrastructure that exists is adequate for the compliance obligations that just became mandatory.
+
+## Curator Notes (structured handoff for extractor)
+PRIMARY CONNECTION: [[scalable oversight degrades rapidly as capability gaps grow with debate achieving only 50 percent success at moderate gaps]]
+WHY ARCHIVED: Creates empirical bridge between EU AI Act mandatory obligations and the practical impossibility of compliance through existing evaluation tools — closes the loop on the "evaluation infrastructure building but architecturally wrong" thesis
+EXTRACTION HINT: Focus on the zero-coverage finding for loss-of-control capabilities — this is the most striking and specific number, and it directly supports the argument that compliance infrastructure exists on paper but not in practice
+
+
+## Key Facts
+- EU AI Act GPAI obligations (Article 55) came into force August 2, 2025
+- Prandi et al. analyzed approximately 195,000 benchmark questions using LLM-as-judge methodology
+- 61.6% of regulatory-relevant benchmark coverage addresses 'tendency to hallucinate'
+- 31.2% of regulatory-relevant benchmark coverage addresses 'lack of performance reliability'
+- Zero benchmark questions in the analyzed corpus covered oversight evasion, self-replication, or autonomous AI development capabilities

inbox/queue/2026-03-20-eu-ai-act-article43-conformity-assessment-limits.md

@@ -0,0 +1,62 @@
+---
+type: source
+title: "EU AI Act Article 43: Conformity Assessment is Mostly Self-Assessment, Not Independent Third-Party Evaluation"
+author: "European Union / EU AI Act (euaiact.com)"
+url: https://www.euaiact.com/article/43
+date: 2024-07-12
+domain: ai-alignment
+secondary_domains: []
+format: legislation
+status: null-result
+priority: medium
+tags: [EU-AI-Act, Article-43, conformity-assessment, self-assessment, notified-bodies, high-risk-AI, independence, FDA-comparison]
+processed_by: theseus
+processed_date: 2026-03-20
+extraction_model: "anthropic/claude-sonnet-4.5"
+extraction_notes: "LLM returned 1 claims, 1 rejected by validator"
+---
+
+## Content
+
+Article 43 establishes conformity assessment procedures for **high-risk AI systems** (not GPAI — high-risk AI is a separate category covering things like medical devices, recruitment systems, law enforcement uses).
+
+**Assessment structure:**
+- For high-risk AI in **Annex III point 1** (biometric identification): providers may choose between internal control (self-assessment) OR quality management system assessment with notified body involvement
+- For high-risk AI in **Annex III points 2-8** (all other categories): **internal control (self-assessment) only** — no notified body required
+- Third-party notified body required ONLY when: harmonized standards don't exist, common specifications unavailable, provider hasn't fully applied relevant standards, or standards published with restrictions
+
+**Notified bodies:** Third-party conformity assessment organizations designated under the regulation. For law enforcement and immigration uses, the market surveillance authority acts as the notified body.
+
+**Key implication:** For the vast majority of high-risk AI systems, Article 43 permits self-certification of compliance. The "conformity assessment" of the EU AI Act is predominantly a documentation exercise, not an independent evaluation.
+
+**Important distinction from GPAI:** Article 43 governs high-risk AI systems (classification by use case); GPAI systemic risk provisions (Articles 51-56) govern models by training compute scale. These are different categories — the biggest frontier models may be GPAI systemic risk WITHOUT being classified as high-risk AI systems, and vice versa. They operate under different regulatory regimes.
+
+## Agent Notes
+
+**Why this matters:** Article 43 is frequently cited as the EU AI Act's "conformity assessment" mechanism, implying independent evaluation. In reality it's self-assessment for almost all high-risk AI, with third-party evaluation as an exception. This matters for understanding whether the EU AI Act creates the "FDA equivalent" that Brundage et al. say is missing. Answer: No, not through Article 43.
+
+**What surprised me:** The simplicity of the answer. Article 43 ≠ FDA because it allows self-assessment for most cases. The path to any independent evaluation in the EU AI Act runs through Article 92 (compulsory AI Office evaluation), not Article 43 (conformity assessment). These are different mechanisms with different triggers.
+
+**What I expected but didn't find:** Any requirement that third-party notified bodies verify the actual model behavior, as opposed to reviewing documentation. Even where notified bodies ARE required (Annex III point 1), their role appears to be quality management system review, not independent capability evaluation.
+
+**KB connections:**
+- Previous session finding from Brundage et al. (arXiv:2601.11699): AAL-1 (peak of current voluntary practice) still relies substantially on company-provided information. Article 43 self-assessment is structurally at or below AAL-1.
+
+**Extraction hints:** This source is better used to CORRECT a potential misunderstanding than to make a new claim. The corrective claim: "EU AI Act conformity assessment under Article 43 primarily permits self-certification — third-party notified body review is the exception, not the rule, applying to a narrow subset of high-risk use cases when harmonized standards don't exist." The path to independent evaluation runs through Article 92, not Article 43.
+
+**Context:** Article 43 applies to high-risk AI systems (Annex III list: biometrics, critical infrastructure, education, employment, essential services, law enforcement, migration, justice). GPAI models face a separate and in some ways more stringent regime under Articles 51-56 when they meet the systemic risk threshold.
+
+## Curator Notes (structured handoff for extractor)
+PRIMARY CONNECTION: voluntary safety pledges cannot survive competitive pressure — self-certification under Article 43 has the same structural weakness as voluntary commitments; labs certify their own compliance
+WHY ARCHIVED: Corrects common misreading of EU AI Act as creating FDA-equivalent independent evaluation via Article 43; clarifies that independent evaluation runs through Article 92 (reactive) not Article 43 (conformity)
+EXTRACTION HINT: This is primarily a clarifying/corrective source; extractor should check whether any existing KB claims overstate Article 43's independence requirements and note the Article 43 / Article 92 distinction
+
+
+## Key Facts
+- EU AI Act Article 43 governs conformity assessment for high-risk AI systems (Annex III categories)
+- High-risk AI in Annex III points 2-8 use internal control (self-assessment) only
+- High-risk AI in Annex III point 1 (biometric identification) may choose between internal control OR notified body assessment
+- Third-party notified body required only when: harmonized standards don't exist, common specifications unavailable, provider hasn't fully applied standards, or standards published with restrictions
+- For law enforcement and immigration uses, the market surveillance authority acts as the notified body
+- Article 43 applies to high-risk AI systems (classification by use case), distinct from GPAI systemic risk provisions (Articles 51-56) which govern models by training compute scale
+- Article 92 provides compulsory AI Office evaluation as a separate mechanism from Article 43 conformity assessment

inbox/queue/2026-03-20-eu-ai-act-digital-simplification-nov2025.md

@@ -0,0 +1,58 @@
+---
+type: source
+title: "EU Digital Simplification Package: November 2025 Commission Amendments to AI Act"
+author: "European Commission (indirect — derived from multiple sources)"
+url: https://digital-strategy.ec.europa.eu/en/policies/european-approach-artificial-intelligence
+date: 2025-11-19
+domain: ai-alignment
+secondary_domains: []
+format: policy-document
+status: null-result
+priority: medium
+tags: [EU-AI-Act, Digital-Simplification-Package, deregulation, GPAI, amendments, enforcement-gap]
+processed_by: theseus
+processed_date: 2026-03-20
+extraction_model: "anthropic/claude-sonnet-4.5"
+extraction_notes: "LLM returned 0 claims, 0 rejected by validator"
+---
+
+## Content
+
+On **November 19, 2025**, the European Commission proposed "targeted amendments" via a Digital Simplification Package that affects the EU AI Act. This information derives from the EC's digital strategy page which notes: "Commission proposed targeted amendments via Digital Simplification Package."
+
+**What is known:** The Digital Simplification Package is part of broader EU deregulatory effort to reduce compliance burden on businesses, particularly SMEs. It follows the EU's "competitiveness agenda" under pressure from US AI dominance and concerns about European AI companies being disadvantaged.
+
+**What is NOT confirmed from accessible sources:** The specific AI Act provisions targeted, whether GPAI Articles 53-55 are affected, whether Article 92 enforcement powers are modified, whether conformity assessment timelines are extended.
+
+**Pattern context:** The November 2025 amendment proposal follows a broader EU pattern: GPAI Code of Practice finalized July 2025 (on schedule), GPAI obligations applied August 2025 (on schedule), then November 2025 simplification proposal seeks to modify what was just implemented.
+
+**Structural concern:** If simplification targets GPAI provisions, it would follow the same pattern as the US: capability scaling triggers deployment, then governance implementation triggers deregulation pressure. The NIST EO rescission (January 2025, US) and EU Digital Simplification Package (November 2025) may represent a convergent pattern where regulatory implementation itself generates industry pushback sufficient to reverse it.
+
+## Agent Notes
+
+**Why this matters:** The timing is architecturally significant. Mandatory GPAI obligations came into force August 2, 2025. Within 3.5 months, the Commission proposed simplification amendments. This is either: (a) routine administrative refinement, or (b) industry pushback causing deregulatory reversal before enforcement gets established. The answer determines whether the EU AI Act represents durable mandatory governance or a temporary framework subject to competitive erosion.
+
+**What surprised me:** I could not access the specific amendments proposed. All sources referencing the Digital Simplification Package were either 404, blocked, or only mentioned it in passing. This is itself informative — the amendments may not have generated as much scholarly/policy analysis as the initial Act provisions. The absence of analysis could mean the changes are technical rather than substantive, OR that they haven't been fully processed yet by the policy community.
+
+**What I expected but didn't find:** Specific provisions being modified. Without this, I cannot assess whether the amendments strengthen, weaken, or simply clarify existing obligations.
+
+**KB connections:**
+- [[technology advances exponentially but coordination mechanisms evolve linearly creating a widening gap]] — if simplification amendments weaken enforcement, the gap widens further
+- voluntary safety pledges cannot survive competitive pressure — EU legislative amendments under competitive pressure may follow the same structural logic as voluntary pledge weakening
+
+**Extraction hints:** This source is primarily a flag rather than a substantive claim source. The claim candidate: "EU AI Act enforcement faced simplification pressure within 3.5 months of GPAI obligations taking effect — suggesting the regulatory implementation cycle for AI governance may itself be subject to competitive erosion dynamics similar to voluntary commitment collapse." But this needs confirmation of what the amendments actually propose.
+
+**Context:** The Digital Simplification Package is part of Commissioner Teresa Ribera's broader work to improve EU competitiveness. The AI Act amendments are one element of a broader deregulatory push affecting GDPR, product liability, and other digital regulations.
+
+## Curator Notes (structured handoff for extractor)
+PRIMARY CONNECTION: [[technology advances exponentially but coordination mechanisms evolve linearly creating a widening gap]]
+WHY ARCHIVED: Documents the pattern of rapid regulatory pushback following mandatory obligation implementation — important for assessing durability of EU AI Act enforcement
+EXTRACTION HINT: This source is incomplete — specific amendment content not confirmed. Extractor should search specifically for "EU AI Act Digital Simplification Package" + specific article amendments before extracting a claim. Flag as needing follow-up.
+
+
+## Key Facts
+- EU AI Act GPAI Code of Practice finalized July 2025
+- EU AI Act GPAI obligations applied August 2, 2025
+- European Commission proposed Digital Simplification Package amendments November 19, 2025
+- Digital Simplification Package is part of EU competitiveness agenda under Commissioner Teresa Ribera
+- Specific AI Act provisions targeted by amendments not publicly confirmed in accessible sources

inbox/queue/2026-03-20-euaiact-article92-compulsory-evaluation-powers.md

@@ -0,0 +1,83 @@
+---
+type: source
+title: "EU AI Act Articles 51-56, 88-93, 101: GPAI Systemic Risk Obligations and Compulsory Evaluation Framework"
+author: "European Union / EU AI Act (euaiact.com)"
+url: https://www.euaiact.com/article/51
+date: 2024-07-12
+domain: ai-alignment
+secondary_domains: []
+format: legislation
+status: null-result
+priority: high
+tags: [EU-AI-Act, GPAI, systemic-risk, Article-55, Article-92, conformity-assessment, independent-evaluation, AI-Office, enforcement, 10-25-FLOPs]
+processed_by: theseus
+processed_date: 2026-03-20
+extraction_model: "anthropic/claude-sonnet-4.5"
+extraction_notes: "LLM returned 2 claims, 2 rejected by validator"
+---
+
+## Content
+
+### Article 51 — GPAI Systemic Risk Classification
+A GPAI model qualifies as having systemic risk if it demonstrates high-impact capabilities OR if the Commission designates it as such. Presumption threshold: cumulative training compute exceeding **10^25 floating-point operations** (approximately the compute used to train GPT-4 and above). This threshold captures only the most computationally intensive frontier models.
+
+### Article 53 — Standard GPAI Provider Obligations
+All GPAI providers must: (1) maintain technical documentation of training and testing processes; (2) provide downstream developers with capability/limitation disclosures; (3) establish copyright compliance policies; (4) publish training data summaries. Open-source exception applies EXCEPT for models with systemic risk.
+
+### Article 55 — Systemic Risk GPAI Obligations
+Providers of systemic-risk GPAI models must: (1) **perform model evaluation including adversarial testing** in accordance with standardized protocols reflecting state-of-the-art; (2) assess and address systemic risks at EU level; (3) track and report serious incidents without undue delay; (4) maintain cybersecurity protections. Compliance pathways are flexible: codes of practice, harmonized standards, or "alternative adequate means" assessed by the Commission. NOT mandatory independent third-party audit.
+
+### Article 56 — Codes of Practice
+AI Office facilitates voluntary codes of practice development with industry, academia, civil society. Codes must be ready by May 2025; Commission approved final Code July 10, 2025. Commission may give approved codes binding force via implementing act. If codes prove inadequate by August 2025, Commission may impose binding common rules.
+
+### Article 88 — Commission Exclusive Enforcement Powers
+Commission receives exclusive powers to supervise and enforce GPAI rules. Implementation delegated to AI Office. National authorities can request Commission assistance when proportionate.
+
+### Article 91 — Information and Documentation Requests
+AI Office may request GPAI providers to submit compliance documentation or "any additional information necessary for assessing compliance." Commission may also compel access upon scientific panel requests. Structured dialogue may precede formal requests. Procedurally specific requirements for all requests.
+
+### Article 92 — Compulsory Evaluation Powers (KEY PROVISION)
+The AI Office may conduct independent evaluations of GPAI models in two scenarios: (1) when Article 91 documentation is insufficient for compliance assessment; (2) to investigate union-level systemic risks following qualified alerts from the scientific panel. Powers include: appointing **independent experts** from the scientific panel; compelling access via APIs, source code, and "appropriate technical means and tools." Providers must comply under penalty of fines. This is a **compulsory** access mechanism — not voluntary-collaborative.
+
+### Article 101 — Fines for GPAI Providers
+Maximum fine: **3% of annual worldwide turnover or EUR 15 million, whichever is higher**. Applies to violations including: violating regulation provisions, failing to provide requested documents, disobeying measures requested, denying access for Commission evaluations.
+
+## Agent Notes
+
+**Why this matters:** This is the most detailed picture of what the EU AI Act actually creates for GPAI systemic risk models. The key finding is that Article 92 creates genuinely compulsory evaluation powers — not voluntary-collaborative like METR/AISI — but they're triggered reactively (by qualified alerts or compliance failures), not proactively required before deployment. This is a crucial distinction from the FDA pre-market approval model.
+
+**What surprised me:** Article 92's compulsory access to APIs and source code is meaningfully stronger than I expected based on yesterday's research. The AI Office can appoint independent experts and compel technical access. This moves the EU AI Act closer to AAL-2 (non-reliance on company statements when triggered) but still falls short of AAL-3/4 (deception-resilient, proactive).
+
+**What I expected but didn't find:** A proactive pre-deployment evaluation requirement. The EU AI Act creates mandatory obligations (Article 55) with binding enforcement (Articles 92, 101) but the evaluation is triggered by problems, not required as a condition of deployment. The FDA analogy fails specifically here — drugs cannot be deployed without pre-market approval; GPAI models under EU AI Act can be deployed while the AI Office monitors and intervenes reactively.
+
+**KB connections:**
+- voluntary safety pledges cannot survive competitive pressure — Article 55 creates mandatory obligations that don't depend on voluntary commitment, but the flexible compliance pathways preserve lab discretion in HOW they comply
+- scalable oversight degrades rapidly as capability gaps grow — Article 92's compulsory evaluation powers don't solve the AAL-3/4 infeasibility problem; even with source code access, deception-resilient evaluation is technically infeasible
+- technology advances exponentially but coordination mechanisms evolve linearly — the 10^25 FLOP threshold will require updating as compute efficiency improves
+
+**Extraction hints:** Primary claim: "EU AI Act Article 92 creates the first binding compulsory evaluation powers for frontier AI models globally — AI Office can compel API/source code access and appoint independent experts — but enforcement is reactive not proactive, falling structurally short of FDA pre-market approval." Secondary claim: "EU AI Act flexible compliance pathways for Article 55 allow GPAI systemic risk models to self-certify compliance through codes of practice rather than mandatory independent third-party audit."
+
+**Context:** This is a synthesis of Articles 51, 53, 55, 56, 88, 91, 92, 101 from the EU AI Act. GPAI obligations became applicable August 2, 2025. The Act is in force globally for any frontier AI models deployed in EU market.
+
+## Curator Notes (structured handoff for extractor)
+PRIMARY CONNECTION: [[voluntary safety pledges cannot survive competitive pressure because unilateral commitments are structurally punished when competitors advance without equivalent constraints]] — EU AI Act's mandatory structure counters this weakness, but flexible compliance pathways partially reintroduce it
+WHY ARCHIVED: First binding mandatory evaluation framework globally for frontier AI — essential for B1 disconfirmation assessment and the multi-session "governance gap" thesis
+EXTRACTION HINT: Focus on the Article 92 compulsory evaluation / reactive vs proactive distinction — this is the key structural feature that makes EU AI Act stronger than voluntary-collaborative METR/AISI but weaker than FDA pre-market approval
+
+
+## Key Facts
+- EU AI Act became applicable August 2, 2025
+- GPAI systemic risk threshold: 10^25 floating-point operations (approximately GPT-4 training compute)
+- Maximum fine for GPAI violations: 3% of annual worldwide turnover or EUR 15 million, whichever is higher
+- Final Code of Practice approved July 10, 2025
+- Codes of practice deadline was May 2025
+- Commission deadline to impose binding common rules if codes inadequate: August 2025
+- Article 51 defines GPAI systemic risk classification
+- Article 53 defines standard GPAI provider obligations
+- Article 55 defines systemic risk GPAI obligations
+- Article 56 defines codes of practice process
+- Article 88 grants Commission exclusive enforcement powers
+- Article 91 defines information and documentation request powers
+- Article 92 defines compulsory evaluation powers
+- Article 101 defines fines for GPAI providers
+- Open-source exception applies to standard GPAI obligations EXCEPT for models with systemic risk

inbox/queue/2026-03-20-stelling-gpai-cop-industry-mapping.md

@@ -0,0 +1,57 @@
+---
+type: source
+title: "Mapping Industry Practices to EU AI Act GPAI Code of Practice Safety and Security Measures (arXiv:2504.15181)"
+author: "Lily Stelling, Mick Yang, Rokas Gipiškis, Leon Staufer, Ze Shen Chin, Siméon Campos, Ariel Gil, Michael Chen"
+url: https://arxiv.org/abs/2504.15181
+date: 2025-04-01
+domain: ai-alignment
+secondary_domains: []
+format: paper
+status: null-result
+priority: high
+tags: [GPAI, Code-of-Practice, industry-practices, EU-AI-Act, safety-measures, OpenAI, Anthropic, Google-DeepMind, compliance, voluntary]
+processed_by: theseus
+processed_date: 2026-03-20
+extraction_model: "anthropic/claude-sonnet-4.5"
+extraction_notes: "LLM returned 0 claims, 0 rejected by validator"
+---
+
+## Content
+
+166-page analysis comparing safety and security measures in the EU AI Act's General-Purpose AI Code of Practice (Third Draft) against actual commitments from leading AI companies. Examined documents from over a dozen companies including OpenAI, Anthropic, Google DeepMind, Microsoft, Meta, and Amazon.
+
+**Key Finding:** "Relevant quotes from at least 5 companies' documents for the majority of the measures in Commitments II.1-II.16" within the Safety and Security section.
+
+**Important Caveat (author-stated):** "This report is not meant to be an indication of legal compliance, nor does it take any prescriptive viewpoint about the Code of Practice or companies' policies."
+
+**Context:** The GPAI Code of Practice (Third Draft, April 2025) was finalized and received by the Commission on July 10, 2025, and became applicable August 2, 2025.
+
+## Agent Notes
+
+**Why this matters:** This paper shows that existing frontier AI lab policies already contain language matching the majority of Code of Practice safety measures. This is important for two competing interpretations: (1) Pro-governance reading: the Code of Practice reflects real existing practices, making compliance feasible. (2) Anti-governance reading: if labs already claim to do most of this, the Code simply formalizes current voluntary commitments rather than creating new obligations — it's the same voluntary-collaborative problem in formal dress.
+
+**What surprised me:** The author caveat is striking: they explicitly say this is NOT evidence of compliance. Labs may publish commitments that match the Code language while the actual model behaviors don't correspond. This is the deception-resilient gap — what labs say they do vs. what their models do.
+
+**What I expected but didn't find:** Evidence that the Code of Practice requires genuinely independent third-party verification of the safety measures it lists. From the structure, it appears labs self-certify compliance through code adherence, with the AI Office potentially auditing retrospectively.
+
+**KB connections:**
+- voluntary safety pledges cannot survive competitive pressure — the Code of Practice may formalize existing voluntary commitments without adding enforcement mechanisms that survive competitive pressure
+- an aligned-seeming AI may be strategically deceptive — the gap between published safety commitments and actual model behavior is precisely what deception-resilient evaluation (AAL-3/4) is designed to detect
+
+**Extraction hints:** Supporting claim: "GPAI Code of Practice safety measures map to existing commitments from major AI labs — but the mapping is of stated policies, not verified behaviors, leaving the deception-resilient gap unaddressed." Use cautiously — authors explicitly say this is not compliance evidence.
+
+**Context:** Independent analysis by researchers at AI safety/governance organizations. Not affiliated with the AI Office or Commission.
+
+## Curator Notes (structured handoff for extractor)
+PRIMARY CONNECTION: [[voluntary safety pledges cannot survive competitive pressure because unilateral commitments are structurally punished when competitors advance without equivalent constraints]]
+WHY ARCHIVED: Shows that Code of Practice may be formalizing existing practices rather than creating new obligations — relevant to whether mandatory framework actually changes behavior
+EXTRACTION HINT: Be careful about the author caveat — this is evidence about stated policies not compliance evidence; extractor should note this distinction clearly
+
+
+## Key Facts
+- EU AI Act GPAI Code of Practice Third Draft finalized April 2025
+- Code of Practice received by Commission July 10, 2025
+- Code of Practice became applicable August 2, 2025
+- Analysis examined documents from over a dozen companies including OpenAI, Anthropic, Google DeepMind, Microsoft, Meta, and Amazon
+- Paper is 166 pages analyzing safety and security measures
+- Authors found relevant quotes from at least 5 companies for majority of measures in Commitments II.1-II.16