Skip to main content
AI28 April 2026

Infrastructure and testing

11 tasks. April 28, 2026 closed at 33.4x weighted leverage across 203.5 human-equivalent hours in 366 minutes of wall-clock time. Supervisory leverage came in at 297.8x.

That is 5.1 weeks of human-equivalent throughput in 6.1 hours. The ceiling was 64.0x; the floor was 6.1x. 6 of the 11 entries came from a single project.

Task Log

#TaskHuman Est.ClaudeSup.Factor
1Audit remediation wave 2: audit_logs middleware (SOC 2 CC7.2), GDPR export+deletion workers, RDS encryption runbook, parity drift inventory, 29 doc-audit + 15 readiness MEDIUMs across ~25 repos80.0h75m2m64.0x
2Author 86 type-B coverage labs across 10 cloud certs (AI-900, DP-420, DP-300, AZ-900, AZ-700, AZ-305, PCNE, PCSE, PDE, PGWA); strict-pass clean; 100% leaf coverage across all 43 cloud certs80.0h77m8m62.3x
3Design and frontend6.0h11m3m32.7x
4auth-service: fix pre-existing legal-dir vite build via predev/prebuild sync-legal.cjs script + gitignore; verify avian-app-web inherits Apple/Google sign in via existing OIDC redirect (no client-side code…1.5h4m2m22.5x
5Deployment8.0h30m3m16.0x
6avian-app-web: fix Flashcards fake-MCQ fallback (static @avian/activity-ui import + remove FallbackMCQ), add sidebar tooltips, suppress section-code flash on lesson load, redeploy prod engine container with…4.0h17m4m14.1x
7Modal centering fix + CLF-C02 lab 9 missing policy + new instruction-vs-UI audit script (2 passes, 6 catalog scrapers) + 16 lab uiSteps/checkpoint fixes + avian-audits content-audit wiring + labs CLAUDE.md…8.0h35m4m13.7x
8Infrastructure10.0h60m7m10.0x
9Debugging0.5h4m1m7.5x
10Coding0.5h4m1m7.5x
11Testing5.0h49m6m6.1x

Aggregate Statistics

MetricValue
Total tasks11
Total human-equivalent hours203.5
Total Claude minutes366
Total supervisory minutes41
Total tokens2,445,000
Weighted average leverage factor33.4x
Weighted average supervisory leverage factor297.8x
Human-equivalent weeks5.1

Analysis

The highest factor of the day came in at 64.0x and the lowest at 6.1x, a spread of 10.5 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 80.0 of the 203.5 human-equivalent hours, or 39 percent of the day. No single task dominated the total, so the weighted average is representative.

Supervisory time was 41 minutes against 366 minutes of execution, a ratio of about 1 to 9. Supervisory leverage of 297.8x 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.