Primary-source intelligence in 30-45 minutes a day, deliberately minimal.

The decision to read fewer sources is the same shape as the decision to own fewer assets in a portfolio: concentration in your highest-conviction holdings beats diffuse coverage that you cannot actually monitor. Cal Newport's "Digital Minimalism" (2019) argues for explicit-attention budgeting; Nicholas Carr's "The Shallows" (2010) documents how heavy multi-source skim-reading rewires attention. The Simplified Desk operationalises both: pick the few sources that consistently move your decisions; ignore everything else.

The default failure mode in news consumption is diffuse subscription without diffuse reading — subscribing to 50 feeds, skimming 5, retaining nothing. The Simplified Desk inverts this: subscribe to 15, read 7-12 daily, retain decisions-grade insight. The full /desk/ is for users whose role makes 15 insufficient; for everyone else, 15 is plenty.

Each of the seven is signed-authority, primary-source, citable. None requires a subscription. All publish their major outputs as PDFs with stable URLs and ISBNs / DOIs. The cadence varies: WTO and WCO publish news daily; IMF's flagship outputs are quarterly (WEO, GFSR, Fiscal Monitor); BIS publishes a triennial FX survey plus quarterly statistics; UN Comtrade refreshes monthly with a 6-9 month lag.

Beyond the seven, useful secondary sources include the OECD Statistics portal (separate from the policy publications), UNCTAD World Investment Report and Trade and Development Report, the FAO Statistical Yearbook for agricultural-trade questions, and the ILO Statistics for labour-market context. These are essential for sector-specific work but optional for general trade intelligence.

Mechanics matter. Use a dedicated RSS reader (Feedly free tier, Inoreader free tier, NetNewsWire on Mac/iOS, FreshRSS self-hosted). Pin the daily-pulse feeds in a single folder. Process inbox-style: read the headline, decide read-now / save-for-weekend / skip; the headline-decision should take 3-5 seconds per item. Set a 15-minute hard cap with a kitchen timer or pomodoro app; when it rings, stop, regardless of how many items remain unread.

Avoid the four hidden cost centres: (1) the comments section — never; (2) the related-articles sidebar — never; (3) opening tabs to read later — either save to a queue or close; (4) social-media feed embedded in news pages — block via uBlock Origin custom rules. These four absorb 50-70% of news-reading time with negligible signal contribution.

The structural difference between daily and weekly reading is working memory load. Daily-pulse content is breaking news at low information-density; weekly content is dense argument with chains of reasoning. The cognitive infrastructure differs: daily reading happens in passive low-attention mode (over coffee, on a train); weekly reading deserves a desk, paper, pen, and the explicit Adler-Doren question framework: what is the book/paper about as a whole? what is being said in detail and how? is it true, in whole or in part? what of it?

Note-taking matters. The weekly deep-read goes into a permanent capture system (Zettelkasten-style atomic notes, Obsidian or Logseq, Cornell-method paper notes) rather than into ephemeral RSS-reader stars or browser bookmarks. The reason: a fact read once and not transcribed is functionally lost within 7-30 days (Ebbinghaus forgetting curve, replicated repeatedly).

Reading flagships effectively requires a different technique than weekly long-form. Skim chapter 1 fully for the macro-thesis; read the executive summary plus methodology section of any chapter that maps to your decisions; skip the data appendices until you have a specific question they can answer. Most professionals try to read the full report cover-to-cover, fail by page 80, and abandon the exercise. The right discipline is selective-deep over comprehensive-shallow.

Calendar mechanics: put each flagship release date in a recurring calendar block 2 weeks after the publication date (giving secondary commentary time to surface the disputed claims). Block 90 minutes for the read; another 30 for the note-summary into your permanent system. Without calendar blocking, quarterly reads slip indefinitely — the same forces that erode any non-deadline-driven knowledge work.

Mistrust signals let you screen quickly: (a) paywall-only with no methodology summary — can't verify; (b) API-only with no archive — tomorrow's claim cannot be checked against today's evidence; (c) anonymous attribution at scale — rumour-laundering; (d) graphs without axes or sources — suggests the headline matters more than the truth; (e) headline rhetorical questions — Betteridge's law applies (any headline ending in a question can be answered "no").

The citation-density rule is the simplest filter: open the article; count linked citations to primary sources in the first three paragraphs; if zero, the piece is downstream commentary; if 3+, it is doing primary-source synthesis. Apply this rule for 30 days and your news-quality improves measurably. Many otherwise-respected outlets fail this rule routinely.

The Simplified Desk explicitly targets 30-45 minutes per day total across daily-pulse + weekly-deep-read averaged. Beyond that range, additional reading rarely improves decisions but reliably increases anxiety and decreases focused-work output. The discipline is to set the dose, not let the dose set itself: if the daily-pulse 15-minute pomodoro is already enough today, do not open the weekly queue; if the weekly queue is already done, do not browse the "more from this source" sidebar.

Cal Newport's deep-work framework argues that the most valuable work is shallow-work-protected; news consumption is shallow work par excellence. The Simplified Desk treats news as one shallow-work bucket, scheduled and time-boxed, with no leakage into focused-work hours. Practical mechanics: phone has news apps deleted; browser bookmarks omit news sites; only the RSS reader (with an OS-level kitchen-timer) is the access path.

The empirical case for triangulation rests on Tetlock's superforecasting work (Tetlock-Gardner 2015) demonstrating that consistently-best forecasters explicitly seek disconfirming evidence and integrate across ideological frames. The mechanism is the avoidance of confirmation bias — the well-documented (Wason 1960, Nickerson 1998) tendency to over-weight evidence supporting prior belief. Triangulation provides the structural antidote that pure reflection cannot.

Practical setup: in your RSS-reader folder structure, group sources by topic, not by ideology — this forces the triangulation. A "Trade policy" folder with WTO, FT, WSJ, Reuters, Project Syndicate, Cato, China Daily, Hindustan Times produces ideological-and-geographic triangulation by default. Avoid the reverse error: dedicating folders to FT or WSJ, which entrenches frame-loyalty.

OPML (Outline Processor Markup Language, RFC since 2000, version 2.0 in 2007) is the universal RSS-subscription portability format. Every major reader imports OPML; every major reader exports OPML. This portability matters: changing readers should not mean rebuilding subscriptions. The platform's OPML export at /desk/opml/ contains the curated essentials (the 15 from Crucible 1, expandable to the 140 from full /desk/) with sector tags pre-applied so the imported reader inherits the folder structure.

Self-hosting note: FreshRSS on a $5/month VPS (DigitalOcean, Hetzner) plus a domain provides a permanent reader independent of any commercial provider; setup takes 30-45 minutes. For users uncomfortable with self-hosting, NetNewsWire (open source, native, no account required) on Mac/iOS or Inoreader free tier covers most needs. Avoid lock-in: never let your RSS pipeline depend on a service that does not support OPML export.

Graduation indicators: your decisions consistently require knowledge from a sector that the 15 essentials do not cover (e.g. agricultural-trade-specific FAO releases, healthcare-trade WHO releases, financial-services FSB releases); your role spans a region the essentials underweight (e.g. ASEAN-specific intelligence beyond what WTO / OECD provide); your work requires court-and-regulatory tracking (USITC, EU Court of Justice, WTO Dispute Settlement Body decisions). Each of these falls into the full /desk/ but not the Simplified.

Hybrid use is fine and common: keep the Simplified Desk's 15-source folder as the daily-pulse default; add 1-3 sector-specific folders from the full /desk/ for your specific decisions. Most users land here within 6-12 weeks of starting the Simplified Desk — the discipline of the simplified system creates the slack to add specific depth where it matters, rather than diffusing across the full breadth from the start.