15 tasks. March 3, 2026 closed at 126.6x weighted leverage across 992.0 human-equivalent hours in 470 minutes of wall-clock time. Supervisory leverage came in at 561.5x.
That is 24.8 weeks of human-equivalent throughput in 7.8 hours. The ceiling was 240.0x; the floor was 24.0x. 15 of the 15 entries came from a single project.
Task Log
| # | Task | Human Est. | Claude | Sup. | Factor |
|---|---|---|---|---|---|
| 1 | Course Mode + Exam Simulator (6 phases: lesson [content generation] engine + runtime personalization API + course UI + exam simulator engine + exam simulator UI + documentation — ~50 files across 5 repos) | 480.0h | 120m | 20m | 240.0x |
| 2 | leverage-manager core package + CLI + DB migrations + CSV import (160 records) — 4 design docs + 25 Python files | 80.0h | 22m | 8m | 218.2x |
| 3 | avian-electron Phase 2 + shared libraries (create @avian/ui library + add demo activities to @avian/activity-ui + migrate avian-demo + build electron Phase 2: 8 screens 50+ files) | 120.0h | 45m | 12m | 160.0x |
| 4 | Lesson caching + content chunking + adaptive toggle (11 files across engine + 3 UI repos) | 16.0h | 6m | 5m | 160.0x |
| 5 | Implement leverage-manager README + FastAPI server + React dashboard with info tooltips | 80.0h | 35m | 8m | 137.1x |
| 6 | Admin Review Tab + Team Leaderboard + Number Formatting (14 files) | 12.0h | 7m | 4m | 102.9x |
| 7 | Wire up avian-demo lesson/exam features (6 tasks: types+API+persistence+LessonScreen+ExamSimScreen+CourseStructure+App.tsx) | 16.0h | 10m | 5m | 96.0x |
| 8 | Leverage manager user/team management expansion | 24.0h | 16m | 5m | 90.0x |
| 9 | Debugging | 24.0h | 19m | 5m | 75.8x |
| 10 | iOS requirements doc for avian-demo clone (18-section comprehensive spec) | 24.0h | 25m | 5m | 57.6x |
| 11 | Knowledge Cosmos visionOS immersive experience (3 new files + 10 modified) | 40.0h | 45m | 8m | 53.3x |
| 12 | [engine subsystem] repair pipeline + docs + AP World History run | 40.0h | 53m | 8m | 45.3x |
| 13 | [engine subsystem] calibration auto-select judge ordering (6 files) | 8.0h | 12m | 4m | 40.0x |
| 14 | Beautiful-mermaid [ip] figure rendering pipeline | 16.0h | 25m | 5m | 38.4x |
| 15 | Runtime lesson personalization + docs + CLF lesson generation | 12.0h | 30m | 4m | 24.0x |
Aggregate Statistics
| Metric | Value |
|---|---|
| Total tasks | 15 |
| Total human-equivalent hours | 992.0 |
| Total Claude minutes | 470 |
| Total supervisory minutes | 106 |
| Total tokens | 3,801,500 |
| Weighted average leverage factor | 126.6x |
| Weighted average supervisory leverage factor | 561.5x |
| Human-equivalent weeks | 24.8 |
Analysis
The highest factor of the day came in at 240.0x and the lowest at 24.0x, a spread of 10.0 times between the two. That is a wide range for a single day, and it usually means the day mixed mechanical work with work that needed real judgement.
The largest single entry accounted for 480.0 of the 992.0 human-equivalent hours, or 48 percent of the day. One task carrying that much of the total is worth noting; the average is doing less work than it appears to.
Supervisory time was 106 minutes against 470 minutes of execution, a ratio of about 1 to 4. Supervisory leverage of 561.5x is the figure I find most honest, because it measures the hours I actually spent rather than the hours a machine spent on my behalf.
Every figure here is recorded at the time the work is done rather than reconstructed afterwards. The human estimate is my own judgement and carries the uncertainty that implies; the minutes and tokens are measured. The full dataset, including this day, is available for download.