A PRACTICAL AI LAB
AI agents.
Real work.
Open notes.
Research, teaching and everyday automation with Codex and OpenClaw. Explore actual work records, deliverables and the evidence used to check them.
An output is the beginning. Verification makes it useful.
THREE FIELDS. ONE QUESTION.
Where can agents make a difference?
FROM THE WORKBENCH
Notes from real projects
Building evidence appendices and editable presentation charts
A Bucheon Sangdong research-presentation workflow linked calculations to evidence appendices and refined chart comparisons and table labels. Windows Codex records were checked alongside a surviving presentation package.
Why an OpenClaw model comparison first needed a better evaluator
Four models completed 96 runs across eight tasks in the same secretary environment. The useful findings included a trace-parser defect, structured-output failures, and limits in the evaluation itself.
Connecting architecture-law slides, teaching notes, and quizzes
A Windows Codex workflow grew from a planned 50-minute lecture into illustrated teaching material and separate student quizzes. The case distinguishes recorded completion from files checked today.
Checking architectural translations at sentence and page level
How a lecturer’s terminology and editing decisions became repeatable Codex checks: translation samples, text-and-illustration separation, and a 337-page layout audit.
Preserving Telegram attachments with a verifiable local record
A Codex-built saving tool turned OpenClaw attachments into local files with an audit index. Account-selection and media-path failures were repaired, then saved files were checked against their recorded sizes and hashes.
Moving Windows to a new SSD: proving the machine boots
A Codex-assisted HP workstation migration, with separate checks for the backup, the data partition and two successful boots from the replacement SSD.
KEEP LEARNING
From the classroom to your next experiment.
Find course access, practical learning resources and answers grounded in the Lab’s notes.