What I am working on
The few things I am giving my attention to, and the points at which I will reconsider them.
This is not a complete project list, just the few directions receiving my attention.
Reading, training, and everyday writing remain part of the blog. There are no new review results for the routine experiment; its original record date remains, without moving the review date forward.
2026-11-10: The two-week training trial has been reviewed, continuing into three strength sessions and one weekend badminton session per week. Training cards 2.4.0 is live. Two-device upgrades, sync, and offline reopening on the phone passed real-device checks; first session creation and exercise entry await actual training. Checklist now reflects the ongoing three-day phase and a reference target of three sessions. Training progress does not establish acceptance of the whole routine experiment or long-term physique results.
2026-10-08: I have abandoned the active investing system project and am no longer working on it. Related past articles remain available.
Edited English version based on the Chinese original. Read the Chinese original
Codex Agents
In progressWhat I am exploring
Current Codex Agents source has two default everyday entry points, manuel and mantou, with optional specialist roles. Colleagues and friends have used earlier versions repeatedly; the new integration and research workflow are on trial.
Next checkpoint
Record goal confirmation, output problems, and manual corrections in real tasks with the new behavior, then judge whether the integration helps.
Last reviewed:
Daily routine experiment V1
In progressWhat I am exploring
Can I reorganize my sleep and energy supply to reliably reclaim the three hours from 5:00 to 8:00 each morning, with at least 90 minutes devoted to my “Beyond Boundaries” reading and critical-thinking practice?
Next checkpoint
Finish the first round and check whether I get enough sleep, have steady daytime energy, meet my weekly reading target, and find that morning exercise helps rather than disrupts reading. After reviewing the data, change only one main variable.
Last reviewed: