Chapter 6 · Part 3
Make it yours — and use it wisely
You've got the workflows. Two things are left: making ChatGPT fit you so you stop repeating yourself, and the handful of habits that keep you safe and effective.
Set it up once
- Custom instructions — a settings field where you tell ChatGPT, once, who you are and how you like answers ("I'm a teacher; be concise; no emojis; explain your reasoning"). It applies to every new chat, so you stop re-typing your context.
- Memory — it can remember facts across conversations (your job, your preferences, ongoing projects). Handy, but check what it has stored, and clear anything you don't want it keeping.
- Files and images — upload a PDF, spreadsheet, or screenshot and ask about it directly. Great for "summarize this document" or "what's wrong in this photo of my code."
- Voice — talk to it hands-free, which is genuinely nice for brainstorming on a walk.
Start fresh when the topic changes
Everything you say stays in the conversation and influences later answers. When you switch to an unrelated task, start a new chat — otherwise leftover context muddies the results. Keep one chat per topic; it also makes things easy to find later.
The habits that matter most
Anything you paste may be stored, and by default may be used to improve the model. Never paste passwords, API keys, financial details, or other people's private information. For sensitive work, check your data-control settings, turn off chat history/training, or use a temporary chat — and follow your employer's policy on what's allowed.
- Verify what matters — the higher the stakes, the more you check. This is the through-line of the whole course.
- Iterate — the first answer is a draft; steer it with follow-ups.
- Give context — the Chapter 2 habit is worth more than any trick.
- Know when not to use it — for a decision only you can own, or where being wrong is costly and unverifiable, ChatGPT is an advisor, not the decider.
Where to go next
You now have the practical skills. Two courses go deeper on the pieces:
- How to Talk to an AI — the craft of prompting, one level beyond the context habits here: few-shot examples, chain-of-thought, and structured prompts.
- How ChatGPT Actually Works — why it's so fluent and why it hallucinates, so the strengths and blind spots in this course finally make sense.
The single sentence to remember: use it boldly for anything where a smart first draft helps, and verify anything where being wrong would cost you. Get that balance right and ChatGPT becomes one of the most useful tools you own.