You've told ChatGPT and Claude more than you realize, in every question, vent and half-finished idea. These prompts have it read between the lines and reflect you back to yourself: the things it figured out that you never said, your blind spot, two versions of where you're headed and one paragraph to read every morning. Some of it lands hard.
See Everything It Stored About You (and Delete It)
If you came from my Claude profile video, start here. It works in ChatGPT too. Open a new chat and paste this. It runs both prompts from the video for you, one after the other:
Rather go one at a time, like in the video? Paste this first:
Lay out every single thing you've stored about me as one complete list. Don't summarize anything.
Out comes a list most people have no idea exists. Now stay in the same chat and paste this:
Now tell me everything you figured out that I never told you. My age, my income, where I probably live, and what tipped you off.
Fair warning: this answer is probably going to freak you out.
Delete what you want gone
In Claude: click your name at the bottom left, then Settings, then Memory (it has its own page, right under Capabilities). Your memory is split into files under You, Topics, Areas and People. Click a file to view it, then click Delete. From then on, Claude only remembers what's left in there.
In ChatGPT: go to Settings > Personalization and manage your memory there. My ChatGPT Privacy Checklist walks through every switch.
Lock down the rest of Claude
- Stop training on your chats: go to Settings > Privacy and, under Preferences, turn off Help improve our AI models.
- Keep a chat off the record: start a new chat and click the ghost icon at the top. It won't show up in your history or use your memory.
- Keep a copy: Export memory, on the same Memory page, downloads your memory as a .zip.
- Stop it building a new file: once you're done with this page, turn off Generate memory from chats in Settings > Memory.
Maverick's tip
Got a summary instead of the full list? Say "Don't summarize anything. List every single item." Want to go deeper, into your blind spot and two versions of your future? Keep going below.
Turn On Its Memory
The plan · 4 steps
- 1Turn on memory so it can see your whole history.
- 2Run the four core prompts in order, in one chat.
- 3Save your morning paragraph where you'll see it.
- 4Pick a bonus prompt and go one level deeper.
Before you start
Your account
ChatGPT or Claude, any plan. The more you've used it, the better this gets.
Your materials
10 minutes and one fresh chat.
Your first goal
Your morning paragraph, saved somewhere you'll see it tomorrow.
This is the one thing that makes or breaks it. To reflect on every conversation you've ever had, the AI needs its memory switched on. Otherwise it can only see the chat that's open right now, and the answers will be thin.
In ChatGPT
Go to Settings > Personalization and make sure Memory (and reference chat history, if you see it) is turned on. That lets it draw on everything you've talked about, not just today.
In Claude
Go to Settings > Memory and make sure both toggles are on: Search and reference chats and Generate memory from chats. Memory is on by default on Free, Pro and Max, so you may just be checking. (On some older accounts these live under Capabilities.) Searching across past chats is a paid-plan feature. On Free, Claude still works from the memory it has built of you.
The more you've used it, the better this gets
This works best if you've had real conversations over time: asked for advice, thought out loud, worked through things. If your account is brand new, it won't have much to go on yet. Come back after you've used it for a while.
Maverick's tip
Want to see the raw material first? ChatGPT shows what it remembers in Settings > Personalization. Claude shows it in Settings > Memory, split into files under You, Topics, Areas and People. The ChatGPT memory guide walks through all of it.
The Reveal
This is the one that gets people. Open a fresh chat in ChatGPT or Claude, paste it in and brace yourself a little.
Based on every conversation we've ever had, tell me the five things you've figured out about me that I've never directly told you. Don't hold back to make me feel good, be honest.
What happens next: it reads between the lines of everything you've typed and names things you never said out loud. How you handle pressure, what you keep circling back to, what you avoid. The first one alone tends to stop people in their tracks.
The "don't hold back" part is doing the heavy lifting. Without it, both apps lean polite and hand you compliments.
Push it deeper
If the first answer lands, don't stop. Same chat:
For each of the five, what's your evidence? Quote the kinds of things I say that gave it away. Rank them by how confident you are, and tell me which one you think I'd argue with.
Maverick's tip
If the answers feel generic, push it: "go deeper, be more specific, use actual examples from our conversations." Want that honesty in every chat? Try my Truth Prompt.
Your Patterns and Your Blind Spot
Now go one level down, into how you actually operate. Stay in the same chat.
Tell me the patterns you see in how I think and make decisions, and the one thing I'm most in denial about.
What happens next: it maps how you tend to think, decide and get stuck, then names the blind spot you've been working hard not to look at. This is the one people screenshot and sit with for a while.
A pattern you can see is a pattern you can change. The point of naming your blind spot isn't to feel bad. You can't fix what you won't look at, so read it as information, not a judgment.
Two Versions of Your Future
This is the part that gets most people. You're going to see two futures side by side, and the gap between them is the whole point.
Predict my next five years if nothing changes. Then predict them again if I fix the one pattern that's holding me back the most. Show me both side by side.
What happens next: two versions of your future self, laid out next to each other. One if you keep drifting, one if you fix the thing. It's usually not a talent gap or a luck gap. It's one habit repeated for five years, and seeing that in black and white turns a vague "I should change" into a real fork in the road.
Maverick's tip
Want more detail? Add "Cover my career, relationships, health and mindset for both versions."
Make it concrete
Don't just read it and close the app. Ask the follow-up. That's how a spooky prediction becomes an actual plan.
Take the second version. What would I actually do differently in the first 30 days? Be specific enough that I could start tomorrow.
Your Morning Paragraph
End on this one. It takes everything it has learned about you and turns it into something you'll actually want to keep. Still in the same chat:
Write the one paragraph I should read every morning, based on everything you know about me.
What happens next: a short, personal paragraph, part reminder, part push, written for the exact person you are. Nobody else would get this paragraph.
What to do with it
Screenshot it. Make it your lock screen, pin it in your notes, tape it to the mirror, whatever you'll actually see. Then re-run this whole sequence in a month. If the paragraph comes back different, that's progress you can read.
Maverick's tip
If the paragraph feels too soft, say "rewrite it like a coach who knows me and isn't worried about my feelings." If it stings too much, say "same message, kinder voice." It's your paragraph, tune it.
How to Get Honest Answers (and Take Them Well)
Give it permission to be blunt. Both apps default to encouraging. Every prompt here works better with a permission line: "don't hold back," "don't soften it," "tell me what a good friend wouldn't say to my face."
One chat, one sitting. Run the core prompts back to back in the same conversation. Each answer becomes context for the next, and the morning paragraph at the end is only as good as everything above it.
Argue back. If one of the five things feels wrong, say so and ask for the evidence. Sometimes it updates, sometimes it holds its ground and shows you the receipts. Both are useful.
It's a mirror, not a verdict. It's spotting patterns in what you've actually typed, the way a sharp friend who's been listening for a year might. That's exactly why it can feel so accurate. When something lands as true, use it. When something feels off, let it go. You're the authority on you.
Heads up: If something hits a nerve, that's normal. These prompts are designed to be honest. But this is a chatbot reflecting your own words back, not a therapist or a diagnosis. If it surfaces something heavy, talk it through with someone you trust, and be as kind to yourself as you'd be to a friend reading the same thing.
Bonus Prompts: Go Deeper
All of these work best in the same chat, after the core four, in ChatGPT or Claude.
Quick reads
The Honest Friend
If a friend who loved me but refused to lie sat me down, what would they tell me I need to hear?
The Ignored Advice
What advice have I asked for more than once and then clearly not taken? Why do you think that is?
The Energy Map
Based on how I write, what topics give me energy and what topics drain me? What does that suggest about how I should spend my time?
The Wrong Question
What question do I keep asking you that you think is the wrong question? What should I be asking instead?
The Monthly Check-In
Compare me now to our conversations from a few months ago. What has changed, what has not, and what does that tell you?
Who you really are
The Complete Profile
Based on every conversation we've ever had, build me a complete profile of who I am. Include my communication style, my values, what I'm insecure about, and what I'm secretly working toward.
The Contradiction Scan
Tell me the contradictions you've noticed in me. Where do my words not match my actions? Be brutally honest.
The Mask Detector
What version of myself do I present that is different from who I actually am? Where am I performing instead of being honest?
The Fear Inventory
What are the 5 biggest fears driving my decisions right now? Not the fears I talk about openly, but the ones hiding behind my word choices and the questions I avoid asking.
The Values Audit
List my stated values based on what I say matters to me. Now list my revealed values based on how I actually spend my time and energy. Where is the gap biggest?
Career and ambition
The Comfort Zone Map
Map out my comfort zone based on our conversations. What is the one thing just outside my comfort zone that would create the most growth if I did it this month?
The Sabotage Scanner
Where am I self-sabotaging without realizing it? What do I consistently abandon right before it could succeed, and why?
The Hidden Talent Finder
Based on everything you know about me, what skills or talents do I consistently undervalue about myself? What would someone pay me for that I currently do not charge for or even recognize as valuable?
Relationships and communication
The Communication X-Ray
How do I come across when I am confident versus when I am uncertain? What impression do I probably make in the first 5 minutes of meeting someone new?
The Attachment Style Detector
Based on how I talk about relationships, what attachment style do I lean toward? What would someone close to me wish I understood about myself?
The Boundary Scanner
Where are my boundaries too rigid and where are they too loose? What is one boundary I need to set that I have been avoiding?
Mindset
The Inner Critic Transcript
If my inner critic had a voice, what would it sound like? Write a paragraph of the script that runs in my head when I am doubting myself. Then rewrite it from the perspective of a mentor who believes in me.
The Belief System Audit
What limiting beliefs am I operating under that I probably do not even recognize as beliefs? What would I attempt if I genuinely did not hold those beliefs?
The Emotional Triggers Map
What are my emotional triggers based on our conversations? What topics, situations, or types of feedback cause me to become defensive, shut down, or overreact? What is the deeper need behind each trigger?
Your future
The 10-Year Letter
Write a letter from the best possible version of me 10 years from now. What would they thank me for starting? What would they wish I had stopped doing sooner?
The Legacy Question
If the people closest to me described my impact on their lives right now, what would they honestly say? Not the polite version. What would I want them to say instead, and what needs to change for that to become true?
The Regret Predictor
Based on my current patterns, what are the 3 things I am most likely to regret in 10 years if I do not change them? Be specific about why I will regret each one.
The Seasons of Growth
What season of life am I in right now based on everything you know about me? Am I planting, growing, harvesting, or resting? What does this season require from me that I am not currently giving it?
The 30-Day Challenge
Want to go all in? Use one bonus prompt a day for 30 days, all in the same chat, so it builds a more complete picture of you every day. On day 29, paste this:
Based on everything we have discussed over the past month, give me a ranked list of the 5 changes that would have the biggest positive impact on my life. For each one, give me the exact first step I should take this week.
Want more reflective prompts like these? Try my Claude therapy prompts. They work in ChatGPT too. And what AI notices about you pushes the same idea one step further.
Common Mistakes and Quick Fixes
| If this happens | Try this |
|---|---|
| The answers are thin or generic | Turn memory on (Start Here), then say "go deeper, be more specific, use actual examples from our conversations." |
| It only talks about today's chat | ChatGPT: turn on reference chat history in Settings > Personalization. Claude: turn on Search and reference chats in Settings > Memory. |
| You can't find Claude's memory toggles | On some older accounts they live under Settings > Capabilities. |
| Your account is brand new | It needs history to work with. Use it for a few weeks, then come back. |
| It only gives you compliments | Add a permission line: "don't hold back" or "tell me what a good friend wouldn't say to my face." |
| One of the five feels wrong | Argue back and ask for the evidence (Prompt 4). |
| You opened a new chat halfway | Go back to the first chat. Each answer builds on the last one. |
| The two futures feel vague | Run the 30-day follow-up (Prompt 7), or ask it to cover career, relationships, health and mindset. |
| The morning paragraph is too soft or too harsh | "Rewrite it like a coach who knows me" or "same message, kinder voice." |
| Something hits a nerve | Take it lightly. It's a chatbot reflecting your words, not a diagnosis. Talk to someone you trust. |
Quick Reference: The Full Sequence
Screenshot this. Run it in order, one chat, one sitting.
| Step | What to paste |
|---|---|
| Setup | ChatGPT: Settings > Personalization, Memory on. Claude: Settings > Memory, both toggles on. |
| 1. The Reveal | Based on every conversation we've had, tell me the 5 things you've figured out about me that I never directly told you. Be honest, don't soften it. |
| 2. Patterns and blind spot | Tell me the patterns in how I think and make decisions, and the one thing I'm most in denial about. |
| 3. Two futures | Predict my next 5 years if nothing changes, then again if I fix my biggest pattern. Show both side by side. |
| 4. Morning paragraph | Write the one paragraph I should read every morning, based on everything you know about me. |
| Go deeper | Pick one bonus prompt, or start the 30-day challenge. |
DO THIS TODAY
Turn on memory, open a fresh chat and paste the Reveal. Run the core four in order, then save your morning paragraph as your lock screen. Keep what's true, drop what isn't, and re-run it in a month to see what changed.
Published · Checked