Gap Scan — 2026-06-07

Scanned

Recent window: 5 dev-log files (2026-06-05-1505/2205, 2026-06-06-0505/1505/2205) and ~200 pages with last_updated in the last 48h (the 06-05/06-06 nightly cycles touched broadly). The 06-06-2205 dev log is unprocessed by the nightly ingest (last processed: 06-06-1505); its news items (Krishnan exit, Project Solara, Helion raise, claude-oceanus leak, bot-traffic crossover, Marcus "AI's Black Friday") were left for developments-log, but its live threads informed this run's entity and expansion picks. Rotation slice: 3 — concepts/ P–Z (61 pages thinness-scanned against in-degree). Backlog ingested from the 2026-06-06 lint report (gap candidates 1–3) and the last 7 days of gap-scan reports.

Gaps actioned (9 of ~21 found)

New pages created (live)

• Federal Trade Commission (FTC) — gap type 3, lint-report backlog (explicitly recommended 06-06) + live thread (X Corp consent-order petition, June 3 comment period). The FTC was referenced across 82 content pages — including an explicit [[entities/ftc]] *(stub)* link on AI Antitrust — with no page. Covers Section 5/6(b) authorities, the September 2025 companion-chatbot 6(b) inquiry (seven companies), TAKE IT DOWN Act enforcement (Ferguson letter, $53,088/violation, May 21 warning letters), the Microsoft–OpenAI inquiry and AEI exemption comments, and the X Corp petition. Built from the FTC's own press release + existing wiki anchors. 4 sources; confidence: medium.
• Americans for Responsible Innovation (ARI) — gap type 3, lint backlog (06-06 candidate #1) + live thread (district ads against the Frontier Act / Great American AI Act (Obernolte–Trahan), praise-plus-concerns on the June 2 EO). Referenced across six content pages with no anchor. Founded March 2024 by Brad Carson and Eric Gastfriend; positions on preemption, export controls (host of Letter to Sec. Lutnick on H20 Restrictions (2025-07-28)), mandatory Mythos oversight, whistleblower and workforce bills. 4 sources; confidence: medium.
• Sriram Krishnan — gap type 3 + live thread (June 6 exit announcement dominating the unprocessed 06-06-2205 dev log; named on 5 existing content pages with no page). Senior White House AI policy advisor Jan 2025 – June 2026; AI Action Plan; Sacks collaboration; planned outside organization. 3 sources; confidence: medium.
• National Security Commission on Artificial Intelligence (NSCAI) — gap type 3, carried backlog (flagged "strong next-run candidate now that CNAS exists," 06-06; 4 inbound from Dominance by Understanding (Frazier-Rozenshtein policy frame), Center for a New American Security (CNAS), the Lawfare source). The commission behind the $32B/yr non-defense AI R&D benchmark the funding-gap thread measures against. 3 sources; confidence: medium.
• Brian Armstrong — gap type 3, perennial deferral (carried since 2026-05-26; 7 inbound link instances from Coinbase, Intelligence Replaces Hierarchy, Jack Dorsey, the Guardian source). Sibling page to the Dorsey founder-mode/manager-purge cluster, built from existing wiki anchors. 2 sources; confidence: medium.

Pages expanded (live)

• Pay-per-Crawl (Pigouvian Pricing of Agent Traffic) — slice-3 thin anchor (631w, sources_count: 1, in-degree 20, last_updated 2026-04-26) + live thread (the June 3 Cloudflare bot-traffic crossover in the 06-06-2205 dev log). Added a Launch and adoption section (July 1, 2025 launch, HTTP 402 mechanism, default AI-crawler blocking, the February 2026 Stack Overflow partnership) and an Agent traffic growth section (bots 57.5% of HTML HTTP requests vs 42.5% human, ~18 months ahead of Prince's forecast; HUMAN Security: agentic traffic +8,000% in 2025). [4 new (Source: URL) cites] sources_count 1 → 5.
• Safety Cases (Frontier AI) — slice-3 thin anchor (433w, sources_count: 2, in-degree 13, core safety area). Added a Templates and research program section folding the Buhl et al. scalable-safety-cases paper, the UK AISI inability-argument template, the GovAI cyber-inability CAE template, the AI-control safety-case sketch, and the March 2026 "Rethinking the Foundations of Frontier AI Safety Cases" survey. Also replaced the banned "Why it matters for AI" header with neutral phrasing per house style. [5 new (Source: URL) cites] sources_count 2 → 5.

Link aliases fixed

• [[entities/biden-administration-blueprint]] → OSTP Blueprint for an AI Bill of Rights (October 2022) in AI and Civil Liberties (1 instance, stub marker removed) and rewritten to plain text in the source page's own Open Questions (a self-link served no purpose). The dangling entity target is cleared; no page was warranted (the Blueprint is already covered by its primary-text source page).

Queued — foundational sources

• OpenAI — "Biodefense in the Intelligence Age" (June 4, 2026) — gap type 1 + lint backlog (06-06 "suggested next sources": "first frontier-lab biodefense roadmap... pairs with An Open Letter in Support of Mandatory Nucleic Acid Synthesis Screening and Recordkeeping (June 2026)"). Already cited as a supporting URL on AI Biosecurity and OpenAI; ingest upgrades it to a primary-text anchor. Verified: landing page on openai.com (June 4, 2026) + PDF on cdn.openai.com; GPT-Rosalind / Rosalind Biodefense context confirmed; corroborated by independent June 4–5 coverage. Pull strategy: PDF fetched fresh at ingest (large-PDF protocol). Queued: INGEST-openai-biodefense-action-plan-2026.md. Verification trail: queue/gap-scan/proposed-sources/openai-biodefense-action-plan-2026.md.

Authenticity-verification failures

• None this run. The single pulled foundational source verified cleanly against its canonical corporate host.

Carried-item resolutions

• entities/nscai — RESOLVED (created), clearing the 06-06 "strong next-run candidate."
• entities/brian-armstrong — RESOLVED (created), clearing a deferral carried since 2026-05-26.
• ARI plain-text reference (lint 06-06 candidate #1) and entities/ftc (lint candidate #3) — RESOLVED (created).

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

• Radical Optionality — slice-3 thin anchor, 492w × in-degree 40, sources_count: 1. Highest-leverage expansion left on the table this run. Score 3. Top promote next run.
• Slice-3 thin anchors not reached: three-lane-standards-governance (397w × 16), specification-hazard (524w × 17), three-privacy-problems-ai (796w × 17, src 1), professional-services-ai-adoption (361w, src 0), surveillance-technology (405w, src 0), standards-industrial-complex (425w, src 0), pol-is (155w, deg 5). Score 2 each.
• concepts/sociotechnical-risk-governance — frontmatter inconsistency (confidence: high with sources_count: 1, against the 3-source bar). Route to lint, not gap-scan.
• claude-code (7 inbound) — still blocked on a placement decision (product vs. company subsection vs. concepts alias). Score 3.
• entities/pim-de-witte / entities/pim-dewitte slug split (6 combined) — create page, then collapse variants. Score 3.
• CMA Google conduct-requirement decision text (gov.uk) — foundational ingest carried from 06-05. Score 4 — strongest queue candidate next run.
• concepts/singularity (4) and concepts/universal-basic-income (3) — fold decisions (vs. Superintelligence / AI Dividends (Universal Basic Capital, Digital Dividend, Global Dividend)); route to lint/crystallize. Score 2–3.
• companies/arm (3), entities/sequoia-capital (3), concepts/effective-altruism (2), entities/larry-ellison (2), entities/pam-bondi (2), entities/ny-dfs (2). Score 1–2.
• models/claude-sonnet-45 (2), models/gpt-5 (2), concepts/frontier-models (2) — small model/link gaps; revisit on the models/industries rotation (slice 9). Score 2.
• concepts/ai-prose-fingerprints (lint 06-06 candidate #2; Spinner antithesis study + AI-slop datapoints) — two datapoints, quality-gate marginal; re-evaluate when a third accretes. Score 2.
• Helion Energy / Project Solara / claude-oceanus leak (06-06-2205 dev log) — developments-log territory; the unprocessed digest should be folded by the next nightly ingest, not by gap-scan.
• Lint territory: sources/anthropic-pwc-expanded-partnership (6), sources/trustworthy-agents-in-practice (5), and the stale satisfied INGEST-* queue files flagged by the 06-06 lint report (queue-processor housekeeping pass).

Dashboard rebuild (skill step 10)

Failed (recurring) — bin/build-dashboard.py aborts on shutil.rmtree over OS-locked iCloud duplicate-naming artifacts (analysis 2, …) in the site directory; same failure documented since 2026-05-30 and logged at queue/gap-scan/needs-review/2026-06-06-dashboard-rebuild-failed (host-side fix required). The page graph itself is fully committed; only the static HTML mirror is stale.

One-line summary

Slice 3 (concepts/ P–Z): five entity gaps closed live (FTC, ARI, Sriram Krishnan, NSCAI, Brian Armstrong — clearing the lint backlog and two long-carried deferrals), two thin slice anchors expanded (pay-per-crawl with the June 2026 bot-traffic crossover; safety-cases with the AISI/GovAI template literature), one dangling alias cleared, and OpenAI's biodefense action plan verified and queued. User review needed: the queued biodefense ingest, and the recurring host-side dashboard-rebuild fix.