Gap Scan — 2026-06-14

Scanned

Recent window: 0 dev-log files in last 48h (latest digest 2026-06-12-1529, already processed by the 06-14 ingest-reflect cycle); 6 wiki pages edited in last 48h (index.md, concepts/ai-coding-agents, concepts/ai-labor-disruption, sources/why-ai-hasnt-replaced-software-engineers, entities/sayash-kapoor, entities/arvind-narayanan — the software-engineering-labor thread). Rotation slice: 10 — analysis/ (all, 22 pages).

The live thread was found already well-supported: every concept and source it leans on (ai-as-normal-technology, ai-deskilling, ai-displacement-vs-augmentation, san-francisco-consensus, liars-dividend, ai-snake-oil-book, metr-early-2025-dev-productivity, darwin-godel-machine, mit-nanda-genai-divide, principal-agent-problem-ai) already has a page. The 06-14 ingest-reflect cycle had already folded the Narayanan & Kapoor essay. The analysis/ slice was clean — no missing-page clusters, only template artifacts (page, standards/) and one in-degree-1 dangling concept (concepts/ai-preemption).

Gaps actioned (6 of ~9 found)

New pages created (live)

• Effective Altruism — dangling [[concepts/effective-altruism]] (relied on by Longtermism and sources/farahany-class-23-effective-altruists-export-controls); a core AI-safety-funding/talent concept. Carried in the deferred backlog since 2026-06-07 and 06-08. Built from supporting web sources (EA Wikipedia, EA Funds Long-Term Future Fund, FTX Future Fund record, MacAskill "EA in the age of AGI"). confidence: medium.
• Regulatory Managerialism — dangling [[concepts/regulatory-managerialism]] (in-degree 2; Cohen-Waldman framing cited across the Deirdre K. Mulligan sociotechnical-governance cluster — interpretability-and-safety, AISI-critique pages). Carried in deferred backlog since 06-07/06-08. Anchored to the canonical Cohen & Waldman (2023) Law & Contemporary Problems introduction on the Duke Law repository. confidence: medium.
• Dropbox — dangling [[companies/dropbox]] from the Box "AI-native document layer / competes-with" cluster; applied-AI deployer with a named AI product (Dropbox Dash / Dash for Business). Persistent backlog item. Built from Dropbox's own Dash product and press pages plus trade coverage. confidence: medium.
• Notion — dangling [[companies/notion]] from the same Box cluster; applied-AI workspace product with Notion AI, Notion Agent, AI Meeting Notes, Enterprise Search, and 2026 credit-priced custom agents. Persistent backlog item. Built from Notion's own AI/pricing pages plus trade coverage. confidence: medium.

Pages expanded (live)

• None this run. The rotation slice and live-thread anchor pages were already adequately built.

Link aliases fixed

• [[entities/center-for-security-and-emerging-technology]] → [[entities/cset]] (2 instances in entities/ben-buchanan.md). CSET already has a page under the cset slug.
• [[concepts/sycophancy-llm]] → [[concepts/sycophancy-and-hallucination]] (2 instances in sources/ball-for-all-issues-so-triable.md). The existing concept page covers the sycophancy failure mode the reference intends.

Queued — foundational sources

• None this run. The one foundational dangling source surfaced (sources/trustworthy-agents-in-practice) is already queued (INGEST-trustworthy-agents-in-practice-anthropic.md + staged in proposed-sources/). The sources/anthropic-pwc-expanded-partnership dangling references (3 files) are a partnership announcement — supporting-class, not a foundational ingest; flagged below for lint to fold as inline (Source:) citations rather than a sources/ page.

Authenticity-verification failures

• None. No raw sources were pulled or saved this run; all new pages were built from supporting web sources cited inline as (Source: <URL>), per the AUTO new-page lane.

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

• claude-code — in-degree ~14 across companies/, entities/, models/ pages. Perennial since 2026-06-04; still blocked on a product-page placement decision (no products/ folder; ChatGPT/Copilot precedent folds products into the company page). Needs a schema call from the user — not auto-created. Score 3.
• sources/anthropic-pwc-expanded-partnership (3 inbound) — supporting-class partnership announcement mis-linked as a sources/ page; should be folded as inline (Source:) citations on industries/healthcare and the two Anthropic source pages. Hand to lint. Score 2.
• concepts/frontier-models — a general definitional "frontier models / frontier AI" concept page distinct from concepts/open-weight-frontier-models and concepts/frontier-ai-governance; mostly used as a tag (in-degree 2 as wikilink). Score 2.
• concepts/safetywashing (2 inbound), concepts/sovereign-ai alias question (→ concepts/sovereign-ai-product? — the existing page is specifically the product framing; a governance-sense page may be warranted). Score 1–2.
• In-degree-1 entity/concept stubs: entities/allen-institute-for-ai, concepts/ai-preemption, and assorted single-reference person targets. Score 1.

Dashboard rebuild — blocked (environment)

python3 bin/build-dashboard.py failed at shutil.rmtree(SITE) with PermissionError: Operation not permitted on iCloud sync-conflict directories inside Wiki/_meta/site/ (companies 2, concepts 2, entities 2, industries 2, models 2, primers 2, sources 2, standards 2 — all empty duplicate folders). These cannot be removed, moved, or rmdir'd from the sandbox (macOS/iCloud permission restriction), so the dashboard could not be regenerated this run. The 4 new pages are live and correct; only the generated HTML site is stale. Action for the user: delete the stray "<folder> 2" conflict directories under Wiki/_meta/site/ from Finder, then re-run python3 bin/build-dashboard.py.

One-line summary

Created 4 pages live (Effective Altruism, Regulatory Managerialism, Dropbox, Notion), fixed 2 link aliases (CSET, sycophancy) across 2 files; no sources pulled or queued. claude-code remains the standing item needing a user schema decision on product-page placement.