EN Humanizer Audit — Google Discover Core Update Feb 2026: X.com +333%

Run 1 — 2026-03-03

Score: 7/10 Verdict: Authentic Patterns detected: 3 Translation artifacts: 1 Pipeline step: humanizer-en

Findings

#	Category	Pattern	Line/Section	Severity	Proposed Fix
1	Cat. 5 — Excessive Hedging	"in some cases, surprising"	Line 94	low	"surprising" or "unexpected in places" — delete "in some cases"
2	Cat. 3 — Vague Claims	"most significant business risk"	Line 190	low	"the highest-impact business risk" or quantify: "the update with the sharpest structural implications"
3	Cat. 9 — Structural	Missing contractions throughout prose	All prose sections	medium	Add natural contractions where they fit: "it's", "you're", "that's", "you'll", "can't" (see Priority Recommendations)

Translation Artifacts

#	Line	Original	Type	Fix
1	All prose	Zero contractions in ~700 words of prose	FR→EN register artifact	Insert natural contractions at 5 natural points (see Priority Recommendations)

Positive Signals

Hook is specific and stakes-driven: "rewrote the rules of content discovery overnight" (line 72)
Concrete named publishers throughout: The Sun, Reuters, The Guardian, The Independent.co.uk, CBSnews, NBCnews, Axios, AP News, CNBC, Forbes (lines 169)
Hard numbers dominate: 333%, 67%, 20%, 57%, 11%, 72.9 chars, 84.2 chars, 48.5 hours, 5.3x, 76.65%, 89.94% (lines 165–182)
Sly sincerity present: "the long-term signal weighting remains opaque" (line 186), "no Discover-specific tags or controls" (line 188), "cannot be resolved through content quality alone" (line 190)
Active voice dominant throughout; passive count = 1 per ~700 words
Direct "you/your" address throughout the implications section (lines 178–182, 190)
Section headers as genuine questions — no template phrasing
No generic opener, no "landscape" clichés, no inflated verbs (zero "leverage", "harness", "streamline")
Paragraph rhythm varies meaningfully; 4-item bullet list (non-AI-typical)
Data sourced and named: "NewzDash DiscoverPulse panel data" with specific methodology reference

Notes

The article is strong — it reads like a reporter who owns the data. The structural backbone (TL;DR, named data source, section-by-section breakdowns, "Our take" with honest caveats) is clean and functional. The single systemic gap is the complete absence of contractions across all prose sections, which is the clearest FR→EN translation fingerprint. This is a single-pass fix requiring 4-6 targeted insertions, not a rewrite. Score is 7/10 solid; with contraction fixes applied, it would reach 8/10.