active · lexical · first recorded 2026-07-02 · registers: cooking, parenting
“usually”, “mostly”, “typically”, “generally”, “slightly” — each is normal English; the density is the tell. Models qualify nearly every claim with one. Any single match means nothing; several per paragraph reads as AI.
Density tell — scored very low per match by design.
model specificity: cross-model; rising across generations
◑ appears in good human writing too (collateral score 0.32 — the closer to 1, the more it also fires on acclaimed pre-AI prose)
…is usually the main factor… mostly a texture issue… typically fine at room temperature…
occurrences per 10k words, measured 2026-08-17 (run #10, 3 current + 3 vintage models)
| current AI | vintage AI (2024/25) | humans (matched) | canon (great pre-AI prose) | |
|---|---|---|---|---|
| pooled | 24.57 | 13.12 | 10.48 | 7.96 |
| hn | 17.62 | — | 8.05 | |
| cooking | 49.96 | — | 20.53 | |
| parenting | 18.63 | — | 7.54 | |
| 3.55 | — | — |
| run | date | AI /10k | vs humans |
|---|---|---|---|
| #2 | 2026-07-02 | 23.91 | ×1.9 vs humans |
| #3 | 2026-07-02 | 25.70 | ×1.8 vs humans |
| #4 | 2026-07-06 | 24.09 | ×1.8 vs humans |
| #5 | 2026-07-13 | 23.20 | ×1.7 vs humans |
| #6 | 2026-07-20 | 23.79 | ×2.0 vs humans |
| #7 | 2026-07-27 | 27.21 | ×1.9 vs humans |
| #8 | 2026-08-03 | 28.06 | ×2.4 vs humans |
| #9 | 2026-08-10 | 27.86 | ×2.2 vs humans |
| #10 | 2026-08-17 | 24.57 | ×2.3 vs humans |
regex: \b(usually|mostly|typically|generally|slightly)\b