Real coding sessions · never self-reported
Where the tokens go
50% building · 1040.4M classified tokens
56 builds predate classification and aren't included.
Worked on albert-portfolio
3768 sessions tracked this month across all builders
2h 51m coding session using gpt-6-astra. 241K tokens across 5 prompts.
Portfolio Refresh Across Core Pages
1206 sessions tracked this week across all builders
Updated navigation, home, about, and projects pages, refined project scene and flight components, refreshed site content, and added a browser check artifact.
Project Page Styling Pass
1.4M tokens vs typical 192.4K
A local Codex session refined the project page styles in one CSS module over 7 prompts and about 70 minutes, with repeated shell checks used to validate the update.
Prototype page polish for albert-portfolio
947 sessions tracked this week across all builders
Updated the prototype page styles and component structure to refine the local build in albert-portfolio, with work centered in the prototype page files.
Coastal image asset and water check added
First build tracked for this project
A local Codex session updated the portfolio with a new coastal sand image, its license note, and a small water-check script after a focused build run.
Worked on explore-x20-2
879 sessions tracked this week across all builders
11m coding session using gpt-5.6-sol. 24K tokens across 2 prompts.
Worked on explore-x20-2
14h 9m
11m coding session using gpt-5.6-sol. 24K tokens across 2 prompts.
Worked on explore-x20-2
28 prompts
11m coding session using gpt-5.6-sol. 24K tokens across 2 prompts.
Worked on explore-x20-2
18 prompts
11m coding session using gpt-5.6-sol. 24K tokens across 2 prompts.
Debugged albertolive
14h 9m
14h 9m session running commands and testing. 516K tokens across 12 prompts using gpt-6-astra.
Debugged albertolive
14h 2m
14h 9m session running commands and testing. 516K tokens across 12 prompts using gpt-6-astra.
Debugged albertolive
1.4M tokens vs typical 194.4K
14h 9m session running commands and testing. 516K tokens across 12 prompts using gpt-6-astra.