← FP3NEWS · Home
source integrity · Source Integrity Layer · Operational Desk

source integrity Watch

The Signal Integrity Desk. Where open-source and social chatter is detected early, gated against contamination, and — when it survives verification — fed into calibrated FP3 forecasts.

Layer · source integrity (source integrity) Function · early-warning / signal triage Grammar · probabilistic, not adjudicative Sign-off · human · Direttore Responsabile

What this desk is — and is not

▲ It is
  • A detection layer: it tells us where to look, not what is true.
  • A tripwire for weak signals before they reach Tier-1 channels.
  • A corroboration engine: surviving signals feed forecasts.
  • A public demonstration of source integrity in motion on real cases.
▼ It is not
  • A source of record. Social never anchors a claim alone.
  • A debunking column. We do not rule items "true / false / absurd".
  • A mass-collection operation. Our edge is curation, not volume.
  • An auto-publishing bot. Nothing ships without a human signature.

How source integrity works this case

— the operational loop

The social stream is the fastest, dirtiest vector of data: high coverage, high speed, low Admiralty grade. SIGNAL INTEGRITY DESK does not trust it — it processes it. A signal travels through five stages before it is allowed anywhere near a published forecast.

01

Recon

Daily directed scan of social and open sources. Candidate signals captured, not yet trusted.

◉ Human · daily
02

Triage

Clustering of recurring claims; noise stripped; each candidate tagged by topic and apparent vector.

◇ AI-assisted
03

Confirm gate

The hard wall. Provenance traced to source-zero; numbers and quotes checked against Tier-1. No pass → capped as rumour.

◼ Gate · mandatory
04

Corroborate

Surviving signal mapped against existing FP3 theses (V12–V18, R7, cascades). Does it confirm, weaken, or complicate the prior?

◇ FP3 Method feed
05

Feed & log

Bayesian update where material; entry archived with falsifiers. Forecast-relevant items scored over time.

◉ Human sign-off
◉ Operational input · the human at stage 01

The desk is fed by directed daily reconnaissance — no less than one hour of structured social and open-source scanning by the Direttore. This is not passive monitoring: it is targeted hunting for weak signals, framing shifts, and pre-Tier-1 movement. SIGNAL INTEGRITY DESK is the discipline that turns that daily hour from browsing into intelligence — every candidate it surfaces is then forced through stages 02–05 before it earns any analytical weight.

The contamination gates

— stage 03, expanded

Stage 03 is where most social signals die — correctly. Four checks, any one of which can demote a claim below usable grade. This is the part that protects a credibility-grade outfit from amplifying a polished fabrication.

G1

Provenance tracing

Walk the claim back to source-zero. If it does not resolve past a social account, it stays a rumour — it never enters FP3 Method as evidence.

G2

Citation & figure verification

Every number, date and quote is checked against a Tier-1 source before use. This is the gate that strips invented specifics from an otherwise-true story.

G3

Synthetic-media check

Reverse-image and manipulation screening — with an honest caveat: detector false-negatives are rising, so absence of a flag is not proof of authenticity. Provenance (G1) stays more robust than the technical scan.

G4

Cross-source triangulation

A single source never promotes a claim above Tier-3. Independent corroboration required before a signal advances to stage 04.

Desk metrics

— KPI definitions · values populated on the register
≥ 1h
Daily directed recon
live
Signals triaged / week
live
Gate-03 promotion rate
live
Contaminated items caught

Counts are populated at the register, not asserted here — §17 discipline. The only fixed value is the committed daily recon cadence.

Signal log

— worked cases + open template slots

Each entry follows one fixed structure, so the reader sees the anatomy of how a signal was handled — not a verdict. Two cases are on the record; the remaining slots are templates, populated only with a verified source.

CW-INT-001 · International · AI / copyright ● On record · verified

"Project Panama" — the destroyed-books story

A true core wrapped in social contamination — the exact case source integrity is built for.

Claim
An AI firm bought millions of physical books, stripped the bindings, scanned them and destroyed the originals to train its model.
Zero-vector
Amplified through social tier (Instagram, Reddit, secondary blogs) before most readers reached the primary reporting.
Verification
Confirmed against Tier-1: Ars Technica (Jun 2025) and the Washington Post (Jan 2026) on unsealed court filings in Bartz v. Anthropic. Internal codename "Project Panama" documented.
Reading
Core: substantiated. The destructive scanning was real and ruled fair use; the larger violation — and the $1.5B settlement — concerned pirated downloads (7M+ titles), not the scanning.
Contamination
Social retelling carried an unverified specific ("hydraulic guillotines") and conflated figures (a "500k–2M" range mixed with the 7M pirated count). G2 strips these.
What it reveals
A story can be true at the core and contaminated at the detail simultaneously. The value is not "true/false" — it is isolating the load-bearing fact from the embellishment.
CW-IT-001 · Domestic · electoral ● On record · scored

Referendum Nordio 2026 — the NO front

Social sentiment as an early read on a binary outcome, resolved and Brier-scored.

Claim
Converging public-mood signals pointed to a NO majority on the justice referendum (22–23 Mar 2026) ahead of resolution.
Zero-vector
Distributed social discourse and mobilisation chatter — a classic detection-layer input, not a record source.
Verification
Resolved against the official result: NO 53.74%. The ex-ante forecast is logged as CL-2026-001.
Reading
Hit · resolved. See PULSE INSIDER N°005bis and the public Calibration Record.
Falsifier (ex-ante)
A YES majority, or quorum failure, would have falsified the call. Neither occurred.
What it reveals
Social signal is usable when it is resolved against an official outcome and scored — the discipline, not the chatter, is the product.
CW-INT-002 · International · [ open ] ◇ Template · awaiting verified source

[ International signal — to populate ]

Drop a verified international social-origin signal here through stages 01–05.

Claim
[ one-line statement of the signal ]
Zero-vector
[ where it first surfaced · which platform / account ]
Verification
[ Tier-1 source(s) · date · what was confirmed ]
Reading
[ probability / directional read · confidence % ]
Falsifier
[ 1–2 signals that would overturn the read ]
What it reveals
[ what the case teaches about the disinformation / signal mechanism ]
CW-IT-002 · Domestic · [ open ] ◇ Template · awaiting verified source

[ Domestic signal — to populate ]

e.g. a movement / sentiment shift that corroborates a standing FP3 thesis — once sourced.

Claim
[ one-line statement of the signal ]
Zero-vector
[ origin platform / account / timestamp ]
Verification
[ Tier-1 corroboration before publication ]
Reading
[ which FP3 thesis it corroborates · how strongly ]
Falsifier
[ disconfirming signals ]
What it reveals
[ mechanism insight ]

The grammar rule

— why this protects the desk

Probabilistic, never adjudicative

SIGNAL INTEGRITY DESK does not declare news items "false" or "absurd". A factual verdict is exposed and litigable; a calibrated probabilistic reading, documented with its source-mix and falsifiers, is a defensible analytical opinion. The grammar shift is the legal shield and the brand differentiator at once:

"This story is fake."  →  "This claim carries ~X% probability of being accurate, given this Tier-mix and these falsifiers."

Every published entry carries a human signature. An AI desk that auto-published a wrong correction on a credibility-grade brand would be the maximum reputational hazard — so the machine detects and drafts; the Direttore Responsabile verifies and signs. Forecast-relevant readings enter the Calibration Ledger with Brier scoring: an honest miss on the record strengthens the desk more than a claim of infallibility ever could.

⌂ Home Access & Engagement Calibration Record Electoral Trajectories The Unit ◈ PULSE Weekly Archive Methodology FP3 Method Spec Editorial Doctrine
MODENART SRLS  ·  P.IVA / C.F. 04522520230  ·  ATECO 70.21  ·  Negrar di Valpolicella (VR)
Method FP3 Method analytic framework · source integrity source-integrity layer  ·  Track record Calibration Ledger · ex-ante · Brier-scored · hits & misses on record
◉ DIREZIONE   Direttore Responsabile · Dott. Iacopo Mutascio  ODG Veneto · MODENART SRLS
SIGNAL INTEGRITY DESK · DETECTION LAYER · PROBABILISTIC GRAMMAR · HUMAN SIGN-OFF
Neither optimists nor pessimists. We take no side: we run the FP3 Method. Trajectories, not verdicts.
Né ottimisti né pessimisti. Non parteggiamo: applichiamo i protocolli FP3. Traiettorie, non verdetti.