You don't need a paragraph of instructions every time you open ChatGPT. Sometimes one specific request is enough to give the conversation a much better direction. These 50 prompts help you ask better questions, challenge an idea, write something useful, learn a subject, or figure out what to do next.
The plan
- 1Start with the eight prompts from the video, in the same order.
- 2Pick from 42 more based on what you're working on.
- 3Replace the brackets, paste your context, and follow up on the answer.
Before you start
Bring something real
A question, draft, idea, decision, or set of notes.
Fill in the brackets
Replace anything in [square brackets] with your details. Add the text or example wherever the prompt asks for it.
Use one at a time
Paste a prompt into a new chat with the relevant context, or use it as a follow-up in the conversation you're already having.
If a prompt says "this," include the thing you want ChatGPT to work on. Give it your actual goal, the audience, and any limits that matter. Clear requests and useful context give it a better starting point.
Maverick's tip
Try one prompt on an answer you already have. Number 2 challenges the idea, number 24 cuts the filler, and number 37 turns it into a first action.
These are reusable instructions, not hidden settings. A role or example can guide the response, but it doesn't give ChatGPT real experience or make an answer automatically correct. Check important facts and open the sources it gives you.
Find What You Need
| Prompts | What they help with |
|---|---|
| 1 to 8 | The eight from the video |
| 9 to 12 | Get clearer answers |
| 13 to 20 | Decisions and tradeoffs |
| 21 to 28 | Writing and communication |
| 29 to 36 | Learning and practice |
| 37 to 44 | Plans and ideas |
| 45 to 50 | Verification and useful follow-ups |
The eight from the video
01. Ask me three questions first
Use this when: The answer depends on details you haven't explained yet.
02. Prove me wrong
Use this when: You want to find the weak points before committing.
Your only job right now is to try to prove me wrong about [my idea or claim]. Give me the strongest objections and the evidence that would matter. Don't invent problems just to disagree.
03. The safe, smart, and bold play
Use this when: You want real options with different levels of risk.
Give me three versions of [my plan or response]: the safe play, the smart play, and the bold play. Explain the main upside and tradeoff of each in one sentence.
04. See it through someone else's eyes
Use this when: You want a different perspective on something you're making.
How might [person or audience] react to [my idea, draft, or offer]? Give me their likely praise, objections, and first question. Use the background I provide or documented public views, and label the reaction as a simulation.
Maverick's tip
For number 4, try Steve Jobs, Warren Buffett, or your toughest customer. A specific customer description is often more useful than a famous name. Treat the response as a perspective to explore, not a prediction of what someone really thinks.
05. Set the quality bar
Use this when: You have an example of what good looks like.
Here's the quality bar: [paste example]. Study what makes it work, then create [my deliverable] to that standard for [my audience]. Match its clarity and level of detail without copying its wording or inventing facts.
06. Explain it like I'm ten
Use this when: The explanation is getting too technical to be useful.
Break down [topic or pasted text] like I'm ten years old. Use plain words and one everyday example. Keep the important details accurate, and explain any technical term you need to use.
07. Make the answer easy to check
Use this when: You need to understand how a solution works.
Work out [problem] carefully. Give me the answer with a short explanation of the key steps, assumptions, and any calculations I need to check it.
08. Get a specialist's perspective
Use this when: You need advice shaped around a specific kind of work.
Answer [my question] from the perspective of a [specific expert] with 20 years of experience in [specific field]. Focus on practical tradeoffs, common mistakes, and what to do first.
Maverick's tip
Number 7 keeps the video's idea of a careful answer, with a more useful request: an explanation you can check. OpenAI says reasoning models already reason internally, so asking them to think out loud is not a reliable upgrade.
Get clearer answers
09. Find the missing context
Use this when: An answer sounds reasonable but might rest on guesses.
For [my request], list the details you're missing and the assumptions you'd otherwise make. Identify the one missing detail most likely to change your answer, and ask me for it.
10. Fix the question
Use this when: You know the goal but aren't sure what to ask.
My goal is [goal], and the question I was going to ask is [question]. Rewrite it as a clearer, more useful question. Point out any assumption in the original that could send us in the wrong direction.
11. Give me the answer in layers
Use this when: You want the headline before deciding how deep to go.
Answer [question] in three layers: one sentence, five bullet points, and one practical example. Keep the same conclusion across all three.
12. Follow this exact format
Use this when: The information is useful but difficult to scan.
Turn [my notes or your last answer] into a table with these columns: [column names]. Use only the information provided, keep each cell brief, and mark missing details as unknown.
Make a better decision
13. Compare the things that matter
Use this when: A long pros-and-cons list isn't helping you choose.
Compare [option A] and [option B] for [my situation] using [my three most important criteria]. Show the tradeoffs, recommend one, and tell me what new information would change that recommendation.
14. Find the hidden cost
Use this when: An option looks good until you count the extra work.
What hidden costs come with [this plan or choice]? Consider time, money, maintenance, and what I'd have to give up. Separate costs supported by my details from possibilities I should investigate.
15. Imagine the plan failed
Use this when: You want to spot preventable problems before starting.
Imagine [my plan] has failed six months from now. Give me five plausible reasons, the early warning sign for each, and one action I could take now to reduce that risk.
16. Test it before committing
Use this when: You need evidence before spending more time or money.
What's the smallest practical test of [my idea] I could run in [time] for under [budget]? Define what to do, what to measure, and what result would support continuing, changing, or stopping.
Think past the obvious
17. Make the strongest case for both
Use this when: You want to understand a disagreement before picking a side.
Make the strongest fair case for and against [decision or claim]. Separate factual disagreements from differences in priorities, then name the evidence that would help resolve the factual part.
18. Work backward from the result
Use this when: You have a destination but no clear path to it.
I want [result] by [date]. Work backward to the milestones that would need to happen, using [my starting point and constraints]. Flag anything that makes the deadline unrealistic.
19. Look at what happens next
Use this when: You're focused on the immediate payoff of a choice.
If I choose [option], what might happen immediately, three months later, and a year later? Show likely benefits, possible side effects, and the assumptions behind this scenario.
20. Decide what you can undo
Use this when: Every part of a decision feels equally permanent.
Break [my decision] into parts I can easily reverse and parts that are expensive or difficult to undo. Suggest what I can test now and what needs more evidence first.
Write something people understand
21. Make it sound like me
Use this when: A draft says the right thing in the wrong voice.
Use these writing samples to revise [my draft] in my voice: [paste samples]. Match the tone, rhythm, and vocabulary while preserving the meaning. Don't add experiences, opinions, or facts I haven't supplied.
22. Write for one actual reader
Use this when: The copy is trying to speak to everyone at once.
Rewrite [this text] for [specific audience] who cares about [priority] and is worried about [concern]. Lead with what matters to them and end with [the action I want them to take].
23. Keep my voice, fix the confusion
Use this when: You want a light edit without a full rewrite.
Edit [my draft] for clarity and flow while changing as little as possible. Keep my voice and meaning, then briefly explain the three most useful edits.
24. Cut the filler
Use this when: The point is buried under too many words.
Cut [this text] to about half its length while keeping the main point, necessary facts, and action. Remove repetition, vague claims, and throat-clearing. Don't replace specific details with generic language.
Make the message land
25. Find a better opening
Use this when: You have a solid piece with a weak first line.
Write five different opening lines for [this piece] aimed at [audience]. Try a question, a contrast, a concrete problem, a surprising detail from the text, and a direct benefit. Don't invent facts or overpromise.
26. Make the abstract concrete
Use this when: The message sounds impressive but doesn't say much.
Find the vague or abstract sentences in [this draft] and rewrite them with concrete language. Use details already provided, and leave a clear placeholder wherever a real example or number is missing.
27. Draft the awkward message
Use this when: You need to be direct without making things worse.
Draft a short message to [person or role] about [situation]. Be [tone], state [my boundary or request] clearly, and avoid blame or unnecessary apologies. Give me a draft to review, not a message to send.
28. Answer the objection
Use this when: Someone has a concern you need to address honestly.
Help me respond to this objection: [objection], about [offer or idea]. Acknowledge the valid concern, explain what my actual evidence supports, and suggest a useful next step. Don't add claims I can't back up.
Learn it so it sticks
29. Find the foundation
Use this when: A new topic feels too big to start.
I want to learn [topic] so I can [goal]. Pick the five foundational ideas I should understand first, explain why each matters, and put them in a sensible learning order.
30. Test what I actually know
Use this when: Reading makes sense but recall is another story.
Quiz me on [topic] at [my level], one question at a time. Wait for my answer, explain what I got right or missed, and adjust the next question based on my response.
31. Check my explanation
Use this when: You want to see whether you truly understand something.
I'll explain [concept] in my own words: [my explanation]. Identify mistakes, missing connections, and anything I've made unnecessarily complicated. Then ask me one question that tests the biggest gap.
32. Connect it to something familiar
Use this when: A definition isn't helping the idea click.
Explain [new concept] using an analogy from [something I understand]. Map the main parts, then tell me where the analogy breaks down so I don't take it too literally.
Turn understanding into practice
33. Show one worked example
Use this when: You understand the rule but can't apply it yet.
Show me one worked example of [skill or problem] at [my level]. Label the key steps and explain why they matter, then give me a similar problem to try without showing its solution yet.
34. Find the beginner mistakes
Use this when: You're starting a skill and want to practice well.
What are five common beginner mistakes in [skill]? For each, show what the mistake looks like, how to spot it, and a small practice exercise that helps correct it.
35. Build a realistic practice plan
Use this when: You have limited time and need a starting routine.
Build a seven-day practice plan for [skill] using [minutes] a day. Start from [my current level], give me one exercise per day, and define an observable result I can check at the end.
36. Untangle two similar ideas
Use this when: You keep mixing up related terms or concepts.
Explain the difference between [concept A] and [concept B]. Give me a simple definition, one example of each, and a borderline case that makes the distinction clearer.
Turn the mess into a plan
37. Find the next ten-minute action
Use this when: A project is so big you keep putting it off.
Here's what I'm trying to do and where I'm stuck: [details]. Give me one useful action I can finish in ten minutes, what I need to start, and what finished looks like.
38. Build a day that actually fits
Use this when: Your to-do list ignores how much time you have.
Plan my day using [tasks, deadlines, fixed commitments, and available hours]. Prioritize what matters, leave room for breaks and delays, and tell me what won't fit instead of squeezing everything in.
39. Turn notes into actions
Use this when: A meeting or brain dump left you with scattered notes.
Turn [these notes] into decisions made, action items, open questions, and deadlines. Include owners only if the notes name them, and mark anything missing as unassigned or not specified.
40. Make a repeatable checklist
Use this when: You're doing the same task differently every time.
Turn [my process] into a checklist someone else could follow. Keep each item to one action, include the essential checks, and define when the task is complete.
Get past your first idea
41. Explore different directions
Use this when: Every suggestion feels like the same idea reworded.
Give me ten distinct ideas for [goal] for [audience], within [constraints]. Vary the underlying approach, not just the wording. Group the ideas by approach and explain the difference briefly.
42. Remove the obvious solution
Use this when: You want alternatives to the default way of doing things.
Solve [problem] without using [the obvious solution or resource]. Give me five workable alternatives within [constraints], with the main tradeoff of each.
43. Borrow from another field
Use this when: Your usual approach has stopped producing useful ideas.
Suggest three ideas from unrelated fields that could help with [my problem]. Explain the connection, how I'd adapt each idea, and one small way to test it in my situation.
44. Combine the strongest parts
Use this when: Several ideas are promising but none feels complete.
Here are my ideas: [list]. Identify the strongest part of each and combine compatible parts into one practical concept for [goal]. Explain what you left out and why.
Check before you trust it
45. Separate facts from assumptions
Use this when: A confident answer blends evidence with interpretation.
Review [this answer or draft]. Separate claims supported by the supplied material, assumptions, and opinions. Flag anything that needs outside verification, and don't label a claim verified unless you've checked its source.
46. Use only what I gave you
Use this when: You need a faithful answer from a document or excerpt.
Answer [my question] using only [the text I provide]. Point to the passage supporting each main point. If the material doesn't answer something, say so instead of filling the gap from memory.
47. Check the current facts
Use this when: Dates, features, rules, or availability could have changed.
Research [question] using current primary sources. Link the sources, note relevant dates, and separate confirmed facts from unresolved points. If you can't browse, say so and tell me what needs checking.
48. Show the numbers I can verify
Use this when: You need to catch a calculation or unit mistake.
Check this calculation: [numbers and method]. Show the formula, substituted values, units, and result. Use a calculator or calculation tool if available, and flag missing inputs or assumptions.
Improve the answer you already have
49. Find the weak spots, then revise
Use this when: The first draft is close but still needs work.
Evaluate [this draft] against [my criteria]. Identify the three most important weaknesses and revise it to address them, without adding unsupported facts. Briefly explain what changed.
50. Save the useful context
Use this when: A long conversation has produced decisions worth keeping.
Summarize this conversation into a reusable brief: goal, context, constraints, decisions, open questions, and next action. Separate confirmed decisions from suggestions so I can paste it into a new chat without losing the important parts.
Maverick's tip
Save the few you actually use. A short personal list is easier to reach for than a folder of prompts you never open.
Put a Few Together
For a decision, use 1 to fill in the context, 3 to explore options, and 16 to design a small test. For a draft, use 5 to set the standard, 21 to match your voice, and 24 to tighten it. For learning, use 6 to get the idea, 33 to see an example, and 30 to test yourself.
Run them as separate follow-ups. Read each answer and add what it missed before moving to the next prompt.
Published · Checked