Author: Jose S
How to check the results of a long, multi-step ChatGPT Work job
Part 3 of the ChatGPT Work tutorial. Run multi-step jobs as a goal-plan-act-checkpoint loop: review the plan, steer mid-run, check usage, and finish with a result checklist that…
What does bad data mean? How to name the problem so you can fix it
Bad data is a shrug, not a diagnosis. Learn five practical quality dimensions (completeness, accuracy, consistency, timeliness, uniqueness) with workplace examples so you name the failure mode before…
How to write a CLAUDE.md or AGENTS.md file so Claude Code follows your rules
Part 6 of the Claude Code tutorial. How CLAUDE.md, AGENTS.md, README, and related instruction files split work between humans and agents, with a sample CLAUDE.md skeleton and the…
How to give ChatGPT Work its first task and approve each step
Part 2 walks one low-risk agentic task with approvals: owned notes into a short status deck, tight tool access, and human review before anything leaves your desk. First…
What Claude Code remembers between sessions, and what it forgets
Part 5 of the Claude Code tutorial. Chat resets with /clear or a new session, CLAUDE.md reloads every time, auto memory can carry learnings if you verify it…
