FINDING #85 · UNIT ID 1321790489
leonickson1/Swiftlet
Swiftlet is a Swift and Metal runtime that runs large Qwen Mixture-of-Experts models locally on Apple devices by streaming expert weights from storage, enabling 35B and 80B models to run with low RAM, including on iPhone.
SURPRISE SCORE
0.00
Score Breakdown
SCORE = SURPRISE × ENGAGEMENT^0.7 × FRESHNESS × VISIBILITY × CONFIDENCE
SURPRISE = WINDOW STARS / DAYS / (AUDIENCE + FLOOR)
10% OF STARS IN ARCHIVE
Growth Telemetry
VELOCITY /D
15.67
ACCEL
-12.50
RETENTION
15.3%
PEAK 2026-08-06 · FORK-RETENTION 0.0% · 47 STARS / WINDOW
Author Audience
AUDIENCE
64
FOLLOWERS
6
OWNER ★
581
Engagement Signals
FORKS
20
ISSUE AUTH
0
PR AUTH
0
UNIQUE STARGAZERS 47 / 47 (DIVERSITY 1.00)
Why This Is A Finding
leonickson1/Swiftlet собрал 47 звёзд за окно, тогда как у автора всего 6 подписчиков — эффективная аудитория ≈ 64. Это даёт surprise-индекс 0.15 (звёзды относительно охвата автора, а не в абсолюте). Удержание форков 0.0% и 0 внешних контрибьюторов отделяют реальный инструмент от разовой вспышки. Акселерация отрицательная — внимание остывает после пика.
Related Findings
RANKS ABOVE 99% OF 6512 FINDINGS
METRICS IN CONTEXT
MEDIAN ACROSS ALL 6512 FINDINGS · Δ vs MEDIAN · PERCENTILE = SHARE RANKED BELOW
METRICVALUEMEDIANΔ MEDPERCENTILE
SCORE0.150.00+0.15ABOVE 99%
VELOCITY15.675.33+10.33ABOVE 83%
RETENTION15.3%16.7%-1.4 PPABOVE 49%
FORKS20419-399ABOVE 9%
SURPRISE0.150.00+0.15ABOVE 97%