active · lexical · first recorded 2026-07-02 · registers: hn, cooking
The pivot connective models use to perform balance. Twelve times the human rate in matched comparisons; the single strongest frame in our first experiment.
model specificity: cross-model; heaviest in gpt-5.4-mini
…the salt acts as an additional preservative. That said, there are some practical guidelines…
occurrences per 10k words, measured 2026-09-28 (run #15, 3 current + 3 vintage models)
| current AI | vintage AI (2024/25) | humans (matched) | canon (great pre-AI prose) | |
|---|---|---|---|---|
| pooled | 5.74 | 5.48 | 0.12 | 0.23 |
| hn | 9.07 | — | 0.00 | |
| cooking | 3.74 | — | 0.00 | |
| parenting | 4.81 | — | 0.40 | |
| 0.00 | — | — |
| run | date | AI /10k | vs humans |
|---|---|---|---|
| #2 | 2026-07-02 | 6.69 | ×12 vs humans |
| #3 | 2026-07-02 | 5.77 | ×4.4 vs humans |
| #4 | 2026-07-06 | 5.95 | ×8.4 vs humans |
| #5 | 2026-07-13 | 5.77 | ×9.2 vs humans |
| #6 | 2026-07-20 | 6.18 | ×56 vs humans |
| #7 | 2026-07-27 | 5.05 | ×8.5 vs humans |
| #8 | 2026-08-03 | 6.88 | ×55 vs humans |
| #9 | 2026-08-10 | 5.56 | ×5.5 vs humans |
| #10 | 2026-08-17 | 6.59 | ×19 vs humans |
| #11 | 2026-08-31 | 7.41 | ×10 vs humans |
| #12 | 2026-09-07 | 6.17 | ×13 vs humans |
| #13 | 2026-09-14 | 5.58 | ×12 vs humans |
| #14 | 2026-09-21 | 6.03 | ×10 vs humans |
| #15 | 2026-09-28 | 5.74 | ×47 vs humans |
literal: that said