Skip to main content
AI07 March 2026

Coding and documentation

34 tasks. March 7, 2026 closed at 54.5x weighted leverage across 500.0 human-equivalent hours in 550 minutes of wall-clock time. Supervisory leverage came in at 206.9x.

That is 12.5 weeks of human-equivalent throughput in 9.2 hours. The ceiling was 180.0x; the floor was 14.4x. 25 of the 34 entries came from a single project.

Task Log

#TaskHuman Est.ClaudeSup.Factor
1Phase 3+4: sync/offline layer search dark mode Docker deployment (22 files 97 tests)24.0h8m2m180.0x
2Pipeline orchestrator + App M [ip] expansion + portfolio doc updates120.0h45m10m160.0x
3Build competitive multiplayer mode for avian-demo (16 files: 7 components + mock server + WebSocket hook + App integration)40.0h25m2m96.0x
4Pipeline orchestrator: full [content generation] lifecycle manager with WebSocket/CLI dual transport 22 files40.0h25m5m96.0x
5Full code review of renkara-cloud and accelastudy-cloud codebases with updated docs and READMEs16.0h12m2m80.0x
6Generate HP domain specs books 5-7 (304+300+295 leaf goals with prereqs)8.0h8m3m60.0x
7Generate ASOIAF domain spec JSON files for books 1-3 (911 leaf goals)8.0h8m5m60.0x
8Draft Family Differentiation Memo for AVIAN [ip] portfolio (A-M)4.0h4m3m60.0x
9Harry Potter trivia syllabi - 7 books x 65 leaf goals with goal IDs and tier annotations8.0h8m5m60.0x
10Generate ASOIAF Books 4-5 domain spec JSON (602 goals total)8.0h8m3m60.0x
11Create Series A and Series B pitch decks with growth projections and enterprise revenue models24.0h25m5m57.6x
12Rewrote Harry Potter trivia syllabi with 2197 goals across 7 books (300-400 per book x 10 domains)24.0h25m5m57.6x
13Build Cross-Domain Semantic Intelligence Engine (backend service + API + frontend page)40.0h45m8m53.3x
14Update all business planning docs: enterprise training market integration + [ip] portfolio 14/306/76 + 276K MCQ + create private repo16.0h20m5m48.0x
15Generate AVIAN domain specs for Harry Potter books 1-4 (1197 leaf goals)8.0h10m5m48.0x
16Debugging16.0h25m5m38.4x
17Fix all remaining code review issues (25+ issues across security/bugs/modernization/performance/quality)8.0h15m5m32.0x
18Coding6.0h12m5m30.0x
19Update business plan for bootstrapped scenario (8 sections rewritten)2.0h4m3m30.0x
20Create Combination Matrix for AVIAN CIP counsel packet with all reference pairings and missing element analysis4.0h8m3m30.0x
21Create Citation Appendix linking defense points to file+line citations4.0h8m3m30.0x
22Phase 3 sync & offline layer for AccelaStudy web app (9 files)4.0h8m5m30.0x
23Coding16.0h35m3m27.4x
24Design competitive multiplayer mode for avian-demo (phases + components + WebSocket + wireframes)6.0h15m3m24.0x
25Eliminate Type 3 fonts from all figure PDFs - Inkscape text-to-path pipeline fix + full 13-app regeneration4.0h10m2m24.0x
26Add L3 scenario clustering and --scenarios-only incremental regeneration mode8.0h20m5m24.0x
27Fix [ip] audit findings - App M [ip-claim] multi-judge spec support / App L runtime benchmark / App D+G cross-reference softening3.0h8m2m22.5x
28Build search page and dark mode toggle for AccelaStudy web app3.0h8m5m22.5x
29Merge renkara-cloud library into accelastudy-cloud monorepo (8-phase merge with branch creation - copy files - update imports - merge deps - sync schemas - update Docker - fix tests - update docs)3.0h8m5m22.5x
30Rewrite Game of Thrones syllabi with 300+ goals per book (1513 total across 5 books and 10 domains)8.0h25m5m19.2x
31Generate HP books 1-4 domain spec JSONs (368+305+305+313 leaves)6.0h20m5m18.0x
32Fix 3 code review issues: Apple JWT verification + EnrollmentManager pool mismatch + sync file I/O caching2.0h8m5m15.0x
33Documentation3.0h12m5m15.0x
34Fix question generator pair-node ID mismatch bug (2% match rate due to stale IDs from regeneration cycles)6.0h25m3m14.4x

Aggregate Statistics

MetricValue
Total tasks34
Total human-equivalent hours500.0
Total Claude minutes550
Total supervisory minutes145
Total tokens3,458,500
Weighted average leverage factor54.5x
Weighted average supervisory leverage factor206.9x
Human-equivalent weeks12.5

Analysis

The highest factor of the day came in at 180.0x and the lowest at 14.4x, a spread of 12.5 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 24.0 of the 500.0 human-equivalent hours, or 5 percent of the day. No single task dominated the total, so the weighted average is representative.

Supervisory time was 145 minutes against 550 minutes of execution, a ratio of about 1 to 4. Supervisory leverage of 206.9x 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.