§ What it observes
Four input streams, read against each other rather than summarised in isolation.
01
Satellite anomaly detection
Global 2° grid of NDVI, LST, SST and thermal signals, scored against rolling baselines. Anomalies surface as cells and clusters, not pixels.
02
Shipping & freight
AIS vessel flows, chokepoint throughput, airspace and freight patterns. Physical-economy movement, read as indices rather than headlines.
03
Regulatory filings
SEC, OFAC, FDA and their European equivalents — surfaced by impact model, not by volume. The filings nobody else is highlighting yet.
04
Structural-shift briefs
Narrative synthesis once the evidence warrants it. Briefs name the regime change, the counter-thesis, and what would falsify both.
Method-led. Every brief passes a three-exit integrity check, carries an explicit counter-thesis, and ships with falsifiable claims and calibration records.
— Editorial standard
§ Method
The brief is a falsifiable object, not a narrative. Four disciplines enforce that.
Integrity check
The three-exit gate
A draft has three opportunities to die: evidence insufficient, counter-thesis stronger than thesis, or no observable proxy would change our mind. A brief that survives all three is published — the rest are held or retired.
Steel man
Explicit counter-thesis
Every brief states the most honest argument against itself, as prominently as the thesis. If the counter-thesis can't be written, the thesis isn't mature enough to publish.
Pre-registration
Falsifiable claims only
Claims are worded so a future observation can confirm or reject them: specific proxies, specific windows. Vague directional language is treated as a drafting failure.
Accountability
Calibration records
Resolved claims are tracked against their pre-registered outcomes. The hit rate is part of the archive, not marketing copy — readers can see where the method works and where it doesn't.
§ Author
Written by one person, under one name.