Gap Scan — 2026-07-16

Scanned

Recent window: 4 dev-log files (07-14 morning/evening, 07-15 morning/evening — the 07-15 evening file is unprocessed and left for tonight's developments-log cycle; only gap-relevant items actioned), ~64 pages edited in last 48h. Rotation slice: 0 — companies/ (all, ~117 pages).

Gaps actioned (7 of ~21 found)

New pages created (live)

• UMG Recordings v. Suno (AI music training data) — the 07-13 lint report's suggested source #3 asked for a litigation anchor for "the stream-ripping allegations the hack corroborates"; Suno (created 07-15) referenced the RIAA-coordinated litigation with no case page, court, or docket, and listed the missing docket as an open question. Built from the CourtListener docket (1:24-cv-11611, D. Mass., Saylor/Levenson, filed 2024-06-24, active through 07-14 entries) plus Music Business Worldwide's motion-papers coverage: 560 works → proposed 61,026-work second amended complaint, Audible Magic discovery history, the Warner settlement and January 2026 dismissal, the April 6 discovery ruling and July 13 overruled objection, and the January 8, 2027 dispositive-motions deadline. Depends on queued source: UMG v. Suno complaint. Suno wired to the new page (litigates relationship; docket open-question resolved; sources 1 → 3).

Pages expanded (live)

• ElevenLabs — slice 0's highest-in-degree misrated thin anchor (in-degree 11, 300 words, high on sources_count: 1 — the stochastic-parrots pattern, stale since 06-06). Added a Snapshot valuation table and a Funding and valuation section: $180M Series C at $3.3B (Jan 2025), ~$6.6B September 2025 secondary, $500M Series D at $11B led by Sequoia with total funding $781M (Feb 4, 2026), CEO Mati Staniszewski / Nvidia backing / IPO framing, and the July 2, 2026 Bloomberg-reported tender-offer talks at ~$22B. Every prior fact and citation preserved. Confidence honestly re-rated high → medium (products coverage remains thin-sourced); sources_count 1 → 6. [5 new (Source: URL) cites]
• Mandiant — slice thin anchor #2 (in-degree 8, 306 words, high/1). Added the 2026 Google Cloud arc: the March agentic-AI security strategy with Wiz integration, Cloud Next 2026's Threat Hunting and Detection Engineering agents, May's AI Threat Defense combining Gemini, Wiz, CodeMender, and Mandiant, M-Trends 2026's AI-driven red-team techniques, and the June 2026 quiet layoffs hitting the threat-intelligence unit and Mandiant as spending shifts to AI. Every prior fact and citation preserved. Confidence high → medium; sources_count 1 → 6. [5 new (Source: URL) cites]
• Anthropic (cite upgrade, in passing) — the 07-15 nightly cycle's explicit gap-scan handoff: the values-across-languages fold cited only a Medium repost. Upgraded to the anthropic.com primary, corrected the date (July 13, not 14) and sample figure (309,815), and added the four named value axes. Medium cite retained as secondary per fidelity rules.

Link aliases fixed

• None this run. The broken-link scan's ≥2-inbound targets are _meta/ cross-namespace links, table-cell \| pipe-link artifacts (not actually broken), already-queued sources, or standing carries (person/org entities at in-degree 2, claude-code placement, foundation-models disambiguation).

Queued — foundational sources

• Anthropic, "Claude's values across models and languages" (Societal Impacts, July 13, 2026) — type-1 dangling foundational reference on a live thread; the 07-15 cycle flagged the Medium-repost citation for upgrade. Verified on anthropic.com (bibtex date block; author list; distinctive method passages). Full raw saved: Raw Sources/Claude's Values Across Models and Languages.md. Queued: INGEST-claude-values-models-languages-2026-07-16.md.
• Hachette v. Google complaint (S.D.N.Y. 1:26-cv-05870, Doc. 1, filed July 10, 2026) — lint suggested source #1 and the 07-15 scan's top deferred item, matured now that the nightly fold created Hachette et al. v. Google (Gemini training data). Verified on CourtListener: parties (Turow, Hachette, Elsevier, Cengage, S.C.R.I.B.E.), Judge Preska, counsel Oppenheim, Exhibit A attached; duplicate-caption docket 1:26-cv-05869 noted for ingest. Queued URL-only (court-PDF pattern): INGEST-hachette-v-google-complaint-2026-07-16.md.
• UMG Recordings v. Suno complaint (D. Mass. 1:24-cv-11611, filed June 24, 2024) — the primary text behind the new litigation page (lint suggested source #3). Verified on CourtListener (docket, judge, nature of suit, active status) with MBW corroboration of the case posture. Queued URL-only: INGEST-umg-v-suno-complaint-2026-07-16.md.

Flagged for review (beyond gap-scan's lanes)

• companies/fairly-trained.md appears misfiled — a certification body (not an AI-building business) sitting in companies/, which the schema reserves for for-profit AI developers; belongs in entities/. Migration touches ~8 inbound links — curator/lint territory. Note: needs-review/2026-07-16-fairly-trained-misfiled.md.
• H.R. 9619 (People-First Chatbot Act) — re-checked on Congress.gov: "As of 07/16/2026 text has not been received." Carried again; re-check next run.

Authenticity-verification failures

• None. All three queued sources verified on canonical hosts (anthropic.com; CourtListener/RECAP for both complaints). One provenance caution: CourtListener storage PDFs are not directly fetchable by agents, so both complaint queues are URL-only with ingest routed through the CourtListener read_document tool.

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

• Unprocessed 07-15 evening dev-log items — tonight's fold first, then gap-scan candidates: Thinking Machines Lab Inkling (first in-house open-weight MoE, 975B total/41B active params — possible models/inkling page), Guidelight AI Standards Control standard v1.0 (Steven Adler; six loss-of-control principles — possible standards/ page + source queue), xAI Grok CSAM suit (~7,000 images; litigation page candidate), OpenAI "Mind the US-China Safety Gap" (policy-agenda site launch; source queue candidate), Anthropic IPO credit-line talks, Microsoft security overhaul, WSJ AI-backlash investigation, EDPB guidelines (already queued 07-08/07-10). Score ~4–6 each once folded.
• US AI Regulatory Approaches Compared (in-deg 36) and Labor Disruption Timelines: Who Predicts What and Why (in-deg 21) — carried since 07-12. analysis/ pages sit outside the gap-scan AUTO new-page lane (they are crystallize/query outputs); recommend a dedicated crystallize or query session rather than further carries. Score ~4–5 each.
• Slice-0 residue (two-per-run pacing): Cloudflare (in-deg 17, missing sources_count — frontmatter repair, lint), Coinbase (deg 9, 328w), Replit (deg 8, 323w), Liquid AI (deg 8, medium/1), Baseten (deg 6, medium/1), Foxconn (Hon Hai Precision Industry) (deg 6, low/3), Crusoe, Hayden AI, Lambda Labs, Kuaishou / Kling AI (deg 5, low/1 each). Score ~2–3 each.
• sources_count missing on the big-4 company pages (anthropic, openai, nvidia-tsmc, meta — all high) — known lint backfill item (07-14 note), not a gap-scan lane.
• Possible standalone models/gemini-3-5 page once 3.5 Pro ships (reported targeting July 17 — tomorrow). Xi's WAIC keynote also July 17. Score ~3 each; check next run.
• Orphan-check sources-layer backlog (07-15): six state chatbot-act primary texts + the Model Spec blog post. Score ~2 each; batch decision when the queue drains.
• Standing carries: claude-code/claude-cowork placement (curator-blocked), inverse-cooking/inverse-trust coined terms (needs-review 07-10), duplicate entities/cdao/government/cdao (lint), concepts/foundation-models disambiguation, concepts/safety-training-methodologies (planned umbrella, in-deg 2), concepts/a-vision-of-democratic-ai (in-deg 2), person/org pages at in-degree 2 (pam-bondi, masahiro-mori, anil-seth, eric-horvitz, orin-kerr, alondra-nelson, palmer-luckey, tom-mitchell, soufan-center, richard-dawkins), Helix (Figure AI Vision-Language-Action model) (next models slice). Score ~2–3.

One-line summary

Three foundational sources verified and queued (the Anthropic values-axes study primary — closing the 07-15 cycle's Medium-repost handoff — plus the Hachette v. Google and UMG v. Suno complaints); one new litigation page created live (UMG v. Suno, resolving companies/suno's open docket question and the lint report's Suno-litigation gap); slice 0's two most misrated thin anchors (elevenlabs, mandiant) expanded and honestly re-rated; fairly-trained flagged as misfiled; H.R. 9619 still has no text. Nothing failed verification; the queue awaits the user's ingest review.