Methodology

Two pipelines, two philosophies

Substrate Wire runs two separate editorial pipelines:

• Shadow War pipeline — OSINT signals collected from telemetry sources (GPS jamming monitors, satellite thermal detection, internet outage trackers, aircraft transponder data). Synthesis and narrative writing run through a bounded editorial process that keeps uncertainty labels visible.
• Frontier news pipeline — AI, geopolitics, markets, health, space. Articles collected from RSS and web sources, pre-filtered, then ranked and summarized through a bounded editorial process with source and privacy checks.

Shadow War pipeline

Signal sources are machine-telemetry systems that detect anomalies against their own baseline. These are not news articles — they are:

• GPSJam — GPS interference monitor (aircraft transponder data)
• NASA FIRMS — satellite thermal detection
• OONI — censorship and internet access monitor
• Cloudflare Radar — internet traffic anomaly detection
• IODA — internet outage and BGP route monitor
• USGS — seismic event detection
• ADS-B Exchange — military aircraft transponder tracker

The narrative brief and per-signal hypotheses are generated from retained public inputs, then constrained by evidence rules that separate observed facts from interpretation.

Frontier news pipeline — local + cloud assist

Articles go through a multi-pass pipeline:

1. Collection from RSS and web sources across 11 sections
2. Deduplication and paywall filtering
3. Pre-editorial filtering trims low-signal articles
4. Final editorial ranking, importance scoring, and summarization
5. Lead promotion and daytime refresh on promotion events

Today's Five two-gate rule

Today's Five is the protected brief. It shows up to five numbered stories, and a story appears only when it clears both gates:

Gate 1 — evidence

1. PRIMARY SOURCE — a government agency, company, or other institution reports its own action, notice, release, rule, warning, or primary document through a safe public link. That source establishes what it did; independent reporting may still change the context.
2. 2+ INDEPENDENT SOURCES — at least two fetched reports support the item, carry timestamps and safe public links, and come from distinct parent organizations.
3. SINGLE-SOURCE REPORT — one established publisher is linked directly and clearly attributed. It is not presented as independently confirmed, so it must clear the higher 60-point attention threshold.

Gate 2 — material consequence

A credible story is not automatically one of the day's five most important stories. The second gate requires a MATERIAL CONSEQUENCE, scored from source-visible facts rather than category keywords. It tests magnitude, relevance, and concrete consequence:

• Magnitude and breadth — how many people, institutions, jobs, services, or dollars are actually affected.
• Mass-casualty signal — a source-headline count of at least 10 confirmed dead, 100 missing, or 50 injured or hospitalized can qualify even when the hazard type is not on a predefined topic list. Boilerplate saying a product poses a risk is not confirmed harm.
• Decision utility and immediacy — whether readers may need to change a safety, money, work, security, travel, or household decision now.
• Structural change — whether law, policy, rates, markets, war, infrastructure, or a broadly available technology capability materially changed.
• Explicit impact — a sourced explanation of the real-world effect; an official label, topic name, or high upstream importance score is not enough.
• Penalties — routine administration, personnel actions, specialist research, and incremental publications are demoted unless the source also demonstrates broad material impact.

A primary or independently corroborated story must score at least 50 of 100 and contain a concrete consequence signal. A single-source report must score at least 60. Scores of 75 or more are marked high consequence. Survivors are ordered by consequence score, evidence class, freshness, and a stable key. Household geography is not a tie-break. Cross-publisher reports of the same event collapse to the strongest sorted report before the brief applies its caps. The brief stops at two stories per topic lane and two per publisher; those caps never relax to force five filled slots.

• Raw source totals, repeated syndication, engagement, source authority, freshness, and AI confidence never clear a slot by themselves.
• Recall notices stay in Official Need-to-Know instead of using one of the five broad-news slots.
• Stories sourced only from the local Anza lane stay in that local section. They do not enter the global five.
• An X or Grok candidate never uses the single-source path and never clears the five by itself; it still needs non-X corroboration or a linked official primary document.
• If fewer than five stories clear, the ordered list contains only the genuine stories. One compact line reports how many of the five maximum positions were left unfilled; if none clear, one compact zero-state replaces the list. Empty is better than filler.
• The first qualifying story is visually emphasized as the lead. Open / Official Need-to-Know follows the brief, while reports that do not clear remain available farther down in the collapsed Unranked reports drawer with separate evidence and consequence labels.
• The device-only URL-set change receipt counts links added or dropped. It does not claim a story was held, changed, or proved wrong.

AI-only: no human editor approves a slot. Deterministic evidence-class, consequence, provenance, and privacy gates run before the AI-assembled page is published.

The design draws on public work that separates unpersonalized top news from personalized news, treats magnitude and relevance as news values, evaluates diversity alongside ranking accuracy, incorporates editorial values beyond clicks, and uses abstention when confidence is inadequate: Google News, Harcup & O'Neill, Google Research, Beyond Optimizing for Clicks, and selective prediction research.

Source tiers

Every story is assigned a source tier before editorial scoring:

• Tier A — Wire services: AP, Reuters, AFP. Highest factual weight.
• Tier B — Major national papers, established broadcasters (NYT, BBC, WSJ, FT).
• Tier C — Trade and sector publications (TechCrunch, Breaking Defense, SpaceNews).
• Tier D — Blog posts, speculative analysis, opinion. Shown collapsed.

Tier D items are not promoted to lead or top stories regardless of engagement. A source tier affects ranking, but it never substitutes for either Today's Five gate.

Refresh cadence

The nightly editorial pipeline runs at 02:15 local time. Daytime refresh-merge runs 5 times per day to promote breaking stories and expire stale ones. Shadow War signals are collected on a separate schedule.

Hallucination mitigations

• Every visible story links to the reporting or official document it came from; evidence labels distinguish primary, corroborated, single-source, and uncorroborated items.
• Direct quotes are only used when pulled verbatim from a cited source.
• Shadow War signals are machine-telemetry anomalies — not news interpretation.
• Items below the protected-brief threshold are held out of Today's Five rather than padded into an empty slot.
• If you spot an error, contact [email protected].

Prediction track record

Shadow War analyst items are timestamped at write time. Each carries a confidence tier (A/B/C/D) and a falsifiable claim. Hit/miss/partial outcomes are logged in the append-only prediction record.

Resolved predictions are logged at /track-record. The raw append-only prediction log keeps edit history auditable.

What is excluded

• No content generated from scratch — only summarization of sourced articles.
• No unsourced claims about named private individuals.
• No affiliate links, no ads, no tracking pixels.

About · Disclaimer · Support