Gap Scan — 2026-07-04

Scanned

Recent window: 4 dev-log files (2026-07-02 ×2, 2026-07-03 ×2), ~40 pages edited in last 48h. Rotation slice: 2 — concepts/ G–O (34 pages).

Gaps actioned (8 of 19 found)

New pages created (live)

• Trump v. Slaughter — top-scored gap (~6: live thread + core area + 07-03 deferred-backlog carry). The June 29, 2026 SCOTUS ruling overruling Humphrey's Executor kept recurring in AI-governance threads (FTC AI-preemption statement, Fathom/CAISI verification op-ed, Warner's FTC-led agent registry) with only a passing wiki mention. Built from the supremecourt.gov slip opinion, SCOTUSblog's opinion analysis, and NPR. confidence: medium. Depends on queued source (slip opinion — cited inline).
• AI Labeling Act of 2026 — live-thread gap in a core area (US AI regulation). S. 4915 (119th), reintroduced June 24, 2026 by Schatz/Curtis/Warner (07-03 dev-log) with no page. Bill number verified via GovInfo; provisions from the sponsors' announcement; Politico and Music Business Worldwide coverage. confidence: medium.
• American AI Sovereign Wealth Fund Act — live-thread gap. Sanders's June 18, 2026 bill (one-time 50% stock tax, ~$7T fund, Independent Commission for Democratic AI) named in the 07-03 dev-log; connects to AI Public Wealth Fund and Government Equity in AI (edited <48h) and the OpenAI 5%-stake thread. Built from the sanders.senate.gov announcement and bill-text PDF, plus Guardian context. confidence: medium.
• Nuclear Regulatory Commission (NRC) — backlog carry (07-03, score ~4) reinforced by the 07-02 dev-log (nine Democratic AGs' NEPA comments on microreactor licensing for data centers). Regulator → entities/. Built from the Federal Register proposed rule (10 CFR Part 57, 2026-05-01), the Bonta comment-letter release (scraped live; EO 14300 implementation, coalition membership), Beveridge & Diamond and Politico Pro analyses, plus the Slaughter dissent naming the NRC. confidence: medium.

Pages expanded (live)

• Monitorability Tax — slice thin anchor: 495 words, confidence: low, sources_count: 1, in-degree 12. Anchored the term to its actual origin — Baker et al. (OpenAI, March 2025), obfuscated reward hacking, the "pay a monitorability tax by not applying strong optimization pressures" recommendation — and added the July 2025 cross-organization position paper (Korbak et al., ~40 authors: UK AISI, Apollo, OpenAI, DeepMind, Anthropic, Meta) urging developers to evaluate and report CoT monitorability. sources_count 1 → 3; confidence low → medium. Every prior fact and citation preserved. [3 new (Source: URL) cites]
• Law-Following AI — slice thin anchor: 510 words, sources_count: 1, in-degree 14. Added a Development and reception section: the Institute for Law & AI research-agenda framing, the SSRN working-paper lineage, the authors' Lawfare version ("AI Agents Must Follow the Law"), and O'Keefe's subsequent legal-duties writing. sources_count 1 → 4. The foundational Fordham article itself was already queued (INGEST-law-following-ai-okeefe-2025.md) — expanded via supporting sources per the precedence rule. [4 new (Source: URL) cites]
• Machine Fluency — slice thin anchor: 644 words, sources_count: 1, in-degree 15. Added a Publication and reception section: SSRN working-paper record and Marginal Revolution's December 2025 feature of the machine-fluency finding. sources_count 1 → 3. [2 new (Source: URL) cites]

Link aliases fixed

• [[queue/INGEST-magnifica-humanitas-leo-xiv-2026]] → [[sources/magnifica-humanitas-encyclical]] (2 instances: concepts/ai-displacement-vs-augmentation, concepts/ai-fluency-divide). The encyclical's source page exists; the pages were linking a since-processed queue task.
• [[concepts/ai-and-legal-services]] → [[industries/legal]] (2 instances: sources/prince-builders-sellers-measurers, sources/uc-berkeley-law-ai-policy-summer-2026). No legal-services concept page exists or is needed; industries/legal covers the sector debate.

Queued — foundational sources

• Trump v. Slaughter slip opinion (No. 25-332, June 29, 2026) — primary text behind the new litigation page. Verified: supremecourt.gov (the Court's own repository; docket number and distinctive passages corroborated by SCOTUSblog/NPR/firm alerts). Large PDF → fetch fresh at ingest. Queued: INGEST-trump-v-slaughter-opinion-2026-06-29.md (+ verification record in proposed-sources/trump-v-slaughter-opinion-2026.md).
• Baker et al., "Monitoring Reasoning Models for Misbehavior and the Risks of Promoting Obfuscation" (OpenAI, March 2025) — dangling foundational reference: Monitorability Tax attributed its defining term to "Baker et al." with no sources/ page or raw file. Verified: arXiv 2503.11926 + first-party OpenAI-hosted PDF; "monitorability tax" appears verbatim in the abstract, matching the NIST AI 800-4 quotation. Queued: INGEST-baker-cot-monitoring-obfuscation-2025.md (+ verification record in proposed-sources/baker-cot-monitoring-obfuscation-2025.md).

Authenticity-verification failures

• None. Both queued sources verified on their canonical hosts; all new/expanded pages were otherwise built from supporting web sources folded as (Source: URL) (senate.gov, govinfo.gov, federalregister.gov, oag.ca.gov, scotusblog.com, arxiv.org, ssrn.com, and press coverage).

Deferred backlog (over the daily cap — re-surfaces next run)

• Warner AI AGENT Act (discussion draft, June 29; detailed in 07-03 dev-logs) — lint rule holds: create a legislation/ page when formally introduced with a bill number. Fold into the Warner thread meanwhile. Score ~4.
• government/california (Newsom–Anthropic SITeS partnership, June 29: Claude at 50% discount for all state agencies, DMV/DHCS/CalOES deployments) — would be the first US-state deployer page in government/ (precedent: government/taiwan-moda). Worth a deliberate schema call; the material currently folds into Anthropic. Score ~4.
• Google FARO proposal (industry-composed frontier-AI regulatory organization, late June) — proposal-class; folds into Google DeepMind / AI Pre-Release Vetting via the nightly cycle unless it institutionalizes. Score ~3.
• entities/akrites (Linux Foundation open-source AI-cyber defense consortium, June 26) — single-thread mention so far. Score ~3.
• Argentina "automated company" bill (Milei/Duprat, detailed July 3) — international, non-core; revisit if it advances. Score ~3.
• concepts/open-weight-frontier-models (slice; 642 words, confidence: high with sources_count: 0, in-degree 14) — frontmatter inconsistency plus expandability; the queued INGEST-estimating-worst-case-frontier-risks-open-weight-llms would deepen it. Flag sources_count for lint. Score ~4.
• concepts/levels-of-autonomy-framework / concepts/normative-competence / concepts/mathematical-impossibility-perfect-fairness (slice) — confidence: high on sources_count: 1 (the lift-ai-act pattern); not short, so the fix is confidence/frontmatter review rather than blind expansion. Flagged for lint. Score ~3 each.
• concepts/long-industrial-revolution (slice; 482 words, medium, src 1, in-degree 10) — next-strongest slice expand. Score ~3.
• [[claude-code]] (in-degree 7, broken) — perennial; placement genuinely ambiguous (product vs. concept); needs a deliberate products/-schema call from the user. Score ~4.
• sources/anthropic-pwc-expanded-partnership (in-degree 6, broken) — supporting partnership announcement; fold into Anthropic (lint/fold task; no sources/ page warranted). Score ~3.
• [[claude-personal-guidance]] (in-degree 2, broken) — Anthropic research line named on concepts/sycophancy-and-hallucination and sources/ask-dont-tell-sycophancy-aisi; no obvious target page; likely resolvable with a supporting URL cite rather than a page. Score ~2.
• concepts/foundation-models (in-degree 2, broken, from concepts/frontier-models) — still needs disambiguation vs. General-Purpose AI (GPAI) before creating or aliasing (the Bommasani-term page would be defensible but low-priority). Score ~2.
• Person pages at in-degree 2 — pam-bondi, masahiro-mori, anil-seth, eric-horvitz, orin-kerr, alondra-nelson, clayton-christensen, palmer-luckey, eric-goldman, soufan-center (org). Confirm substance before creating. Score ~2–3 each.
• Duplicate company pages companies/reflection.md / companies/reflection-ai.md — merge/supersede remains out of gap-identifier scope (flagged since 06-28). Score ~4.

One-line summary

Four new pages (the Slaughter removal-power ruling, two live-thread federal AI bills, and the NRC as AI-infrastructure regulator), three slice thin-anchor expansions in concepts G–O, four alias fixes, and two verified foundational sources queued (the Slaughter slip opinion; Baker et al. 2025, origin of the monitorability tax) — the queue's INGEST backlog (70+ tasks) remains the main item needing the user's review.