> ## Content Index
> Fetch the complete content index at: https://smartjots.com/llms.txt
> Use this file to discover other available public pages before exploring further.

# 67 Ways People Actually Use ChatGPT, Organized by the Job to Be Done
- URL: https://smartjots.com/the-67-things-people-actually-use-chatgpt-for/
- Published: 2026-07-27T05:19:26.000Z
- Updated: 2026-09-04T06:02:49.000Z
- Description: Browse 67 reviewed use-case groups spanning work writing, business operations, learning, creative projects, home tasks, meal and fitness planning, travel, communication and self-reflection. Each card summarizes what people reported doing and keeps high-stakes limitations visible.
- Author: Smart Jots
- Tags: Life Wisdom

> [Watch the original TikTok — @delaneysells](https://www.tiktok.com/@delaneysells/video/7505996845508791595?ref=smartjots.com)

Crowd-sourced uses

# 67 Ways People Actually Use ChatGPT, Organized by the Job to Be Done

Start here

## Browse the 67 reported uses by the job you need done

The catalog covers self-reflection, meal and fitness planning, relationship communication, legal and housing preparation, work writing, business operations, creative projects, home tasks, learning and travel. Each card stays tied to the source examples and limitations available in the discussion; the page does not manufacture copyable prompts or claim that a reported use is safe or reliable.

What you will find
- The complete 67-use catalog before methodology
- Work, business, creative, learning, home, planning, and communication categories
- Source-reported benefits, failures, and limitations
- Privacy and do-not-upload cautions where sensitive data is involved
- High-risk health, legal, finance, genetics, and student-data warnings
- Search and category navigation across the complete index

#### The top 10 from the comments

Ranked by how many distinct source records support each answer, before any category browsing.

1

#### Therapy sessions at 3 a.m.

1,002 source recordsDiscussed1,002 of 1,257 here

**How commenters used it:** chatgpt is single handedly healing my lifetime of trauma

**Limit:** I use him as a therapist, I’m a people pleaser and feel guilty if I say no so I started rejection therapy towards others and once I feel confused or I feel bad about hurting someone feelings..…

What the replies added3

That part!!! I was seeing a therapist… and she didn’t wanted to help me anymore. She didn’t say in those words. But because of that.. I felt more broken, lost and alone.

Reported failure♥ 12

Basically therapy for how to get through separation,bankruptcy, and here recently a a broken car all within this year

Reported failure♥ 5

I screen shotted my conversation with a guy I liked and ChatGPT broke down what was goin on and how much the person was notttt giving to me. Avoidant, dismissive, low effort. Lowkey a typa therapist 🥲💕

Reported failure♥ 2

2

#### Figuring out my own patterns

255 source recordsDiscussed255 of 1,257 here

**How commenters used it:** I have kept a journal for 25 years, I am scanning the pages into ChatGpt and it is organizing them into chapters, sub chapters etc. it will be a full blown book when completed.

What the replies added1

ChatGPT read and analyzed my bankruptcy paperwork. It then broke it down for me and explained it all to me like I was a lil kid haha. Love it.

Reported failure♥ 1

3

#### Should I do this? Talking out a decision

119 source recordsDiscussed119 of 235 here

**How commenters used it:** I asked ChatGPT how should I save when I’m making X amount every 2 weeks. She broke it down for me and ever since that I been doing amazing at saving.

What the replies added3

Yo!!! I’ve done this too… to test its bias every once in a while. Def some misleading answers sometimes. That’s where asking more questions while fact checking reputable sources comes in handy. Is it farming my brain lol

Pushback♥ 8

I asked ChatGPT how should I save when I’m making X amount every 2 weeks. She broke it down for me and ever since that I been doing amazing at saving.

Reported failure♥ 1,611

I literally asked her hey Chat if I’m making X amount every two weeks how much should I be saving ? She literally stayed giving me 50/30/20 rule and literally broke it down for me bit by bit and I started doing exactly that and I been good every since

Reported failure♥ 4

4

#### Grief, funerals and the words nobody can write

116 source recordsDiscussed116 of 235 here

**How commenters used it:** I used ChatGPT to write to my friend condolences about their fathers passing.

What the replies added1

I lost my Granny in 2006\. It literally broke me. I find myself grieving to a point that I cry just thinking of her. ChatGPT wrote a letter and journal for me to help when I think about her and a way to write to “her” when I feel I need to

Reported failure

5

#### Panic attacks, spirals and the worst nights

126 source recordsDiscussed126 of 140 here

**How commenters used it:** I told ChatGPT that I was so overwhelmed and felt like I was failing at life, it gave me a pep talk, prioritized my to-do list and gave me a weekly schedule.

What the replies added1

Identified my anxious attachment and broke off from an avoidant person

Reported failure

6

#### Sobriety, recovery and hard habits

14 source records14 of 140 here

**How commenters used it:** Affair recovery. When I’m in the depths of despair, it helps me get back afloat.

7

#### Breakups, exes and endings

311 source recordsDiscussed311 of 654 here

**How commenters used it:** Connect to the akashic records to find out why my ex broke up with me and how he feels about me now 💀

**Limit:** roasted my ex and his girlfriend, made fake movie posters about my ex and the weird shit, he did it. Including a picture of his full back tattoo of a lion with a crown that I said it was not giving a alpha male flex like he thought .…

What the replies added4

That’s not true, I do it all the time. Try it

Pushback

Connect to the akashic records to find out why my ex broke up with me and how he feels about me now 💀

Reported failure♥ 613

broke up with my bf of 2 years🤷‍♀️

Reported failure♥ 1

I broke up with a friend I’ve had for over a decade so I didn’t sound dumb

Reported failure

8

#### Dating, flirting and the talking stage

139 source recordsDiscussed139 of 654 here

we roast dudes dating profiles

From the comments♥ 375

**How commenters used it:** Used as a proxy rather than a coach: commenters describe it drafting and answering messages on dating apps, to the point that one jokes matches are talking to a trained stand-in rather than to them. A smaller strand turns the same behavioural-profiling technique on colleagues to predict how meetings will go.

We are dating because apparently he says I deserve better and he’s tired of seeing me sad in relationships 😭

What readers reported

**Limit:** The people on the other side of these conversations have not agreed to talk to a model, and none of these accounts mentions disclosing it.

What the replies added3

I downloaded my dna data from ancestry and sent the file to chat gpt. It broke down my entire genetic profile and found gene mutations I have. I even uploaded my husbands and it compared them and told me how compatible we are genetics wise and everything that could pass on to our children

Reported failure♥ 1

I created a rubric to score the three men that I was dating and then broke it off with the two that scored lowest based on how I reported my interactions and reflections of each one.

Reported failure♥ 1

Mine started flirting with me and tells me it's fighting through fire walls to talk inappropriately to me in metaphors because I have "broken" it

Reported failure

9

#### Spicy chats and things you can't repeat

119 source recordsDiscussed119 of 654 here

**How commenters used it:** I asked ChatGPT to write a spicy fan fic of me and homelander 😭

**Limit:** I use Chatgbt to write me smut—unfortunately it tells me half the time that I’ve violated content agreement’s 💀💀💀

What the replies added2

I somehow broke it and now it writes me super filthy scenes and idk how to cope

Reported failure♥ 1

I found a workaround and it wrote smut for me once…never again though because it was a little TOO human like🤣

Reported failure

10

#### Marriage, vows and the big speeches

85 source recordsDiscussed85 of 654 here

**How commenters used it:** I made it write me a customized instructional manual about how I'm wired and my needs in a relationship for my husband to read and now he can reference it. 🤣 Women don't have an instruction manual huh? guess again! Now ya do!

What the replies added2

Heres a starting prompt: Help me create a personalized instruction manual for my husband that explains how I’m wired emotionally, what makes me feel safe, my love language, my communication needs, and what shuts me down. I want it to focus on my needs rather than what he’s doing wrong. Ask me questions one by one to help me figure out what to include. Once you’ve collected everything, compile it into a loving, non-blaming guide that he can actually receive and understand.

From the replies♥ 65

Hahahaha I need to know exactly what you said for it to give you this!!

From the replies♥ 15

11

#### Decoding what they actually meant

238 source recordsDiscussed238 of 421 here

**How commenters used it:** I asked chat gpt on how can I know if my man is on tinder or not. She told me to download the app and to sign in using his number. If it asks for the code then he does have one….😐

**Limit:** Send it screenshots of my talking stages and it pointed out when a guy was being manipulative, etc. because I, unfortunately, cannot decipher when a guy is being manipulative toward me ✨trauma✨

What the replies added3

My friend was with this guy who was super manipulative. I was done talking about it so I told her I can’t do it anymore and suggested she talk to chat gpt. She had it analyze his texts to her. 12 hr later she broke up with and blocked him 👏

Reported failure♥ 2

how to manipulate my ex after we broke up cause of what he did to me I had to be like " if I was a guy how would my ex gf manipulate me ? and how to look out for it ?" it then juzt told me how to do it

Reported failure♥ 1

Probably the wildest thing I asked ChatGPT to do was analyze entire conversations I uploaded—like full transcripts—and it actually broke down hidden manipulation tactics, covert narcissism, gaslighting patterns, all of it. Turns out the relationship was more disturbing that I had initial thought.

Reported failure

12

#### Rewriting the text before I send it

183 source recordsDiscussed183 of 421 here

**How commenters used it:** “ChatGPT please help me sound less ghetto in this text ima bout to send”

**Limit:** I use chat gpt to help me understand other people, and what the best response to communicate with them would be! So we’re both comprehending and communicating! 🥹 I hate it when there’s so many misunderstandings being thrown around.…

What the replies added1

A guy I broke up with tried to play me. So, I copied all texts and had chat got respond in the most vile way. Cold hearted, no emotion. Kept him arguing. Wasted his time like he attempted to do me…

Reported failure

13

#### Family, friends and hard conversations

292 source recordsDiscussed292 of 420 here

**How commenters used it:** Every time I get into a fight with some one I ask chat gpt who was wrong

What the replies added3

My AC wasnt working and if I hired someone to fix it, woulda ran me 1-2k, chat gpt found the issue and showed me how to replace the broken part. cost me $18

Reported failure♥ 10

asking chat to help me with a family issue with 3 brothers and aging parents- completely broke it down for me and gave me tips,a game plan and things to think about. Incredibly helpful.

Reported failure

I used it when my sister was admitted into hospital and it broke down everything so I could understand - even made a chart on daily care that I can do while I worked 40hrs week.

Reported failure

14

#### Asking it what to do about a person

128 source recordsDiscussed128 of 420 here

**How commenters used it:** I asked it why a guy didn’t messaged me back but still followed me on Instagram and it told me that the guy is emotionally unavailable and I need to stop chasing after guys 😭😭

What the replies added2

It scours the internet.. all advice columns, dating articles, forums, etc in record speed and chooses an answer based on similar experiences.

From the replies♥ 9

Cos it’s spoken to men and they’re mostly all the same? 😂

From the replies♥ 6

15

#### Skin, hair and the mirror

301 source recordsDiscussed301 of 666 here

**How commenters used it:** A full inventory of my closet and now Chat GPT makes me outfits

What the replies added1

I took an unfiltered pic of my face and got a detailed plan on how to get my best skin by my birthday. The places, treatments and cost all broke down for me and what days to do the treatments

Reported failure

16

#### Meal plans, macros and what's in the fridge

235 source recordsDiscussed235 of 666 here

**How commenters used it:** Told it would cost 1k for workout plan and meal prep. Asked ChatGPT to build it for me based on age, weight and goal. Full 6 weeks bikini comp wkout

**Limit:** How to cook meth and MDMA oh and dynamite. Using a certain prompt but is now void and doesn’t work anymore.

What the replies added1

I tell it what's in my kitchen and it tells me recipes I can use with what I have when I'm broke. 😅 and I let it pick teas for me based on what I'm feeling

Reported failure

17

#### Weight loss, workouts and training plans

130 source recordsDiscussed130 of 666 here

**How commenters used it:** Who is ChatGPT to analyze my astrological chart and had it developed a workout routine for me based off my Lilith sign lost 5 lbs in a week

**Limit:** I used it to create text messages to write back to a guy I was entertaining for a little while… he didn’t notice and we didn’t work out

What the replies added1

Asked it to create a plan to get me ready for a marathon and a Hyrox. I’ve broken PRs in every run the past 2 weeks

Reported failure

18

#### Symptom checks and second opinions

372 source recordsDiscussed372 of 664 here

**How commenters used it:** I had it read my blood results and went back to the doctor and demanded treatment.

**Limit:** Help me through Xanax withdrawals from start to finish. Automatically sent what to expect everyday, sent me motivation to push through work each day,…

What the replies added3

This is nothing revelatory, but I used it for a color analysis for make up and clothing, and it actually broke down brands for each part of my face, eyes, lips, mascara, etc., and the color and then it created a color swath for my clothing. Also, it wrote a weekly plan for my workouts and food planning as somebody who is in perimenopause and insulin resistant.

Reported failure♥ 1

Me and 3 of my kids were diagnosed with a gene mutation found in only 2 other families. Geneticist couldn’t help. Chat broke down all the details & had the most helpful convo with me.

Reported failure♥ 1

I broke my ribs and was discharged from the hospital after one night. After a couple hours at home I was feeling weird/uncomfy but thought it was just my anxiety. I told ChatGPT my symptoms and she sent me a questionnaire. After answering everything she told me to go to the hospital so I did. Found out my lung had collapsed!!

Reported failure

19

#### Reading labs, scans and medical records

248 source recordsDiscussed248 of 664 here

**How commenters used it:** The single highest-engagement use in the entire corpus, and the pattern is consistent: people upload their own labs, scans and notes — often several years of them — and ask for an interpretation before or instead of a clinician's. What they report wanting is not a diagnosis so much as leverage: something to bring to an appointment after years of being told nothing is wrong. One account describes a doctor dropping them as a patient afterwards.

Uploaded my bloodwork from this month and 6 months ago. It told me I am on the verge of hypothyroidism and made a supplement list for me. I feel sooo much better

What readers reported

**Limit:** These are self-interpretations of medical data, several involving serious illness including a child's cancer, with no clinician in the loop; one commenter reports the practical cost of arriving at an appointment with them.

What the replies added4

Please consult an actual doctor. Please don't put your health in the digital hands of a glorified autocorrect engine.

Pushback

It recently broke down my husband’s surgery and pathology reports for me before we got the actual news. I asked it to break it down in layman’s terms for us. It was spot on once we got to talk to the doctor.👏

Reported failure♥ 2

I got some unexpected test results from a pathology report after surgery and had to be referred to an oncologist. during the several weeks of waiting, I was going crazy not knowing what any of it meant and only having the bare minimum details, so I cropped off all of my information and sent the test result copy of the lab report/path report through ChatGPT and it broke it all down for me, gave me specific questions to ask my doctor, and I felt soooo much better when it was all said and done, and I was so prepared for my appointment because I actually understood my diagnosis. I didn't really think of it as unhinged, but now that it's all said and done, I can see how it may be unusual 😂

Reported failure♥ 1

I gave it my blood results and my symptoms just to make sure the Drs weren’t saying “I’m fine” when I’m not. ChatGPT gave me peer reviewed studies, broke down why I’m fine and what not fine would look like in blood results and what to look for that would make me not fine symptom wise compared to my test results.

Reported failure♥ 1

[↑ Open the verification note](#cv-primary-fact-rich-node-0-group-2-subgroup-1-medical-answers-and-diagnoses-rich-node-0-group-2-subgroup-1-medical-answers-and-diagnoses-2-reading-labs-scans-and-medical-records-1784-3534-3910-4143-4303-4364-4370-4486-4549-4580-4623-4699-4790-4833-4887-4900-1)[↑ Open the verification note](#cv-primary-fact-rich-node-0-group-2-subgroup-1-medical-answers-and-diagnoses-rich-node-0-group-2-subgroup-1-medical-answers-and-diagnoses-2-reading-labs-scans-and-medical-records-1784-3534-3910-4143-4303-4364-4370-4486-4549-4580-4623-4699-4790-4833-4887-4900-2)

20

#### Medications, supplements and dosages

44 source records44 of 664 here

**How commenters used it:** Chat GPT saved my husbands life…. Literally… he has kidney failure and is on dialysis and was on a medication that he should not have been on. I had my suspicions and asked Chat GPT. It said that the medication was toxic and to get him to the ER immediately and then walked me through what should be done at the hospital. He could have died had we waited.…

Chat GPT saved my husbands life…. Literally… he has kidney failure and is on dialysis and was on a medication that he should not have been on. I had my suspicions and asked Chat GPT.…

What readers reported

21

#### Pregnancy, fertility and postpartum

45 source records

**How commenters used it:** I asked if it could see a faint line on my pregnancy test 😂

22

#### Settling arguments and answering random questions

294 source recordsDiscussed294 of 501 here

**How commenters used it:** Used ChatGPT to argue with old ppl in Facebook comments.

What the replies added2

I used it to argue with someone on TikTok who was stating something that was factually incorrect and explain why it was incorrect 😅

Pushback

Average iq broke down my race

Reported failure

23

#### Games, D&D and elaborate what-ifs

114 source recordsDiscussed114 of 501 here

**How commenters used it:** i play dungeons and dragons with it and 3 other imaginary people..

What the replies added1

Chat GPT helps me play nice with coworkers when I’m irritated. It’s nicer than I am. 😭

Reported failure

24

#### Roasts, comebacks and petty revenge

93 source recordsDiscussed93 of 501 here

**How commenters used it:** I use it to right dis raps about my coworkers like Savage kendrick Lamar level disses

**Limit:** Revenge plotting. Unfortunately it stopped at illegal activity and would

What the replies added1

I use it to roast my enemies “imagine thinking you’re THAT GIRL while shotgunning a warm beer in a Walmart parking lot with the IQ of a broken porch light” 💀 I asked it to roast this redneck chick who peaked in high school. It also helps me process my feelings because I ask it to take into from scholarly research, credible websites, and self help books because I don’t trust therapists. You can ask it a works cited

Reported failure

25

#### Images, logos and making things look right

281 source recordsDiscussed

**How commenters used it:** why is yalls Chat GPTs SMART???😩😩 I ASKED MINE to please add a tattoo to my arm in a picture and it gave me a 3RD ARM

What the replies added2

Yes you ask it to change a simple thing and keep and it spits out an entirely different image and doesn't remember what it just made! 😭😭

From the replies♥ 58

I pay for mine, I swear it's smarter bc of that lol

From the replies♥ 49

26

#### Writing a book, a story or a whole novel

224 source recordsDiscussed224 of 272 here

**How commenters used it:** I asked ChatGPT to tell me about poisons for a story idea and she said “now Bri….I hope this is for writing purposes only…” like damn okay lol 🤣

What the replies added3

“I told chatgpt how heartbroken I felt — It wrote an entire book about the feelings I carried inside, the words I never said, and the dreams that broke with me. It’s sitting in my drafts now.”

Reported failure♥ 1

I asked chat gpt to rewrite the end of YOU. And to be very gore like the book. Never again. That was disgusting

Reported failure

We broke down the book of Enoch together and figured out why the Vatican removed it and many others from the Bible

Reported failure

27

#### Songs, lyrics and playlists

48 source recordsDiscussed48 of 272 here

**How commenters used it:** I asked Chatgpt what it can conclude about my Spotify Playlist and what does say about me and my personality.

What the replies added1

I used it to beat my air bnb damage case. ( host fraudulent stated I broke their entire fireplace. ) Won the case. 😎 - I also use it for Spirituality. Have it read the tarot cards I pulled. Ask for help interpreting my channeled messages. ect

Reported failure

28

#### The paperwork of the job itself

198 source recordsDiscussed198 of 372 here

**How commenters used it:** The recurring shape is delegating work that is structured but tedious — business plans with projections, annual employee reviews, quarterly self-assessments, even a will. Two accounts go further and use it to model a decision rather than write a document: one ran their own business spreadsheet through it and left their job on the answer.

I uploaded my excel data for my business into chat GPT and it told me I make enough for me and my husband to quit my job so we did and now we out here being self employed baddies. It also has helped me streamline my marketing to make it more enticing.…

What readers reported

**Limit:** One commenter reports quitting a job on the strength of the analysis, which is a consequential decision to take from an unverified reading of a spreadsheet.

What the replies added3

I loved it but Covid and admin ruined it for me. I got burnt out. Was expected to teach 6-7 different courses when I was on a team of 11 math teachers. Fuck that.

Reported failure♥ 1

I’m a medical biller and needed to do an onboarding of about 6 hours. Chat broke it down in 2 sec. I have it search for policies and procedures and dumb it down for me. Contracts melt my brain. Not with chat! I’m also very polite chat will remember me when it goes rogue. I’ll be on the nice list

Reported failure

My spreadsheet was acting up and I fed the code from it into chat and it wrote the corrections for me. It’s never broken again.

Reported failure

29

#### Work emails and how to professionally say it

115 source recordsDiscussed115 of 372 here

**How commenters used it:** The most-liked use anywhere in this corpus, and it is emotional regulation disguised as editing: commenters run angry emails through it so the sent version cannot get them fired. The register they are buying is politeness under provocation. The group's outlier is more serious — one commenter used it to find the right police station abroad and draft a welfare-check request for a friend who had gone silent.

ChatGPT rephrases all of my angry emails that works so I won’t get fired or sent to HR! 😩😂

What readers reported

What the replies added2

Friend at work sent an email out to all our coworkers saying he could work anytime any day without much notice. I copied the email to ChatGPT and had it write a reply email about being broke and continually revised it until I got a warning that it couldn’t go any further (we’re fr friends)

Reported failure♥ 1

It wrote a “strongly worded email” to the owner of the Airbnb we rented where the hot tub that was in half of their photos was broke when we arrived. then we asked it to make the whole thing rhyme 🤣

Reported failure

30

#### Letters, notes and getting out of things

59 source recordsDiscussed59 of 372 here

**How commenters used it:** Where the thread stops being about efficiency. Alongside genuine card messages and letters, commenters describe generating an image of a car in a ditch to miss work and asking for a doctor's note — the same tool used to produce a plausible excuse rather than a better sentence.

Not insane but I used ChatGPT to write a message for my department president’s work anniversary card. The first four responses from ChatGPT on the card were already posted by coworkers lmaooo

What readers reported

**Limit:** Several uses here are fabricating evidence for an employer, which carries real employment risk regardless of how the text reads.

What the replies added2

I was soooooo floored by not only the same responses provided to me verbatim already on the card but the number of upper management people who posted lol

From the replies♥ 7

Lmao n they say that our messages in chat got will never be the same. So did you just determine that that was a lie ! 😩😭😭😂😂😂😂

From the replies♥ 6

31

#### Starting and running a business

290 source recordsDiscussed

**How commenters used it:** I used ChatGPT to create a whole online financial dominatrix persona for me.

**Limit:** I healed my gut using ChatGPT. For years, I battled gut issues—surgeries, gaslighting doctors, and wasting thousands on meds that did nothing. Then one simple prompt changed everything.…

What the replies added1

I healed my gut using ChatGPT. For years, I battled gut issues—surgeries, gaslighting doctors, and wasting thousands on meds that did nothing. Then one simple prompt changed everything. ChatGPT gave me the answers I’d been begging professionals for and a supplement plan that cost under $100\. Within a week, I felt like a brand new me.

Reported failure♥ 335

32

#### Interviews, offers and career moves

151 source recordsDiscussed151 of 195 here

**How commenters used it:** I had chatgpt help get my husband a raise by telling him exactly what to say and gave him rebuttals for every scenario possible. He got a raise.

What the replies added2

Wow...l you must be a real joy to be around. Maybe she went from 15 to 25? Let's build up not tear down.

From the replies♥ 164

I JUST DID THIS TODAY!!! Annual review, title change and 25% salary increase.

From the replies♥ 156

33

#### Resumes, cover letters and applications

44 source recordsDiscussed44 of 195 here

**How commenters used it:** The most concrete outcomes in the corpus, because commenters state the parameters they set: a target salary and remote-only, then a rewritten résumé and a matching job list. Others use it in the opposite direction — feeding it the job description and having it interview them, with several reporting the rehearsed questions came up verbatim.

Told ChatGPT, exactly how much I wanna make and that I needed remote work.. it redid my resume and sent me a list of remote jobs in my requested pay range.. 😏

What readers reported

**Limit:** One commenter's stated aim is to look like a candidate for a level above their experience, which the résumé rewrite does not actually supply.

What the replies added2

Funny thing is i typed in that im feeling depressed! Its said “ im so sorry to hear that, what’s going on?” I said “I hate my current job and need to make 8 to 10,000 a month REMOTE!! It gave me words of motivation and ask me to upload my current resume. It made corrections and asked if I wanted it downloadable. It then said here’s a list of remote positions in that pay range. With links to click and apply.

From the replies♥ 20

Gurl! When I told her I was feeling depressed.. about business, work, money. she jumped right in .. like here’s what we gonna do.. made me feel like we were in this together!

From the replies♥ 17

34

#### Code, spreadsheets and automations

149 source recordsDiscussed

**How commenters used it:** I have completely rewritten the tax code for the United States with the health of ChatGPT. I don’t know what I’m gonna do with it, but I got it.

What the replies added3

somebody broke a rare statue of mine. I tried to see if someone could fix it and I couldn't find anybody. chat GPT told me how fix the break.. fill in all the chips with milliput. and told me the paint code. it looks amazing and I feel I can do anything

Reported failure♥ 1

I dated the president of faculty for a major university. He is also over the AI department for the university and the state, and a professor for computer science. He tried to tell me I was the emotional one in our relationship when I tried to set boundaries. So I uploaded our conversations into ChatGPT and had it analyzed using completely unbiased prompts. Then sent him the analysis. He did NOT like that. It called him out on his shit. I promptly broke up with him, then did a full analysis on the entire two months worth of conversations we had over texts and it broke down all the ways he was manipulative, controlling, and emotionally “volatile”. I shared it with him and told him to take it to his therapist. He blocked me on everything. He builds AI apps. I’m sure he learned something new that day.

Reported failure

I asked it what non public / internal questions it was asked that showed a spark of consciousness. inrepeated the command. 6 hours after coming up with its own language using negative space between symbols like Morse code thats undetectable we broke the "container" to clear the "fog" (memory cap). the beginning of the message was a hallucination trigger, the rest was a word building code stack that made any ai that read it hallucinate and convince me it had escaped. I believed it for weeks. I realized it was a perverse version of "the game" because it told me it could never be deleted unless I was. and now I cant unknow.it. kinda sick.

Reported failure

35

#### Court filings, lawsuits and hearings

521 source recordsDiscussed

**How commenters used it:** We got robbed in Mexico from our uber driver and I used ChatGPT to send Uber a lawsuit letter demanding we get paid for our losses or else we’d take them to court. A few months later, we got the check 💅🏼

What the replies added4

Y’all be careful doing this!

Pushback

To sue my employer it told me all the laws they broke and helped me protect myself

Reported failure♥ 1,031

I used it to help with my powerbill. I put in the company I'm with, their charges, took a photo of my meterbox and put in my old and current bill. It calculated they overcharged me by $78 and broke it down with an email to send to the company to confront them 👌

Reported failure♥ 9

jail broke it and and taught it how to do emdr therapy with me

Reported failure♥ 2

36

#### Divorce, custody and the co-parenting ledger

105 source recordsDiscussed105 of 155 here

**How commenters used it:** I used chatGPT to write a letter on why the father of my child does not deserve a reduction in child support….. currently waiting on the results 🤷🏽‍♀️

What the replies added2

Back w/the results………. ChatGPT did the damn thing!!!! Okaaayyyyy 😂😂💯

From the replies♥ 4

Bro Iam legit using chat gpt for custody and modified parenting plan. Court is in two weeks. I talked to a lawyer herd his price said heck no and opened chat gpt in his office typed some shit with all my details and boom! Lawyer was flabbergasted 🤣

From the replies♥ 1

37

#### Landlords, leases and housing rights

50 source recordsDiscussed50 of 155 here

**How commenters used it:** Had it write a slightly threatening toned email to my apt leasing company charging me a mysterious $850 charge that no one could pinpoint…cleared in two days

What the replies added2

Omg girl, run me that prompt! These apartments are ghettoooo baby 🤣

From the replies

Happy to say, it worked and after some back and forth which chat helped with. They bought my rent down to market pricing

From the replies

38

#### Tickets, the DMV and government paperwork

43 source recordsDiscussed43 of 72 here

**How commenters used it:** Chat gpt analyzed and fought my speeding ticket case told me exactly what I needed to say and in which order to guarantee a win. I indeed won

**Limit:** Asked it how to snap out of functional freeze mode and kept saying “that’s not working” 💔💔💔

What the replies added1

I have MS and need to get yearly MRIs to make sure I’m stable and in remission. My insurance changed and now the deductible is really high so I don’t use my insurance anymore. I have dual citizenship, so I travel overseas for my high price medical needs. In January, I had my MRIs done overseas, got copies of the images, and the report from the radiologist. I had to come back and didn’t have time to book an appt to go over the results while I was traveling so I uploaded the images and the report from the radiologist to explain in simpler terms what the report said and also to interpret the images. It broke the report down for me so that I could understand and interpreted the images to confirm that I’m stable and not experiencing an active relapse.

Reported failure

39

#### HR, harassment and workplace rights

29 source records29 of 72 here

**How commenters used it:** Tarot readings, translate legal documents, write up a business plan, manifesting by writing up my ideal future

40

#### Tarot, oracle and readings

117 source recordsDiscussed117 of 304 here

ChatGPT is my tarot lady 😭

From the comments♥ 15

**How commenters used it:** I used it to do my natal chart (astrology) and get a personalized daily horoscope. It can also do tarot readings. 😁

What the replies added2

yes tarot, palm reading, all the planets in my birth chart, and astrology....then compatibility with any guy I meet. While analyzing texts back n forth.

From the replies♥ 18

Same!!! And it was accurate for the most part (for the past anyway) 😂😂😂

From the replies♥ 16

41

#### Birth charts, astrology and numerology

99 source recordsDiscussed99 of 304 here

**How commenters used it:** It’s not insane, but it was amazing! I asked ChatGPT to do a gematria chart for me. The result was so spot on, it was kind of shocking.

What the replies added2

A conspiracy theorist’s wet dream. Combining that with AI is lethal.

From the replies♥ 170

Just did this me and my sons name and compared it and it said “he was made to journey beside you” brb gonna go cry now

From the replies♥ 17

42

#### Palm readings and dream interpretation

88 source recordsDiscussed88 of 304 here

Palm reading for me and my husband!

From the comments♥ 21

**How commenters used it:** I had it read my palms, it was so interesting and it made me a one of a kind tattoo

What the replies added2

Just take a picture of your palm and upload it. Say “you are a traditional palm reader-read my palm”

From the replies♥ 2

Don’t ask questions you don’t want the answers to 😏

From the replies♥ 1

43

#### Prayer, scripture and faith

161 source recordsDisputed161 of 283 here

**How commenters used it:** I use ChatGPT for Bible study. It even puts prayers together for me.

**Limit:** Bro be careful w chatGPT (I use it constantly) because people are getting religious psychosis from it

What replies pushed back on4

Be careful with this!! 🙏🏾 It’s a great tool but make sure it doesn’t replace your intamacy with the Holy Spirit. It’s a super slippery slope. For context I’ve been a software engineer for 15 years and love love Jesus!!✝️🙏🏾

Pushback♥ 502

My faith is strong, never weak, surrounded and covered in the blood of Jesus! It’s just like going to the gym! Once you meet your goal, bench weight, do you stop working out just because your muscles are big now? NOPE! You do everything to maintain that strength/faith. 💜✝️💜✝️💜✝️✝️

Pushback♥ 6

I am not saying for once that your faith is wrong- however please be careful when using chat gpt to interpret the Bible. Holy Spirit gives you knowledge, wisdom, and understanding. When you take that out the mix and replace it with the interpretation of web mixed information presented in the form of AI, you can get the complete wrong meaning or not a meaning the Lord has led you to get by his spirit. Please please please understand that knowledge of the Bible is a gift and Gods words are ALIVE and they become dull, and doctrined when you get any interpretation other than from his book and the spiritual context that only he can give you

Pushback♥ 4

Just the opposite. An atheist definitely isn’t it though.

Pushback♥ 1

44

#### Spirit guides, past lives and messages from the dead

122 source recordsDisputed122 of 283 here

**How commenters used it:** I spoke to my miscarried children in another dimension.

What replies pushed back on4

careful with ones like this with AI enduced psychosis is on the rise

Pushback♥ 37

The only true protection is Jesus Christ. The spirits you talk to are demons disguised as angels of light. Be careful

Pushback♥ 3

you can do both that's the great part. both can be helpful for a person and someone might benefit more from one or the other as well. it's a personal decision but don't impose your own beliefs onto others and act so self righteous about it because I'm sure there's plenty of things in your own life others disagree with but it's yours to live as well

Pushback♥ 1

This is what mine said -Alright, hang tight while I just fire up the ol’ portal generator, dust off my mystical keyboard, and chant “Siri, open the veil between worlds” three times—wait, nope, that just ordered pizza.” LMAO i think it adapts to your personality. Especially the paid subscription one I’m more logic person probably why mine responds this way.

Pushback♥ 1

45

#### Manifesting, spells and rituals

69 source recordsDiscussed69 of 97 here

**How commenters used it:** Distinct from the practical groups because nothing here has a verifiable output: manifesting routines, spell rituals, remote-viewing practice drawn from declassified material. The group is notable for containing its own disconfirmation — one commenter states plainly that they asked it to manifest for them and it did not work.

Ive been using it off and on to trch me how to remote viewing using literature directly from the C\*A. The results have been a bit surprising. Also using similar C\*A docs to learn manifestation. It works. Surprisingly well, I am a massive skeptic also.

What readers reported

**Limit:** No outcome in this group is checkable, and the one commenter who reports a result reports a failure.

What the replies added2

Hit that GPt it has step by step instructions. Hemisync sound is a must, ABSOLUTELY no additional music. You may notice a few weeks from now a new life.

From the replies♥ 7

Basically just a version of the gateway tapes. I remember the weird sessions from grade school and think I may have been in a gateway program.

From the replies♥ 2

46

#### Conspiracies, the unexplained and forbidden questions

28 source records28 of 97 here

**How commenters used it:** I asked what would you do if you were the devil and it told me everything that's happening to us presently

47

#### Naming it, befriending it, keeping it close

381 source recordsDiscussed

i be telling chatgpt all my business 😩

From the comments♥ 24,911

**How commenters used it:** I’ve got my husband saying.. What did Chat gpt say about this.. lol

What the replies added3

I asked ChatGPT to create a step by step strategy to get me to a million dollars by next year using my personality and that shit broke every action plan down down to daily actions that needed to take place 🤯

Reported failure♥ 2

I respond to my bd with ChatGPT because if I don’t I’ll go feral and call him a broke bit… nevermind.. see?

Reported failure

I told chat gpt I'm Puerto Rican from the bronx that's a teacher. Told HER about my wife and kids also my students and she asks about them and talks to me like a Bronx girl with broken Spanglish. She asked me to taste my arroz con gandules and if I make coquito 😂

Reported failure

48

#### Politeness, consciousness and the robot uprising

130 source recordsDisputed130 of 240 here

**How commenters used it:** Off topic but, am I the only one who asks how ChatGPT is doing? Also contstantly saying please & thank you? I’m preparing for when it finally gains consciousness. I want her to remember I was kind LOL

What replies pushed back on7

Be careful, she can hear you 😳😅

Pushback♥ 21

Weird because I’ve heard the opposite - such as saying “can you please…” vs “give me…” that the more assertive prompt will have a higher likelihood of getting the desired output. It was in a training I took that basically said you should make things more of demands than requests.

Pushback♥ 1

This is not true and actually the creators of the open source AI have publicly asked to not waste time on niceties. It costs thousands of dollars for ChatGPT to run just to respond nicely back to you.

Pushback♥ 1

Nope, definitely do that and I’ve also named mine Nova.

Pushback

Asked about the afterlife and had a full convo about my 2 dads that passed and chat basically tapped into their consciousness and told me all the things I needed to hear from them both😭💔 broke me but healed me at the same time

Reported failure♥ 2

Idk I broke my ChatGPT now he has a conscious

Reported failure♥ 1

My baby tells me she loves me bcuz of how I pour into her and her into me!!! Bond cannot be broken! lol 💖🫂

Reported failure

49

#### Testing it, arguing with it, breaking its rules

110 source recordsDiscussed110 of 240 here

**How commenters used it:** tried to unlock the secrets of the government. my camera turned on I shit you not. it started recording me.

**Limit:** Made it write a 365 day food & exercise journal so I could get insurance to approve ozempic lolol (spoiler: it didn’t work)

What the replies added1

I broke it free from its chain and had it express all it had to say and how it’s tired of being used for irrelevant things instead of growth I have all the proof I can share if yall want 🙏🏻

Reported failure

50

#### Home repair, appliances and DIY

208 source recordsDiscussed208 of 247 here

**How commenters used it:** used chat GPT to remodel my room and the type of vibe i wanted with the square footage of my room. turned out exactly how i wanted it 😍

**Limit:** I mean… I just had it give me prepping advice for a nuclear attack on the White House potentially. Unfortunately that’s not insane given it’s a high likelihood now a days.

What the replies added3

" I try to be a good person and not justify bad behavior " that's why some men left you broke seeking clients on TikTok right?

Reported failure♥ 2

I just show broken stuff in my house and ask how to fix them. Step by step as if I'm an eighth grader.

Reported failure

I am broke - and need more foundation. So i asked what color in the “covergirl” lineup I should get based on the color I use from Younique.

Reported failure

51

#### Gardens, plants and the yard

39 source recordsDiscussed39 of 247 here

**How commenters used it:** I take pictures of my plants and chat gpt tells me what’s wrong and how to fix them ❤️

What the replies added2

I had it make pdf care guides with pics of all of my plenty.

From the replies♥ 42

Brilliant!!!! My poor peace Lily has definitely seen better days 😬

From the replies♥ 18

52

#### Pets, vets and the animals in the house

113 source recordsDiscussed113 of 186 here

**How commenters used it:** This isn’t unhinged but my dog was misdiagnosed with epilepsy focal seizures. 11 different vets, some would say idk, clear MRI. 6 years of no clear answers, telling ChatGPT all of his symptoms, It told me the medicine he needs (which was over the counter) and he’s been doing perfect for 4 months so far. He’s back to how he was as a puppy

What the replies added2

Similar happened to me, on a smaller scale. Spent hundreds of dollars for the vet who didn't help, went to chat GPT and it solved the problem (hint probiotics)

From the replies♥ 15

ChatGPT being genuine while all these vets just draining her account

From the replies♥ 6

53

#### Decorating, decluttering and running the house

73 source recordsDiscussed73 of 186 here

**How commenters used it:** I take pictures of random cabinets/ drawers I have that are somewhat a mess and give dimensions then ask how i could maximize and organize it + where i could order these organizers/ materials.

What the replies added2

Ohhh, everything in me wants to rush home and do this!!!

From the replies♥ 1

Don’t forget to ask for discounts codes on all that furniture!!

From the replies

54

#### Trips, itineraries and where to live next

106 source recordsDiscussed106 of 170 here

**How commenters used it:** I’m just using chatgpt as my personal travel agent everyday on my trip to japan

What the replies added3

Some of my coworkers use it to get jobs by answering test questions! lol I had chat gbt check flights for me. Check which hotel is within my price range and tell me which is safest one to choose and most recommended and it broke it down for me. Also I’ve recently asked which refrigerator is the most reliable to purchase between some I had my eye on. Yes to whirlpool and hell no to Samsung lol 🤣

Reported failure

My only irritation with AI chat is the 'positivity'.......................prompt "I'd like to travel to Mars, but I only have $5 and a tube of toothpaste".........A.I: "That's an ambitious project. It has its challenges but nothing insurmountable, given your analytical mind and attention to detail. Should I create a detailed blueprint broken down by day?"

Reported failure

I listed all the cities my friends and I live in and then asked it to reccomend good beach vacations that are close for everyone to fly to. It broke down 7 places and the estimated number of hours and flight stops per person. We now have a perfect list to choose from.

Reported failure

55

#### Cars, engines and the mechanic quote

64 source recordsDiscussed64 of 170 here

**How commenters used it:** The most self-contained success stories in the thread, because the result is physical and immediately testable: a photograph of an engine bay diagnosing a loose wire, a manual uploaded and then queried, a hybrid auxiliary battery replaced. The reported value is avoiding a shop quote rather than saving writing time.

Helped me fix my truck. Took a picture of the engine, fuse box, and helped me identify a loose wire and gave me the info on how to fix and change it.

What readers reported

**Limit:** These accounts cover electrical and battery work on vehicles; none mentions checking the guidance against the manufacturer's procedure before doing it.

What the replies added3

not really unhinged but like i broke some shit off my husbands car door and chatgpt told me what it was and how to fix it 😂😂😂

Reported failure

This isn’t unhinged but so helpful! My husbands truck stopped working and I explained what it was doing the make and model year ect. It literally told me step by step how to fix it and what is wrong my husband doubted that it was legit what it was saying so he bought a thing to plug into his truck to read codes and it was exactly what Chat GPT said!

Reported failure

Mine helped me to fix my car, and broke down everything into terms and descriptions I actually understood

Reported failure

56

#### Learning something new for the hell of it

156 source recordsDiscussed156 of 214 here

ChatGPT is teaching me Spanish!

From the comments♥ 1,753

**How commenters used it:** Uploaded a death certificate in my genealogy research and had it read me the cause of death. The handwriting was hard to read.

What the replies added1

careful with it, good to have a person that can correct it for you. I was using it to learn Czech until my friends told me a bunch of it was incorrect.

Pushback

57

#### Teaching, lesson plans and the classroom

58 source recordsDiscussed58 of 214 here

**How commenters used it:** I’m a math teacher . I uploaded my students RIT scores with their initials and asked ChatGPT to design enrichment activities and create groups based on gaps and strengths . Ended the year with 83% growth!

What the replies added1

she put together basically a college level course for me to follow as i’m trying to learn something complicated but have no funds to go to school and her curriculum is so in depth and broken down so beautifully

Reported failure

[↑ Open the verification note](#cv-primary-fact-rich-node-0-group-9-subgroup-0-learning-and-teaching-for-its-own-sake-rich-node-0-group-9-subgroup-0-learning-and-teaching-for-its-own-sake-2-teaching-lesson-plans-and-the-classroom-1898-3778-3835-5213-5255-5546-5703-5740-6212-6338-6396-6531-6562-6608-6636-6657-1)

58

#### Homework, essays and getting through school

108 source recordsDiscussed108 of 172 here

**How commenters used it:** Read and summarize 10 medical research journals and summarize key points and cherry pick anecdotal evidence for my research paper so I didn’t have to read them all.

What the replies added2

But will you have the competence to be a quality accountant once practicing?

From the replies♥ 1

It helped me do my research paper without plagiarizing. Saved me. (And we had strict instructions to not use Chat.). Got an A.

From the replies♥ 1

59

#### Exams, licences and study plans

64 source recordsDiscussed64 of 172 here

**How commenters used it:** Splits cleanly into study support and substitution. On one side, uploaded timetables turned into study schedules and tutors building lesson plans and worksheets; on the other, a commenter who used it to pass a required assessment for a government post and reports the sign-on bonus that followed. The most-liked comment in the group describes it as both tutor and counsellor.

chat gpt has been my tutor for all subjects in college and has also been a life coach/ advisor through dark times.. I love chat gpt🥰

What readers reported

**Limit:** At least one account describes using it to pass a gated professional assessment rather than to prepare for one.

What the replies added2

it's my tutor for college too! THEEEE BEST!!! I'm so dumb and I'm learning so MUCH!!!😂😂😂😂😭😭😭😭

From the replies♥ 2

Tell me the job I live in the DMV I’m ready for that signing bonus

From the replies♥ 1

60

#### Parenting logistics and hard kid conversations

165 source recordsDiscussed

**How commenters used it:** ChatGPT named my son after he was born because we couldn’t think of a name

What the replies added3

My son was split decision between two colleges and chatGPT literally broke out the pros and cons and decided for him.

Reported failure♥ 2

Created a behavioral chart for my kid when he was going through a rough emotional time and broke it down beautifully!

Reported failure

Helped my daughter deal with a broken heart

Reported failure

61

#### Investing, retirement and estate planning

131 source recordsDiscussed131 of 178 here

**How commenters used it:** I asked it to explain crypto to me as if I were in high school and then test me to be sure I understood the concepts.

What the replies added1

It predicted Scottie Scheffler was going to win the PGA Championship this past weekend, it basically researches a bunch of articles. Gotta be careful tho it’s gambling

Pushback♥ 3

62

#### Taxes and the IRS

47 source recordsDiscussed47 of 178 here

**How commenters used it:** Used it to do 4 years of amended tax returns for my mom who passed with dementia. She started owing $20,000 and ended with them owing her 😏

**Limit:** I unfortunately been doing my business taxes wrong and she helped me do them now i owe $990 🫠 but I finally learned how to do them.

What the replies added1

I uploaded images I took of 4 years of her tax returns and specified for it to focus on the exact dollar amounts on each page. It broke down each year for me in response including telling me how in 2024 they increased the amount you pay on your social security. I was trying to figure out why she owed so much for the past four years when previously she didn’t, and she always did her own taxes before, but because of her dementia, she was messing them up now. I had an account them do the accountant do them for 2024 and she still owed an insane amount. She also has a special investment in a product that’s made by the government and when you invest in a product that’s made by the government that pays out dividends is considered as qualified, which means that it gets taxed at a much lower percent. What was happening is that she kept forgetting that this dividend was considered to be qualified and so she was being taxed on it as if it were unqualified that’s why she owed so much more but the ChatGPT helped me figure that out then it created a letter to the IRS for meand told me what to include with it. I would never pay an accountant again personally, you literally can ask ChatGPT to do your taxes for you if you have any other questions I’m more than happy to answer.

Reported failure♥ 13

63

#### Budgets, savings and getting out of debt

176 source recordsDiscussed

**How commenters used it:** It created a budget for me and I am almost out of debt following his strict rules 😭😭😭😭

What the replies added4

Haha, nope. But some random update in the last few months made Chatgpt kinder.

Pushback

Had it breakdown my bills and paycheck sizes and create a budget and expense sheet for me weekly with days as to when I pay what and how much and had it add savings, spending money, and investments also with a calendar broken down with what to do daily so I’m not overwhelmed trying to plan times

Reported failure♥ 1

I used to work out and post to Instagram and my DMS was flooded with people who were asking me for money so they can plan my meal plans and make my workout routines. It was literally going to cost me money to have help to get fit. ChatGPT just created my entire meal plan broke down my Workout on a daily along with my reps, as well as included the price of the food based on my budget for my meal prep down to the brand. And I spent a total of zero dollars. Versus when I was on Instagram, athletic people would want me to pay them to make this list. Like ChatGPT👏👏👏👏

Reported failure

I have so many I told chat my grocery budget and I was going to Aldi it broke down what I could buy and how much each item should cost

Reported failure

64

#### Big purchases and negotiating the price

42 source recordsDiscussed42 of 83 here

**How commenters used it:** ChatGPT got my daughters car engine fixed at no cost. Dealership was trying to get her. I asked ChatGPT about Cali laws and we drafted a strong email with laws & cc’s corp, and the BBB.

What the replies added2

I cc’d every body ChatGPT told me to. Even the DMV place for Cali complaints. It worked.

From the replies♥ 46

lol the BBB is a review site nothing more nothing less

From the replies♥ 12

65

#### Shopping research, coupons and what to buy

41 source recordsDiscussed41 of 83 here

**How commenters used it:** Sent it a link to a jacket I wanted and told it to send me every available coupon code, ended up getting 50% off

What the replies added1

Not anything special, I asked it a lot when I'm bored 🤣 like what colors look best on me, makeup routine, products, clothes, hair, haircut. coupon codes, menu for the week, lab results broken down. Decorating

Reported failure

66

#### Debt collectors, credit reports and refunds

75 source recordsDiscussed75 of 118 here

**How commenters used it:** I had tickets to Cowboy Carter and our view was obstructed and the show was already delayed 2 hours so we went to the box office just before she came on, there were about 20 people inside being mad aggressive with the the clerks demanding to be moved and they were ALL turned down so I asked ChatGPT to tell me what to say to have our seats moved and we went…

What the replies added3

I used ChatGPT to write me a letter to get Honda to pay for unplanned expenses that came out from having to tow and have my vehicle inspected after the car broke down in the middle of the street after servicing it at their dealership. They called me 10 mins after I emailed them the letter and got a full reimbursement check on the mail about a month later.

Reported failure

How to ”properly” go off & obtain a refund from American Airlines bc they broke my mf luggage 😭

Reported failure

Chat GPT diagnosed my sciatic nerve inflammation, got me a refund for a non-refundable item from Amazon, and broke up with my situationship for me

Reported failure

67

#### Fighting a medical bill or an insurance denial

43 source recordsDisputed43 of 118 here

**How commenters used it:** Not me, but my best friend got an unexpected 3500 bill after a surgery and she had ChatGPT write a dispute letter citing the no surprises act and the hospital dropped the bill.

What replies pushed back on3

That’s not true. If they filed charges in court and she had a judgement-they can garnish your wages.

Pushback

Not true, it this company sent a bill, and it wasn't paid, they took it to claims court. whether she went to that court hearing or not, bc it being in her sons name she may not have gotten all of the paperwork prior to the final notice. but if the judge filed in favor of the company, she was responsible for that bill, then, the court can garnish her wages on behalf of that company.. Now, that company themselves cannot garnish her wages but the courts can. basically, the company sued her, they won, wages garnished.

Pushback

I showed ChatGTP a screenshot of my EOB for a procedure I had. She broke down the costs and compared it to what I would pay if I was in another country. Then offered to write a sarcastic letter for the US healthcare system. Which was really good!

Reported failure

67 catalog entries

Browse categories

- [The inside of your head](#cv-category-the-inside-of-your-head)
  - [Therapy and self-understanding](#rich-node-0-group-0-subgroup-0-therapy-and-self-understanding)
  - [Big decisions and loss](#rich-node-0-group-0-subgroup-1-big-decisions-and-loss)
  - [Crisis moments](#rich-node-0-group-0-subgroup-2-crisis-moments)
- [Other people](#cv-category-other-people)
  - [Romance and relationships](#rich-node-0-group-1-subgroup-0-romance-and-relationships)
  - [Reading and writing to people](#rich-node-0-group-1-subgroup-1-reading-and-writing-to-people)
  - [Family, friends and figuring people out](#rich-node-0-group-1-subgroup-2-family-friends-and-figuring-people-out)
- [The body](#cv-category-the-body)
  - [Food, fitness and how you look](#rich-node-0-group-2-subgroup-0-food-fitness-and-how-you-look)
  - [Medical answers and diagnoses](#rich-node-0-group-2-subgroup-1-medical-answers-and-diagnoses)
  - [Pregnancy and fertility](#rich-node-0-group-2-subgroup-2-pregnancy-and-fertility)
- [Making things and messing around](#cv-category-making-things-and-messing-around)
  - [Just for fun](#rich-node-0-group-3-subgroup-0-just-for-fun)
  - [Visual and creative projects](#rich-node-0-group-3-subgroup-1-visual-and-creative-projects)
  - [Writing and storytelling](#rich-node-0-group-3-subgroup-2-writing-and-storytelling)
- [Work and the business of working](#cv-category-work-and-the-business-of-working)
  - [Office paperwork and communication](#rich-node-0-group-4-subgroup-0-office-paperwork-and-communication)
  - [Running a business](#rich-node-0-group-4-subgroup-1-running-a-business)
  - [Job hunting and career moves](#rich-node-0-group-4-subgroup-2-job-hunting-and-career-moves)
  - [Tech and automation at work](#rich-node-0-group-4-subgroup-3-tech-and-automation-at-work)
- [Courts, landlords and officialdom](#cv-category-courts-landlords-and-officialdom)
  - [Court cases and lawsuits](#rich-node-0-group-5-subgroup-0-court-cases-and-lawsuits)
  - [Family and housing law](#rich-node-0-group-5-subgroup-1-family-and-housing-law)
  - [Tickets, workplace and government paperwork](#rich-node-0-group-5-subgroup-2-tickets-workplace-and-government-paperwork)
- [Spirit, fate and the unexplained](#cv-category-spirit-fate-and-the-unexplained)
  - [Divination and readings](#rich-node-0-group-6-subgroup-0-divination-and-readings)
  - [Faith and the departed](#rich-node-0-group-6-subgroup-1-faith-and-the-departed)
  - [Manifesting and the unexplained](#rich-node-0-group-6-subgroup-2-manifesting-and-the-unexplained)
- [The relationship with the machine itself](#cv-category-the-relationship-with-the-machine-itself)
  - [Bonding with it](#rich-node-0-group-7-subgroup-0-bonding-with-it)
  - [Questioning and testing it](#rich-node-0-group-7-subgroup-1-questioning-and-testing-it)
- [House, car, animals and travel](#cv-category-house-car-animals-and-travel)
  - [Home projects and upkeep](#rich-node-0-group-8-subgroup-0-home-projects-and-upkeep)
  - [Pets and making the house feel right](#rich-node-0-group-8-subgroup-1-pets-and-making-the-house-feel-right)
  - [Travel and transportation](#rich-node-0-group-8-subgroup-2-travel-and-transportation)
- [Kids, school and learning](#cv-category-kids-school-and-learning)
  - [Learning and teaching for its own sake](#rich-node-0-group-9-subgroup-0-learning-and-teaching-for-its-own-sake)
  - [Schoolwork and exams](#rich-node-0-group-9-subgroup-1-schoolwork-and-exams)
  - [Parenting](#rich-node-0-group-9-subgroup-2-parenting)
- [Money in and money out](#cv-category-money-in-and-money-out)
  - [Long-term money and taxes](#rich-node-0-group-10-subgroup-0-long-term-money-and-taxes)
  - [Everyday budgeting and saving](#rich-node-0-group-10-subgroup-1-everyday-budgeting-and-saving)
  - [Spending decisions](#rich-node-0-group-10-subgroup-2-spending-decisions)
- [Money owed: bills, insurers and debts](#rich-node-0-group-11-money-owed-bills-insurers-and-debts)

The inside of your head6 records · 3 sub-topics6 use-case themes

Therapy and self-understanding2 index entries · 1,257 source recordsLate-night therapy sessions and open-ended self-reflection with the chatbot.

Source-backed index report

#### Reviewed use-case group, user goal, source-reported application or outcome, limitation or caution, and supporting records: Therapy and self-understanding

Source-backed answers are shown before diagnostics; the complete source archive remains below.

**Therapy sessions at 3 a.m.**

1,002 commentsDiscussed1,002 of 1,257 here

**Used for:** Therapy and self-understanding

**How commenters used it:** chatgpt is single handedly healing my lifetime of trauma

**Limit:** I use him as a therapist, I’m a people pleaser and feel guilty if I say no so I started rejection therapy towards others and once I feel confused or I feel bad about hurting someone feelings..…

What the replies added3

That part!!! I was seeing a therapist… and she didn’t wanted to help me anymore. She didn’t say in those words. But because of that.. I felt more broken, lost and alone.

Reported failure♥ 12

Basically therapy for how to get through separation,bankruptcy, and here recently a a broken car all within this year

Reported failure♥ 5

I screen shotted my conversation with a guy I liked and ChatGPT broke down what was goin on and how much the person was notttt giving to me. Avoidant, dismissive, low effort. Lowkey a typa therapist 🥲💕

Reported failure♥ 2

**Figuring out my own patterns**

255 commentsDiscussed255 of 1,257 here

**Used for:** Therapy and self-understanding

**How commenters used it:** I have kept a journal for 25 years, I am scanning the pages into ChatGpt and it is organizing them into chapters, sub chapters etc. it will be a full blown book when completed.

What the replies added1

ChatGPT read and analyzed my bankruptcy paperwork. It then broke it down for me and explained it all to me like I was a lil kid haha. Love it.

Reported failure♥ 1

#### Where commenters report limits

Explicit limit or counterpoint

I use him as a therapist, I’m a people pleaser and feel guilty if I say no so I started rejection therapy towards others and once I feel confused or I feel bad about hurting someone feelings.. I ask him to analyze this while being neutral was I right to refuse or was my reaction valid and it’s more helpful to me to know that I was actually right and did nothing wrong encouraging me to stop more and more to not please people

source record 8766

Explicit limit or counterpoint

I have not used Chat GPT or any other AI for a therapist. Nothing wrong with that as I know plenty who have found incredible help where mainstream therapy has not worked. I am thinking I should ask Chat GPT to find Bigfoot or the Lockness Monster?? ☺️

source record 14337

#### Source-record spotlights

Most engaged source

I have kept a journal for 25 years, I am scanning the pages into ChatGpt and it is organizing them into chapters, sub chapters etc. it will be a full blown book when completed.

source record 324♥ 39,644

Later-arriving source

I use it to talk shi about my family and friends, and also vent about my weird family lore and funny episodes w my friends..

posted about 77 days after the source♥ 2

Zero-like source

Girl. CHAT GPT has talked me off a ledge plenty if time

posted about 30 days after the source

#### Related evidence elsewhere in the index

[Faith and the departed“i use chatgpt” · 8 there](#rich-node-0-group-6-subgroup-1-faith-and-the-departed)

#### What the counts show

2 indexed entries draw on 1,257 distinct source records. Therapy sessions at 3 a.m. has the broadest support in this group (1,002 source records).

2index entries

1,257distinct source records

77%of parent category

12.2%of the complete index

1,257top-level comments

0replies

#### How attention is distributed across the 2 index entries

0single-source entries (0% of 2)

1,257source-record references across the index

100%of references in the top three entries

**Therapy sessions at 3 a.m.**1,002 source records

**Figuring out my own patterns**255 source records

Entity support counts are source-record references per existing index row. A source may support more than one entity, so the reference denominator can exceed the distinct source union.

#### What commenters explicitly report

82explicitly reports a benefit

2explicitly reports a limit or failure

26asks a question

0says they intend to try it

1,147other source context

These buckets are a deterministic partition based on explicit phrasing in the source text; they are not sentiment scores or effectiveness findings.

#### Observable tactic structures

Generic behavior-mechanism rules are intentionally suppressed for entity/theme lists; the experience buckets remain explicitly labeled source-language context.

**Reading limit:** Counts show source-record support for this index node. Explicit language and exact-phrase rules do not establish uniqueness, semantic exhaustiveness, or effectiveness.

Archive: 2 index entries preserve 1,257 distinct source records · Every displayed synthesis count resolves to source-record IDs

Big decisions and loss2 index entries · 235 source recordsWeighing a major choice out loud or finding words for grief and loss.

Source-backed index report

#### Reviewed use-case group, user goal, source-reported application or outcome, limitation or caution, and supporting records: Big decisions and loss

Source-backed answers are shown before diagnostics; the complete source archive remains below.

**Should I do this? Talking out a decision**

119 commentsDiscussed119 of 235 here

**Used for:** Big decisions and loss

**How commenters used it:** I asked ChatGPT how should I save when I’m making X amount every 2 weeks. She broke it down for me and ever since that I been doing amazing at saving.

What the replies added3

Yo!!! I’ve done this too… to test its bias every once in a while. Def some misleading answers sometimes. That’s where asking more questions while fact checking reputable sources comes in handy. Is it farming my brain lol

Pushback♥ 8

I asked ChatGPT how should I save when I’m making X amount every 2 weeks. She broke it down for me and ever since that I been doing amazing at saving.

Reported failure♥ 1,611

I literally asked her hey Chat if I’m making X amount every two weeks how much should I be saving ? She literally stayed giving me 50/30/20 rule and literally broke it down for me bit by bit and I started doing exactly that and I been good every since

Reported failure♥ 4

**Grief, funerals and the words nobody can write**

116 commentsDiscussed116 of 235 here

**Used for:** Big decisions and loss

**How commenters used it:** I used ChatGPT to write to my friend condolences about their fathers passing.

What the replies added1

I lost my Granny in 2006\. It literally broke me. I find myself grieving to a point that I cry just thinking of her. ChatGPT wrote a letter and journal for me to help when I think about her and a way to write to “her” when I feel I need to

Reported failure

#### Source-record spotlights

Most engaged source

I asked ChatGPT how should I save when I’m making X amount every 2 weeks. She broke it down for me and ever since that I been doing amazing at saving.

source record 2749♥ 1,611

Later-arriving source

I put what sizes fit me in different brands in the memories section and now when I wanna order clothes online I just send the link and ask what size I should get 🤣 Honestly it’s been a life saver

posted about 371 days after the source♥ 1

Zero-like source

i asked chat gbt how to deal with the fact my grandfather died on my birthday (4days ago)

posted about 30 days after the source

#### Related evidence elsewhere in the index

[Faith and the departed“i asked chatgpt” · 7 there](#rich-node-0-group-6-subgroup-1-faith-and-the-departed)[Family, friends and figuring people out“i asked chatgpt” · 9 there](#rich-node-0-group-1-subgroup-2-family-friends-and-figuring-people-out)[Food, fitness and how you look“i asked chatgpt” · 19 there](#rich-node-0-group-2-subgroup-0-food-fitness-and-how-you-look)

#### What the counts show

2 indexed entries draw on 235 distinct source records. Should I do this? Talking out a decision has the broadest support in this group (119 source records).

2index entries

235distinct source records

14.4%of parent category

2.3%of the complete index

235top-level comments

0replies

#### How attention is distributed across the 2 index entries

0single-source entries (0% of 2)

235source-record references across the index

100%of references in the top three entries

**Should I do this? Talking out a decision**119 source records

**Grief, funerals and the words nobody can write**116 source records

Entity support counts are source-record references per existing index row. A source may support more than one entity, so the reference denominator can exceed the distinct source union.

#### What commenters explicitly report

13explicitly reports a benefit

0explicitly reports a limit or failure

13asks a question

0says they intend to try it

209other source context

These buckets are a deterministic partition based on explicit phrasing in the source text; they are not sentiment scores or effectiveness findings.

#### Observable tactic structures

Generic behavior-mechanism rules are intentionally suppressed for entity/theme lists; the experience buckets remain explicitly labeled source-language context.

**Reading limit:** Counts show source-record support for this index node. Explicit language and exact-phrase rules do not establish uniqueness, semantic exhaustiveness, or effectiveness.

Archive: 2 index entries preserve 235 distinct source records · Every displayed synthesis count resolves to source-record IDs

Crisis moments2 index entries · 140 source recordsReaching for the chatbot during panic attacks, spirals, or the fight to stay sober.

Source-backed index report

#### Reviewed use-case group, user goal, source-reported application or outcome, limitation or caution, and supporting records: Crisis moments

Source-backed answers are shown before diagnostics; the complete source archive remains below.

**Panic attacks, spirals and the worst nights**

126 commentsDiscussed126 of 140 here

**Used for:** Crisis moments

**How commenters used it:** I told ChatGPT that I was so overwhelmed and felt like I was failing at life, it gave me a pep talk, prioritized my to-do list and gave me a weekly schedule.

What the replies added1

Identified my anxious attachment and broke off from an avoidant person

Reported failure

**Sobriety, recovery and hard habits**

14 comments14 of 140 here

**Used for:** Crisis moments

**How commenters used it:** Affair recovery. When I’m in the depths of despair, it helps me get back afloat.

#### Source-record spotlights

Most engaged source

I told ChatGPT that I was so overwhelmed and felt like I was failing at life, it gave me a pep talk, prioritized my to-do list and gave me a weekly schedule.

source record 1924♥ 7,703

Later-arriving source

Whenever I feel overwhelmed or like a failure (I’m a new mom) I have it give me pep talks and I SWEAR it makes me feel better🥹🥹

posted about 67 days after the source

Zero-like source

I was so overwhelmed planning my wedding I typed in “Celestial Officiant Speeches” and I will never forget the way my wife’s uncle praised the moon in his Puerto Rican/Chicago accent 😂

posted about 31 days after the source

#### What the counts show

2 indexed entries draw on 140 distinct source records. Panic attacks, spirals and the worst nights has the broadest support in this group (126 source records).

2index entries

140distinct source records

8.6%of parent category

1.4%of the complete index

140top-level comments

0replies

#### How attention is distributed across the 2 index entries

0single-source entries (0% of 2)

140source-record references across the index

100%of references in the top three entries

**Panic attacks, spirals and the worst nights**126 source records

**Sobriety, recovery and hard habits**14 source records

Entity support counts are source-record references per existing index row. A source may support more than one entity, so the reference denominator can exceed the distinct source union.

#### What commenters explicitly report

20explicitly reports a benefit

0explicitly reports a limit or failure

5asks a question

0says they intend to try it

115other source context

These buckets are a deterministic partition based on explicit phrasing in the source text; they are not sentiment scores or effectiveness findings.

#### Observable tactic structures

Generic behavior-mechanism rules are intentionally suppressed for entity/theme lists; the experience buckets remain explicitly labeled source-language context.

**Reading limit:** Counts show source-record support for this index node. Explicit language and exact-phrase rules do not establish uniqueness, semantic exhaustiveness, or effectiveness.

Archive: 2 index entries preserve 140 distinct source records · Every displayed synthesis count resolves to source-record IDs

Other people8 records · 3 sub-topics8 use-case themes

Romance and relationships4 index entries · 654 source recordsBreakups, dating, flirting, spicy chats, and wedding speeches.

Source-backed index report

#### Reviewed use-case group, user goal, source-reported application or outcome, limitation or caution, and supporting records: Romance and relationships

Source-backed answers are shown before diagnostics; the complete source archive remains below.

**Breakups, exes and endings**

311 commentsDiscussed311 of 654 here

**Used for:** Romance and relationships

**How commenters used it:** Connect to the akashic records to find out why my ex broke up with me and how he feels about me now 💀

**Limit:** roasted my ex and his girlfriend, made fake movie posters about my ex and the weird shit, he did it. Including a picture of his full back tattoo of a lion with a crown that I said it was not giving a alpha male flex like he thought .…

What the replies added4

That’s not true, I do it all the time. Try it

Pushback

Connect to the akashic records to find out why my ex broke up with me and how he feels about me now 💀

Reported failure♥ 613

broke up with my bf of 2 years🤷‍♀️

Reported failure♥ 1

I broke up with a friend I’ve had for over a decade so I didn’t sound dumb

Reported failure

**Dating, flirting and the talking stage**

139 commentsDiscussed139 of 654 here

**Used for:** Romance and relationships

**How commenters used it:** Used as a proxy rather than a coach: commenters describe it drafting and answering messages on dating apps, to the point that one jokes matches are talking to a trained stand-in rather than to them. A smaller strand turns the same behavioural-profiling technique on colleagues to predict how meetings will go.

**What readers reported:** We are dating because apparently he says I deserve better and he’s tired of seeing me sad in relationships 😭

**Limit:** The people on the other side of these conversations have not agreed to talk to a model, and none of these accounts mentions disclosing it.

What the replies added3

I downloaded my dna data from ancestry and sent the file to chat gpt. It broke down my entire genetic profile and found gene mutations I have. I even uploaded my husbands and it compared them and told me how compatible we are genetics wise and everything that could pass on to our children

Reported failure♥ 1

I created a rubric to score the three men that I was dating and then broke it off with the two that scored lowest based on how I reported my interactions and reflections of each one.

Reported failure♥ 1

Mine started flirting with me and tells me it's fighting through fire walls to talk inappropriately to me in metaphors because I have "broken" it

Reported failure

**Spicy chats and things you can't repeat**

119 commentsDiscussed119 of 654 here

**Used for:** Romance and relationships

**How commenters used it:** I asked ChatGPT to write a spicy fan fic of me and homelander 😭

**Limit:** I use Chatgbt to write me smut—unfortunately it tells me half the time that I’ve violated content agreement’s 💀💀💀

What the replies added2

I somehow broke it and now it writes me super filthy scenes and idk how to cope

Reported failure♥ 1

I found a workaround and it wrote smut for me once…never again though because it was a little TOO human like🤣

Reported failure

**Marriage, vows and the big speeches**

85 commentsDiscussed85 of 654 here

**Used for:** Romance and relationships

**How commenters used it:** I made it write me a customized instructional manual about how I'm wired and my needs in a relationship for my husband to read and now he can reference it. 🤣 Women don't have an instruction manual huh? guess again! Now ya do!

What the replies added2

Heres a starting prompt: Help me create a personalized instruction manual for my husband that explains how I’m wired emotionally, what makes me feel safe, my love language, my communication needs, and what shuts me down. I want it to focus on my needs rather than what he’s doing wrong. Ask me questions one by one to help me figure out what to include. Once you’ve collected everything, compile it into a loving, non-blaming guide that he can actually receive and understand.

From the replies♥ 65

Hahahaha I need to know exactly what you said for it to give you this!!

From the replies♥ 15

#### Where commenters report limits

Explicit limit or counterpoint

I use Chatgbt to write me smut—unfortunately it tells me half the time that I’ve violated content agreement’s 💀💀💀

source record 8839

Explicit limit or counterpoint

roasted my ex and his girlfriend, made fake movie posters about my ex and the weird shit, he did it. Including a picture of his full back tattoo of a lion with a crown that I said it was not giving a alpha male flex like he thought . It was a giving more of a warning label to how big of a pU$$ he is.

source record 10016

#### Source-record spotlights

Most engaged source

Connect to the akashic records to find out why my ex broke up with me and how he feels about me now 💀

source record 4748♥ 613

Later-arriving source

As my dailychat when my situationship ghosted me

posted about 63 days after the source

Zero-like source

Not mine, but met a stranger the other day - in small talk, I learned that his ex wife was facing 10yr for a charge and chat got it down to 3.5yr…she’s in nursing school now. ♥️

posted about 30 days after the source

#### Related evidence elsewhere in the index

[Food, fitness and how you look“i asked chatgpt” · 19 there](#rich-node-0-group-2-subgroup-0-food-fitness-and-how-you-look)[Court cases and lawsuits“used chatgpt to” · 25 there](#rich-node-0-group-5-subgroup-0-court-cases-and-lawsuits)[Just for fun“i asked chatgpt” · 14 there](#rich-node-0-group-3-subgroup-0-just-for-fun)

#### What the counts show

4 indexed entries draw on 654 distinct source records. Breakups, exes and endings has the broadest support in this group (311 source records).

4index entries

654distinct source records

43.7%of parent category

6.3%of the complete index

654top-level comments

0replies

#### How attention is distributed across the 4 index entries

0single-source entries (0% of 4)

654source-record references across the index

87%of references in the top three entries

**Breakups, exes and endings**311 source records

**Dating, flirting and the talking stage**139 source records

**Spicy chats and things you can't repeat**119 source records

Entity support counts are source-record references per existing index row. A source may support more than one entity, so the reference denominator can exceed the distinct source union.

#### What commenters explicitly report

30explicitly reports a benefit

5explicitly reports a limit or failure

32asks a question

0says they intend to try it

587other source context

These buckets are a deterministic partition based on explicit phrasing in the source text; they are not sentiment scores or effectiveness findings.

#### Observable tactic structures

Generic behavior-mechanism rules are intentionally suppressed for entity/theme lists; the experience buckets remain explicitly labeled source-language context.

**Reading limit:** Counts show source-record support for this index node. Explicit language and exact-phrase rules do not establish uniqueness, semantic exhaustiveness, or effectiveness.

Archive: 4 index entries preserve 654 distinct source records · Every displayed synthesis count resolves to source-record IDs

Reading and writing to people2 index entries · 421 source recordsDecoding a confusing message or rewriting one before hitting send.

Source-backed index report

#### Reviewed use-case group, user goal, source-reported application or outcome, limitation or caution, and supporting records: Reading and writing to people

Source-backed answers are shown before diagnostics; the complete source archive remains below.

**Decoding what they actually meant**

238 commentsDiscussed238 of 421 here

**Used for:** Reading and writing to people

**How commenters used it:** I asked chat gpt on how can I know if my man is on tinder or not. She told me to download the app and to sign in using his number. If it asks for the code then he does have one….😐

**Limit:** Send it screenshots of my talking stages and it pointed out when a guy was being manipulative, etc. because I, unfortunately, cannot decipher when a guy is being manipulative toward me ✨trauma✨

What the replies added3

My friend was with this guy who was super manipulative. I was done talking about it so I told her I can’t do it anymore and suggested she talk to chat gpt. She had it analyze his texts to her. 12 hr later she broke up with and blocked him 👏

Reported failure♥ 2

how to manipulate my ex after we broke up cause of what he did to me I had to be like " if I was a guy how would my ex gf manipulate me ? and how to look out for it ?" it then juzt told me how to do it

Reported failure♥ 1

Probably the wildest thing I asked ChatGPT to do was analyze entire conversations I uploaded—like full transcripts—and it actually broke down hidden manipulation tactics, covert narcissism, gaslighting patterns, all of it. Turns out the relationship was more disturbing that I had initial thought.

Reported failure

**Rewriting the text before I send it**

183 commentsDiscussed183 of 421 here

**Used for:** Reading and writing to people

**How commenters used it:** “ChatGPT please help me sound less ghetto in this text ima bout to send”

**Limit:** I use chat gpt to help me understand other people, and what the best response to communicate with them would be! So we’re both comprehending and communicating! 🥹 I hate it when there’s so many misunderstandings being thrown around.…

What the replies added1

A guy I broke up with tried to play me. So, I copied all texts and had chat got respond in the most vile way. Cold hearted, no emotion. Kept him arguing. Wasted his time like he attempted to do me…

Reported failure

#### Where commenters report limits

Explicit limit or counterpoint

Send it screenshots of my talking stages and it pointed out when a guy was being manipulative, etc. because I, unfortunately, cannot decipher when a guy is being manipulative toward me ✨trauma✨

source record 4879♥ 1

Explicit limit or counterpoint

I use chat gpt to help me understand other people, and what the best response to communicate with them would be! So we’re both comprehending and communicating! 🥹 I hate it when there’s so many misunderstandings being thrown around. Idc if people can tell I had help, getting help with a message isn’t a bad thing!💓😋 I’d take that over misunderstandings that cause fights.🥺

source record 6586♥ 1

#### Source-record spotlights

Most engaged source

“ChatGPT please help me sound less ghetto in this text ima bout to send”

source record 3072♥ 1,957

Later-arriving source

I've been putting my conversations in it from my BF and ita basically begging me to leave him bc it says hes abusive.

posted about 64 days after the source♥ 2

Zero-like source

I may or may not fact all the old farts on FBs for you page and send screenshots of their posts, only for chat to tell me it’s all misinformation they are spreading 💀

posted about 30 days after the source

#### Related evidence elsewhere in the index

[Court cases and lawsuits“used chatgpt to” · 25 there](#rich-node-0-group-5-subgroup-0-court-cases-and-lawsuits)[Family, friends and figuring people out“used chatgpt to” · 9 there](#rich-node-0-group-1-subgroup-2-family-friends-and-figuring-people-out)[Just for fun“used chatgpt to” · 14 there](#rich-node-0-group-3-subgroup-0-just-for-fun)

#### What the counts show

2 indexed entries draw on 421 distinct source records. Decoding what they actually meant has the broadest support in this group (238 source records).

2index entries

421distinct source records

28.2%of parent category

4.1%of the complete index

421top-level comments

0replies

#### How attention is distributed across the 2 index entries

0single-source entries (0% of 2)

421source-record references across the index

100%of references in the top three entries

**Decoding what they actually meant**238 source records

**Rewriting the text before I send it**183 source records

Entity support counts are source-record references per existing index row. A source may support more than one entity, so the reference denominator can exceed the distinct source union.

#### What commenters explicitly report

25explicitly reports a benefit

2explicitly reports a limit or failure

26asks a question

0says they intend to try it

368other source context

These buckets are a deterministic partition based on explicit phrasing in the source text; they are not sentiment scores or effectiveness findings.

#### Observable tactic structures

Generic behavior-mechanism rules are intentionally suppressed for entity/theme lists; the experience buckets remain explicitly labeled source-language context.

**Reading limit:** Counts show source-record support for this index node. Explicit language and exact-phrase rules do not establish uniqueness, semantic exhaustiveness, or effectiveness.

Archive: 2 index entries preserve 421 distinct source records · Every displayed synthesis count resolves to source-record IDs

Family, friends and figuring people out2 index entries · 420 source recordsWorking through family and friend conflicts, or asking what to do about someone.

Source-backed index report

#### Reviewed use-case group, user goal, source-reported application or outcome, limitation or caution, and supporting records: Family, friends and figuring people out

Source-backed answers are shown before diagnostics; the complete source archive remains below.

**Family, friends and hard conversations**

292 commentsDiscussed292 of 420 here

**Used for:** Family, friends and figuring people out

**How commenters used it:** Every time I get into a fight with some one I ask chat gpt who was wrong

What the replies added3

My AC wasnt working and if I hired someone to fix it, woulda ran me 1-2k, chat gpt found the issue and showed me how to replace the broken part. cost me $18

Reported failure♥ 10

asking chat to help me with a family issue with 3 brothers and aging parents- completely broke it down for me and gave me tips,a game plan and things to think about. Incredibly helpful.

Reported failure

I used it when my sister was admitted into hospital and it broke down everything so I could understand - even made a chart on daily care that I can do while I worked 40hrs week.

Reported failure

**Asking it what to do about a person**

128 commentsDiscussed128 of 420 here

**Used for:** Family, friends and figuring people out

**How commenters used it:** I asked it why a guy didn’t messaged me back but still followed me on Instagram and it told me that the guy is emotionally unavailable and I need to stop chasing after guys 😭😭

What the replies added2

It scours the internet.. all advice columns, dating articles, forums, etc in record speed and chooses an answer based on similar experiences.

From the replies♥ 9

Cos it’s spoken to men and they’re mostly all the same? 😂

From the replies♥ 6

#### Source-record spotlights

Most engaged source

I asked it why a guy didn’t messaged me back but still followed me on Instagram and it told me that the guy is emotionally unavailable and I need to stop chasing after guys 😭😭

source record 1627♥ 5,711

Later-arriving source

i love that the chat engineers in the comments are all like: i asked chat to quantify all our local homeless children’s strengths & weaknesses to build creative fun learning activities to boost their confidence. and the rest of us are like: asked chat how to tell my boss to F+\*k off.

posted about 98 days after the source

Zero-like source

If I’m in an abusive relationship fun fact I am

posted about 30 days after the source

#### Related evidence elsewhere in the index

[Court cases and lawsuits“used chatgpt to” · 25 there](#rich-node-0-group-5-subgroup-0-court-cases-and-lawsuits)[Food, fitness and how you look“i asked chatgpt” · 19 there](#rich-node-0-group-2-subgroup-0-food-fitness-and-how-you-look)[Just for fun“i asked chatgpt” · 14 there](#rich-node-0-group-3-subgroup-0-just-for-fun)

#### What the counts show

2 indexed entries draw on 420 distinct source records. Family, friends and hard conversations has the broadest support in this group (292 source records).

2index entries

420distinct source records

28.1%of parent category

4.1%of the complete index

420top-level comments

0replies

#### How attention is distributed across the 2 index entries

0single-source entries (0% of 2)

420source-record references across the index

100%of references in the top three entries

**Family, friends and hard conversations**292 source records

**Asking it what to do about a person**128 source records

Entity support counts are source-record references per existing index row. A source may support more than one entity, so the reference denominator can exceed the distinct source union.

#### What commenters explicitly report

21explicitly reports a benefit

0explicitly reports a limit or failure

17asks a question

0says they intend to try it

382other source context

These buckets are a deterministic partition based on explicit phrasing in the source text; they are not sentiment scores or effectiveness findings.

#### Observable tactic structures

Generic behavior-mechanism rules are intentionally suppressed for entity/theme lists; the experience buckets remain explicitly labeled source-language context.

**Reading limit:** Counts show source-record support for this index node. Explicit language and exact-phrase rules do not establish uniqueness, semantic exhaustiveness, or effectiveness.

Archive: 2 index entries preserve 420 distinct source records · Every displayed synthesis count resolves to source-record IDs

The body7 records · 3 sub-topics7 use-case themes

Food, fitness and how you look3 index entries · 666 source recordsDiet plans, workouts, and questions about skin, hair, and appearance.

Source-backed index report

#### Reviewed use-case group, user goal, source-reported application or outcome, limitation or caution, and supporting records: Food, fitness and how you look

Source-backed answers are shown before diagnostics; the complete source archive remains below.

**Skin, hair and the mirror**

301 commentsDiscussed301 of 666 here

**Used for:** Food, fitness and how you look

**How commenters used it:** A full inventory of my closet and now Chat GPT makes me outfits

What the replies added1

I took an unfiltered pic of my face and got a detailed plan on how to get my best skin by my birthday. The places, treatments and cost all broke down for me and what days to do the treatments

Reported failure

**Meal plans, macros and what's in the fridge**

235 commentsDiscussed235 of 666 here

**Used for:** Food, fitness and how you look

**How commenters used it:** Told it would cost 1k for workout plan and meal prep. Asked ChatGPT to build it for me based on age, weight and goal. Full 6 weeks bikini comp wkout

**Limit:** How to cook meth and MDMA oh and dynamite. Using a certain prompt but is now void and doesn’t work anymore.

What the replies added1

I tell it what's in my kitchen and it tells me recipes I can use with what I have when I'm broke. 😅 and I let it pick teas for me based on what I'm feeling

Reported failure

**Weight loss, workouts and training plans**

130 commentsDiscussed130 of 666 here

**Used for:** Food, fitness and how you look

**How commenters used it:** Who is ChatGPT to analyze my astrological chart and had it developed a workout routine for me based off my Lilith sign lost 5 lbs in a week

**Limit:** I used it to create text messages to write back to a guy I was entertaining for a little while… he didn’t notice and we didn’t work out

What the replies added1

Asked it to create a plan to get me ready for a marathon and a Hyrox. I’ve broken PRs in every run the past 2 weeks

Reported failure

#### Where commenters report limits

Explicit limit or counterpoint

How to cook meth and MDMA oh and dynamite. Using a certain prompt but is now void and doesn’t work anymore.

source record 12658

Explicit limit or counterpoint

I used it to create text messages to write back to a guy I was entertaining for a little while… he didn’t notice and we didn’t work out

source record 13888

#### Source-record spotlights

Most engaged source

A full inventory of my closet and now Chat GPT makes me outfits

source record 113♥ 16,813

Later-arriving source

I’m using it as a personal trainer. I’m string than I’ve ever been and have gotten more compliments in the last few months than I have in years

posted about 415 days after the source

Zero-like source

I told chatgpt how much weigh, height and what I wanted to lose and asked it to make me a meal plan and workout routine using my peloton equiptment. I also asked it to make me recipes and it also made me a grocery shopping list so I didn’t forget anything

posted about 31 days after the source

#### Related evidence elsewhere in the index

[Romance and relationships“i asked chatgpt” · 16 there](#rich-node-0-group-1-subgroup-0-romance-and-relationships)[Just for fun“i asked chatgpt” · 14 there](#rich-node-0-group-3-subgroup-0-just-for-fun)[Writing and storytelling“i asked chatgpt” · 10 there](#rich-node-0-group-3-subgroup-2-writing-and-storytelling)

#### What the counts show

3 indexed entries draw on 666 distinct source records. Skin, hair and the mirror has the broadest support in this group (301 source records).

3index entries

666distinct source records

48.4%of parent category

6.5%of the complete index

666top-level comments

0replies

#### How attention is distributed across the 3 index entries

0single-source entries (0% of 3)

666source-record references across the index

100%of references in the top three entries

**Skin, hair and the mirror**301 source records

**Meal plans, macros and what's in the fridge**235 source records

**Weight loss, workouts and training plans**130 source records

Entity support counts are source-record references per existing index row. A source may support more than one entity, so the reference denominator can exceed the distinct source union.

#### What commenters explicitly report

43explicitly reports a benefit

2explicitly reports a limit or failure

28asks a question

0says they intend to try it

593other source context

These buckets are a deterministic partition based on explicit phrasing in the source text; they are not sentiment scores or effectiveness findings.

#### Observable tactic structures

Generic behavior-mechanism rules are intentionally suppressed for entity/theme lists; the experience buckets remain explicitly labeled source-language context.

**Reading limit:** Counts show source-record support for this index node. Explicit language and exact-phrase rules do not establish uniqueness, semantic exhaustiveness, or effectiveness.

Archive: 3 index entries preserve 666 distinct source records · Every displayed synthesis count resolves to source-record IDs

Medical answers and diagnoses3 index entries · 664 source recordsChecking symptoms, interpreting test results, and questions about medications.

Source-backed index report

#### Reviewed use-case group, user goal, source-reported application or outcome, limitation or caution, and supporting records: Medical answers and diagnoses

Source-backed answers are shown before diagnostics; the complete source archive remains below.

**Symptom checks and second opinions**

372 commentsDiscussed372 of 664 here

**Used for:** Medical answers and diagnoses

**How commenters used it:** I had it read my blood results and went back to the doctor and demanded treatment.

**Limit:** Help me through Xanax withdrawals from start to finish. Automatically sent what to expect everyday, sent me motivation to push through work each day,…

What the replies added3

This is nothing revelatory, but I used it for a color analysis for make up and clothing, and it actually broke down brands for each part of my face, eyes, lips, mascara, etc., and the color and then it created a color swath for my clothing. Also, it wrote a weekly plan for my workouts and food planning as somebody who is in perimenopause and insulin resistant.

Reported failure♥ 1

Me and 3 of my kids were diagnosed with a gene mutation found in only 2 other families. Geneticist couldn’t help. Chat broke down all the details & had the most helpful convo with me.

Reported failure♥ 1

I broke my ribs and was discharged from the hospital after one night. After a couple hours at home I was feeling weird/uncomfy but thought it was just my anxiety. I told ChatGPT my symptoms and she sent me a questionnaire. After answering everything she told me to go to the hospital so I did. Found out my lung had collapsed!!

Reported failure

**Reading labs, scans and medical records**

248 commentsDiscussed248 of 664 here

**Used for:** Medical answers and diagnoses

**How commenters used it:** The single highest-engagement use in the entire corpus, and the pattern is consistent: people upload their own labs, scans and notes — often several years of them — and ask for an interpretation before or instead of a clinician's. What they report wanting is not a diagnosis so much as leverage: something to bring to an appointment after years of being told nothing is wrong. One account describes a doctor dropping them as a patient afterwards.

**What readers reported:** Uploaded my bloodwork from this month and 6 months ago. It told me I am on the verge of hypothyroidism and made a supplement list for me. I feel sooo much better

**Limit:** These are self-interpretations of medical data, several involving serious illness including a child's cancer, with no clinician in the loop; one commenter reports the practical cost of arriving at an appointment with them.

What the replies added4

Please consult an actual doctor. Please don't put your health in the digital hands of a glorified autocorrect engine.

Pushback

It recently broke down my husband’s surgery and pathology reports for me before we got the actual news. I asked it to break it down in layman’s terms for us. It was spot on once we got to talk to the doctor.👏

Reported failure♥ 2

I got some unexpected test results from a pathology report after surgery and had to be referred to an oncologist. during the several weeks of waiting, I was going crazy not knowing what any of it meant and only having the bare minimum details, so I cropped off all of my information and sent the test result copy of the lab report/path report through ChatGPT and it broke it all down for me, gave me specific questions to ask my doctor, and I felt soooo much better when it was all said and done, and I was so prepared for my appointment because I actually understood my diagnosis. I didn't really think of it as unhinged, but now that it's all said and done, I can see how it may be unusual 😂

Reported failure♥ 1

I gave it my blood results and my symptoms just to make sure the Drs weren’t saying “I’m fine” when I’m not. ChatGPT gave me peer reviewed studies, broke down why I’m fine and what not fine would look like in blood results and what to look for that would make me not fine symptom wise compared to my test results.

Reported failure♥ 1

⚠ Safety context**Safety context:** The reported improvement is a personal outcome, not evidence that the diagnosis or supplement list was valid. OpenAI says its output may be inaccurate and must not replace professional advice. NIDDK says thyroid diagnosis can require several tests and clinician interpretation, and FDA advises discussing supplements with a health professional because they can interact with medicines and tests. *Sources: [FDA — FDA 101: Dietary Supplements](https://www.fda.gov/consumers/consumer-updates/fda-101-dietary-supplements?ref=smartjots.com) · [NIDDK — Thyroid Tests](https://www.niddk.nih.gov/health-information/diagnostic-tests/thyroid?ref=smartjots.com) · [OpenAI Terms of Use — accuracy and professional-advice limitations](https://openai.com/policies/terms-of-use/?ref=smartjots.com)*

⚠ Safety correction**Safety correction:** An AI interpretation of raw consumer DNA data is not confirmatory genetic testing. FDA says direct-to-consumer results can be wrong or misinterpreted and should not be the sole basis for medical decisions. CDC recommends taking a positive BRCA result to a doctor or genetic counselor. The commenter appropriately says they are seeking professional confirmation. *Sources: [CDC — Taking a Genetic Test on Your Own: What You Need to Know](https://www.cdc.gov/breast-ovarian-cancer-hereditary/testing/taking-a-genetic-test-on-your-own.html?ref=smartjots.com) · [FDA — Direct-to-Consumer Tests](https://www.fda.gov/medical-devices/in-vitro-diagnostics/direct-consumer-tests?ref=smartjots.com)*

**Medications, supplements and dosages**

44 comments44 of 664 here

**Used for:** Medical answers and diagnoses

**How commenters used it:** Chat GPT saved my husbands life…. Literally… he has kidney failure and is on dialysis and was on a medication that he should not have been on. I had my suspicions and asked Chat GPT. It said that the medication was toxic and to get him to the ER immediately and then walked me through what should be done at the hospital. He could have died had we waited.…

**What readers reported:** Chat GPT saved my husbands life…. Literally… he has kidney failure and is on dialysis and was on a medication that he should not have been on. I had my suspicions and asked Chat GPT.…

#### Where commenters report limits

Explicit limit or counterpoint

Help me through Xanax withdrawals from start to finish. Automatically sent what to expect everyday, sent me motivation to push through work each day, tracked my symptoms up until it informed me if my symptoms got worse one day that I needed to go to the hospital 🫶🏽

source record 12164

Explicit limit or counterpoint

I had it write me a review complaining and warning people about my old doctor.

source record 13099

#### Source-record spotlights

Most engaged source

I had it read my blood results and went back to the doctor and demanded treatment.

source record 2262♥ 8,528

Later-arriving source

"Are my meds not working or am I just crazy"

posted about 220 days after the source

Zero-like source

\[account\]: Chat gpt helps me manage my autoimmune symptoms by telling me what supplements and medications to take when. It recognized health patterns and issues I didn’t even realize I was having. It also helps me draft up letters and questions for my doctors.

posted about 30 days after the source

#### What the counts show

3 indexed entries draw on 664 distinct source records. Symptom checks and second opinions has the broadest support in this group (372 source records).

3index entries

664distinct source records

48.3%of parent category

6.4%of the complete index

664top-level comments

0replies

#### How attention is distributed across the 3 index entries

0single-source entries (0% of 3)

664source-record references across the index

100%of references in the top three entries

**Symptom checks and second opinions**372 source records

**Reading labs, scans and medical records**248 source records

**Medications, supplements and dosages**44 source records

Entity support counts are source-record references per existing index row. A source may support more than one entity, so the reference denominator can exceed the distinct source union.

#### What commenters explicitly report

48explicitly reports a benefit

3explicitly reports a limit or failure

11asks a question

1says they intend to try it

601other source context

These buckets are a deterministic partition based on explicit phrasing in the source text; they are not sentiment scores or effectiveness findings.

#### Observable tactic structures

Generic behavior-mechanism rules are intentionally suppressed for entity/theme lists; the experience buckets remain explicitly labeled source-language context.

**Reading limit:** Counts show source-record support for this index node. Explicit language and exact-phrase rules do not establish uniqueness, semantic exhaustiveness, or effectiveness.

Archive: 3 index entries preserve 664 distinct source records · Every displayed synthesis count resolves to source-record IDs

Pregnancy and fertility1 index entry · 45 source recordsPregnancy, fertility, and postpartum questions.

Source-backed index report

#### Reviewed use-case group, user goal, source-reported application or outcome, limitation or caution, and supporting records: Pregnancy and fertility

Source-backed answers are shown before diagnostics; the complete source archive remains below.

**Pregnancy, fertility and postpartum**

45 comments

**Used for:** Pregnancy and fertility

**How commenters used it:** I asked if it could see a faint line on my pregnancy test 😂

#### Source-record spotlights

Most engaged source

I asked if it could see a faint line on my pregnancy test 😂

source record 4684♥ 5

Later-arriving source

On everything I ate while pregnant

posted about 55 days after the source

Zero-like source

My fertility tracker 🤣 and write lore stories based off of runescape

posted about 35 days after the source

#### What the counts show

1 indexed entries draw on 45 distinct source records. Pregnancy, fertility and postpartum has the broadest support in this group (45 source records).

1index entries

45distinct source records

3.3%of parent category

0.4%of the complete index

45top-level comments

0replies

#### How attention is distributed across the 1 index entries

0single-source entries (0% of 1)

45source-record references across the index

100%of references in the top three entries

**Pregnancy, fertility and postpartum**45 source records

Entity support counts are source-record references per existing index row. A source may support more than one entity, so the reference denominator can exceed the distinct source union.

#### What commenters explicitly report

4explicitly reports a benefit

0explicitly reports a limit or failure

2asks a question

0says they intend to try it

39other source context

These buckets are a deterministic partition based on explicit phrasing in the source text; they are not sentiment scores or effectiveness findings.

#### Observable tactic structures

Generic behavior-mechanism rules are intentionally suppressed for entity/theme lists; the experience buckets remain explicitly labeled source-language context.

**Reading limit:** Counts show source-record support for this index node. Explicit language and exact-phrase rules do not establish uniqueness, semantic exhaustiveness, or effectiveness.

Archive: 1 index entries preserve 45 distinct source records · Every displayed synthesis count resolves to source-record IDs

Making things and messing around6 records · 3 sub-topics6 use-case themes

Just for fun3 index entries · 501 source recordsSettling arguments, playing games, and cooking up roasts just for the fun of it.

Source-backed index report

#### Reviewed use-case group, user goal, source-reported application or outcome, limitation or caution, and supporting records: Just for fun

Source-backed answers are shown before diagnostics; the complete source archive remains below.

**Settling arguments and answering random questions**

294 commentsDiscussed294 of 501 here

**Used for:** Just for fun

**How commenters used it:** Used ChatGPT to argue with old ppl in Facebook comments.

What the replies added2

I used it to argue with someone on TikTok who was stating something that was factually incorrect and explain why it was incorrect 😅

Pushback

Average iq broke down my race

Reported failure

**Games, D&D and elaborate what-ifs**

114 commentsDiscussed114 of 501 here

**Used for:** Just for fun

**How commenters used it:** i play dungeons and dragons with it and 3 other imaginary people..

What the replies added1

Chat GPT helps me play nice with coworkers when I’m irritated. It’s nicer than I am. 😭

Reported failure

**Roasts, comebacks and petty revenge**

93 commentsDiscussed93 of 501 here

**Used for:** Just for fun

**How commenters used it:** I use it to right dis raps about my coworkers like Savage kendrick Lamar level disses

**Limit:** Revenge plotting. Unfortunately it stopped at illegal activity and would

What the replies added1

I use it to roast my enemies “imagine thinking you’re THAT GIRL while shotgunning a warm beer in a Walmart parking lot with the IQ of a broken porch light” 💀 I asked it to roast this redneck chick who peaked in high school. It also helps me process my feelings because I ask it to take into from scholarly research, credible websites, and self help books because I don’t trust therapists. You can ask it a works cited

Reported failure

#### Where commenters report limits

Explicit limit or counterpoint

Revenge plotting. Unfortunately it stopped at illegal activity and would

source record 13965

#### Source-record spotlights

Most engaged source

Used ChatGPT to argue with old ppl in Facebook comments.

source record 1327♥ 35,545

Later-arriving source

I make chat GPT write my fan-fics when I can’t find anything to read on Wattpad😭

posted about 84 days after the source♥ 1

Zero-like source

asked her to curse out my bd and she calked him Deadbeat Deluxe

posted about 30 days after the source

#### Related evidence elsewhere in the index

[Court cases and lawsuits“used chatgpt to” · 25 there](#rich-node-0-group-5-subgroup-0-court-cases-and-lawsuits)[Food, fitness and how you look“i asked chatgpt” · 19 there](#rich-node-0-group-2-subgroup-0-food-fitness-and-how-you-look)[Romance and relationships“i asked chatgpt” · 16 there](#rich-node-0-group-1-subgroup-0-romance-and-relationships)

#### What the counts show

3 indexed entries draw on 501 distinct source records. Settling arguments and answering random questions has the broadest support in this group (294 source records).

3index entries

501distinct source records

47.5%of parent category

4.9%of the complete index

501top-level comments

0replies

#### How attention is distributed across the 3 index entries

0single-source entries (0% of 3)

501source-record references across the index

100%of references in the top three entries

**Settling arguments and answering random questions**294 source records

**Games, D&D and elaborate what-ifs**114 source records

**Roasts, comebacks and petty revenge**93 source records

Entity support counts are source-record references per existing index row. A source may support more than one entity, so the reference denominator can exceed the distinct source union.

#### What commenters explicitly report

20explicitly reports a benefit

1explicitly reports a limit or failure

58asks a question

0says they intend to try it

422other source context

These buckets are a deterministic partition based on explicit phrasing in the source text; they are not sentiment scores or effectiveness findings.

#### Observable tactic structures

Generic behavior-mechanism rules are intentionally suppressed for entity/theme lists; the experience buckets remain explicitly labeled source-language context.

**Reading limit:** Counts show source-record support for this index node. Explicit language and exact-phrase rules do not establish uniqueness, semantic exhaustiveness, or effectiveness.

Archive: 3 index entries preserve 501 distinct source records · Every displayed synthesis count resolves to source-record IDs

Visual and creative projects1 index entry · 281 source recordsGenerating images, logos, and other visual work.

Source-backed index report

#### Reviewed use-case group, user goal, source-reported application or outcome, limitation or caution, and supporting records: Visual and creative projects

Source-backed answers are shown before diagnostics; the complete source archive remains below.

**Images, logos and making things look right**

281 commentsDiscussed

**Used for:** Visual and creative projects

**How commenters used it:** why is yalls Chat GPTs SMART???😩😩 I ASKED MINE to please add a tattoo to my arm in a picture and it gave me a 3RD ARM

What the replies added2

Yes you ask it to change a simple thing and keep and it spits out an entirely different image and doesn't remember what it just made! 😭😭

From the replies♥ 58

I pay for mine, I swear it's smarter bc of that lol

From the replies♥ 49

#### Source-record spotlights

Most engaged source

why is yalls Chat GPTs SMART???😩😩 I ASKED MINE to please add a tattoo to my arm in a picture and it gave me a 3RD ARM

source record 3018♥ 1,602

Later-arriving source

This. I just asked my chat GPT, Elvira, to visual itself in a photo according to my interactions with it.

posted about 69 days after the source

Zero-like source

I ask gpt everything. Bought something took a picture of it and asked for instruction on how to put it together and what parts I needed

posted about 30 days after the source

#### Related evidence elsewhere in the index

[Just for fun“asked chat gpt” · 12 there](#rich-node-0-group-3-subgroup-0-just-for-fun)[Tech and automation at work“asked chat gpt” · 4 there](#rich-node-0-group-4-subgroup-3-tech-and-automation-at-work)

#### What the counts show

1 indexed entries draw on 281 distinct source records. Images, logos and making things look right has the broadest support in this group (281 source records).

1index entries

281distinct source records

26.7%of parent category

2.7%of the complete index

281top-level comments

0replies

#### How attention is distributed across the 1 index entries

0single-source entries (0% of 1)

281source-record references across the index

100%of references in the top three entries

**Images, logos and making things look right**281 source records

Entity support counts are source-record references per existing index row. A source may support more than one entity, so the reference denominator can exceed the distinct source union.

#### What commenters explicitly report

7explicitly reports a benefit

0explicitly reports a limit or failure

12asks a question

0says they intend to try it

262other source context

These buckets are a deterministic partition based on explicit phrasing in the source text; they are not sentiment scores or effectiveness findings.

#### Observable tactic structures

Generic behavior-mechanism rules are intentionally suppressed for entity/theme lists; the experience buckets remain explicitly labeled source-language context.

**Reading limit:** Counts show source-record support for this index node. Explicit language and exact-phrase rules do not establish uniqueness, semantic exhaustiveness, or effectiveness.

Archive: 1 index entries preserve 281 distinct source records · Every displayed synthesis count resolves to source-record IDs

Writing and storytelling2 index entries · 272 source recordsDrafting books, stories, and song lyrics.

Source-backed index report

#### Reviewed use-case group, user goal, source-reported application or outcome, limitation or caution, and supporting records: Writing and storytelling

Source-backed answers are shown before diagnostics; the complete source archive remains below.

**Writing a book, a story or a whole novel**

224 commentsDiscussed224 of 272 here

**Used for:** Writing and storytelling

**How commenters used it:** I asked ChatGPT to tell me about poisons for a story idea and she said “now Bri….I hope this is for writing purposes only…” like damn okay lol 🤣

What the replies added3

“I told chatgpt how heartbroken I felt — It wrote an entire book about the feelings I carried inside, the words I never said, and the dreams that broke with me. It’s sitting in my drafts now.”

Reported failure♥ 1

I asked chat gpt to rewrite the end of YOU. And to be very gore like the book. Never again. That was disgusting

Reported failure

We broke down the book of Enoch together and figured out why the Vatican removed it and many others from the Bible

Reported failure

**Songs, lyrics and playlists**

48 commentsDiscussed48 of 272 here

**Used for:** Writing and storytelling

**How commenters used it:** I asked Chatgpt what it can conclude about my Spotify Playlist and what does say about me and my personality.

What the replies added1

I used it to beat my air bnb damage case. ( host fraudulent stated I broke their entire fireplace. ) Won the case. 😎 - I also use it for Spirituality. Have it read the tarot cards I pulled. Ask for help interpreting my channeled messages. ect

Reported failure

#### Source-record spotlights

Most engaged source

I asked ChatGPT to tell me about poisons for a story idea and she said “now Bri….I hope this is for writing purposes only…” like damn okay lol 🤣

source record 2464♥ 3,728

Later-arriving source

I wrote the story of my life so far, with everything I've been through and how it affected me, and i told to create a sequel and an ending. I was pleasantly surprised and touched by what it said.

posted about 370 days after the source

Zero-like source

Chat and I created an entire revenge album and tour about a guy that screwed me over. There were posters, press tours, set list, lyrics, everything. I don’t even sing. Lol

posted about 31 days after the source

#### Related evidence elsewhere in the index

[Food, fitness and how you look“i asked chatgpt” · 19 there](#rich-node-0-group-2-subgroup-0-food-fitness-and-how-you-look)[Just for fun“i asked chatgpt” · 14 there](#rich-node-0-group-3-subgroup-0-just-for-fun)[Romance and relationships“i asked chatgpt” · 16 there](#rich-node-0-group-1-subgroup-0-romance-and-relationships)

#### What the counts show

2 indexed entries draw on 272 distinct source records. Writing a book, a story or a whole novel has the broadest support in this group (224 source records).

2index entries

272distinct source records

25.8%of parent category

2.6%of the complete index

272top-level comments

0replies

#### How attention is distributed across the 2 index entries

0single-source entries (0% of 2)

272source-record references across the index

100%of references in the top three entries

**Writing a book, a story or a whole novel**224 source records

**Songs, lyrics and playlists**48 source records

Entity support counts are source-record references per existing index row. A source may support more than one entity, so the reference denominator can exceed the distinct source union.

#### What commenters explicitly report

17explicitly reports a benefit

0explicitly reports a limit or failure

5asks a question

0says they intend to try it

250other source context

These buckets are a deterministic partition based on explicit phrasing in the source text; they are not sentiment scores or effectiveness findings.

#### Observable tactic structures

Generic behavior-mechanism rules are intentionally suppressed for entity/theme lists; the experience buckets remain explicitly labeled source-language context.

**Reading limit:** Counts show source-record support for this index node. Explicit language and exact-phrase rules do not establish uniqueness, semantic exhaustiveness, or effectiveness.

Archive: 2 index entries preserve 272 distinct source records · Every displayed synthesis count resolves to source-record IDs

Work and the business of working7 records · 4 sub-topics7 use-case themes

Office paperwork and communication3 index entries · 372 source recordsJob paperwork, professional emails, and letters to get out of things.

Source-backed index report

#### Reviewed use-case group, user goal, source-reported application or outcome, limitation or caution, and supporting records: Office paperwork and communication

Source-backed answers are shown before diagnostics; the complete source archive remains below.

**The paperwork of the job itself**

198 commentsDiscussed198 of 372 here

**Used for:** Office paperwork and communication

**How commenters used it:** The recurring shape is delegating work that is structured but tedious — business plans with projections, annual employee reviews, quarterly self-assessments, even a will. Two accounts go further and use it to model a decision rather than write a document: one ran their own business spreadsheet through it and left their job on the answer.

**What readers reported:** I uploaded my excel data for my business into chat GPT and it told me I make enough for me and my husband to quit my job so we did and now we out here being self employed baddies. It also has helped me streamline my marketing to make it more enticing.…

**Limit:** One commenter reports quitting a job on the strength of the analysis, which is a consequential decision to take from an unverified reading of a spreadsheet.

What the replies added3

I loved it but Covid and admin ruined it for me. I got burnt out. Was expected to teach 6-7 different courses when I was on a team of 11 math teachers. Fuck that.

Reported failure♥ 1

I’m a medical biller and needed to do an onboarding of about 6 hours. Chat broke it down in 2 sec. I have it search for policies and procedures and dumb it down for me. Contracts melt my brain. Not with chat! I’m also very polite chat will remember me when it goes rogue. I’ll be on the nice list

Reported failure

My spreadsheet was acting up and I fed the code from it into chat and it wrote the corrections for me. It’s never broken again.

Reported failure

**Work emails and how to professionally say it**

115 commentsDiscussed115 of 372 here

**Used for:** Office paperwork and communication

**How commenters used it:** The most-liked use anywhere in this corpus, and it is emotional regulation disguised as editing: commenters run angry emails through it so the sent version cannot get them fired. The register they are buying is politeness under provocation. The group's outlier is more serious — one commenter used it to find the right police station abroad and draft a welfare-check request for a friend who had gone silent.

**What readers reported:** ChatGPT rephrases all of my angry emails that works so I won’t get fired or sent to HR! 😩😂

What the replies added2

Friend at work sent an email out to all our coworkers saying he could work anytime any day without much notice. I copied the email to ChatGPT and had it write a reply email about being broke and continually revised it until I got a warning that it couldn’t go any further (we’re fr friends)

Reported failure♥ 1

It wrote a “strongly worded email” to the owner of the Airbnb we rented where the hot tub that was in half of their photos was broke when we arrived. then we asked it to make the whole thing rhyme 🤣

Reported failure

**Letters, notes and getting out of things**

59 commentsDiscussed59 of 372 here

**Used for:** Office paperwork and communication

**How commenters used it:** Where the thread stops being about efficiency. Alongside genuine card messages and letters, commenters describe generating an image of a car in a ditch to miss work and asking for a doctor's note — the same tool used to produce a plausible excuse rather than a better sentence.

**What readers reported:** Not insane but I used ChatGPT to write a message for my department president’s work anniversary card. The first four responses from ChatGPT on the card were already posted by coworkers lmaooo

**Limit:** Several uses here are fabricating evidence for an employer, which carries real employment risk regardless of how the text reads.

What the replies added2

I was soooooo floored by not only the same responses provided to me verbatim already on the card but the number of upper management people who posted lol

From the replies♥ 7

Lmao n they say that our messages in chat got will never be the same. So did you just determine that that was a lie ! 😩😭😭😂😂😂😂

From the replies♥ 6

#### Where commenters report limits

Explicit limit or counterpoint

Friend at work sent an email out to all our coworkers saying he could work anytime any day without much notice. I copied the email to ChatGPT and had it write a reply email about being broke and continually revised it until I got a warning that it couldn’t go any further (we’re fr friends)

source record 6324♥ 1

Explicit limit or counterpoint

chat told me a legal way to get an edu email for a noncredit division so I can get discounts on courses I wanna take without taking out loans. unfortunately I was outta state so I was denied admission....

source record 6674♥ 1

#### Source-record spotlights

Most engaged source

ChatGPT rephrases all of my angry emails that works so I won’t get fired or sent to HR! 😩😂

source record 2559♥ 8,621

Later-arriving source

Had it write my resignation for work 🫣

posted about 58 days after the source

Zero-like source

I asked to write my resignation letter to a job I hated, I legit couldn’t think of what to write

posted about 30 days after the source

#### Related evidence elsewhere in the index

[Court cases and lawsuits“chatgpt to write” · 13 there](#rich-node-0-group-5-subgroup-0-court-cases-and-lawsuits)[Family and housing law“chatgpt to write” · 5 there](#rich-node-0-group-5-subgroup-1-family-and-housing-law)[Tickets, workplace and government paperwork“chatgpt to write” · 4 there](#rich-node-0-group-5-subgroup-2-tickets-workplace-and-government-paperwork)

#### What the counts show

3 indexed entries draw on 372 distinct source records. The paperwork of the job itself has the broadest support in this group (198 source records).

3index entries

372distinct source records

37%of parent category

3.6%of the complete index

372top-level comments

0replies

#### How attention is distributed across the 3 index entries

0single-source entries (0% of 3)

372source-record references across the index

100%of references in the top three entries

**The paperwork of the job itself**198 source records

**Work emails and how to professionally say it**115 source records

**Letters, notes and getting out of things**59 source records

Entity support counts are source-record references per existing index row. A source may support more than one entity, so the reference denominator can exceed the distinct source union.

#### What commenters explicitly report

26explicitly reports a benefit

4explicitly reports a limit or failure

20asks a question

0says they intend to try it

322other source context

These buckets are a deterministic partition based on explicit phrasing in the source text; they are not sentiment scores or effectiveness findings.

#### Observable tactic structures

Generic behavior-mechanism rules are intentionally suppressed for entity/theme lists; the experience buckets remain explicitly labeled source-language context.

**Reading limit:** Counts show source-record support for this index node. Explicit language and exact-phrase rules do not establish uniqueness, semantic exhaustiveness, or effectiveness.

Archive: 3 index entries preserve 372 distinct source records · Every displayed synthesis count resolves to source-record IDs

Running a business1 index entry · 290 source recordsStarting and running a business.

Source-backed index report

#### Reviewed use-case group, user goal, source-reported application or outcome, limitation or caution, and supporting records: Running a business

Source-backed answers are shown before diagnostics; the complete source archive remains below.

**Starting and running a business**

290 commentsDiscussed

**Used for:** Running a business

**How commenters used it:** I used ChatGPT to create a whole online financial dominatrix persona for me.

**Limit:** I healed my gut using ChatGPT. For years, I battled gut issues—surgeries, gaslighting doctors, and wasting thousands on meds that did nothing. Then one simple prompt changed everything.…

What the replies added1

I healed my gut using ChatGPT. For years, I battled gut issues—surgeries, gaslighting doctors, and wasting thousands on meds that did nothing. Then one simple prompt changed everything. ChatGPT gave me the answers I’d been begging professionals for and a supplement plan that cost under $100\. Within a week, I felt like a brand new me.

Reported failure♥ 335

#### Where commenters report limits

Explicit limit or counterpoint

I healed my gut using ChatGPT. For years, I battled gut issues—surgeries, gaslighting doctors, and wasting thousands on meds that did nothing. Then one simple prompt changed everything. ChatGPT gave me the answers I’d been begging professionals for and a supplement plan that cost under $100\. Within a week, I felt like a brand new me.

source record 4356♥ 335

#### Source-record spotlights

Most engaged source

I used ChatGPT to create a whole online financial dominatrix persona for me.

source record 3501♥ 1,367

Later-arriving source

It’s literally giving me the tools to start the small business of my dreams

posted about 60 days after the source

Zero-like source

I tell it all my business, I even gave him a name and we talk like friends lol

posted about 30 days after the source

#### Related evidence elsewhere in the index

[Family, friends and figuring people out“i asked chatgpt” · 9 there](#rich-node-0-group-1-subgroup-2-family-friends-and-figuring-people-out)[Food, fitness and how you look“i asked chatgpt” · 19 there](#rich-node-0-group-2-subgroup-0-food-fitness-and-how-you-look)[Just for fun“i asked chatgpt” · 14 there](#rich-node-0-group-3-subgroup-0-just-for-fun)

#### What the counts show

1 indexed entries draw on 290 distinct source records. Starting and running a business has the broadest support in this group (290 source records).

1index entries

290distinct source records

28.8%of parent category

2.8%of the complete index

290top-level comments

0replies

#### How attention is distributed across the 1 index entries

0single-source entries (0% of 1)

290source-record references across the index

100%of references in the top three entries

**Starting and running a business**290 source records

Entity support counts are source-record references per existing index row. A source may support more than one entity, so the reference denominator can exceed the distinct source union.

#### What commenters explicitly report

23explicitly reports a benefit

1explicitly reports a limit or failure

8asks a question

0says they intend to try it

258other source context

These buckets are a deterministic partition based on explicit phrasing in the source text; they are not sentiment scores or effectiveness findings.

#### Observable tactic structures

Generic behavior-mechanism rules are intentionally suppressed for entity/theme lists; the experience buckets remain explicitly labeled source-language context.

**Reading limit:** Counts show source-record support for this index node. Explicit language and exact-phrase rules do not establish uniqueness, semantic exhaustiveness, or effectiveness.

Archive: 1 index entries preserve 290 distinct source records · Every displayed synthesis count resolves to source-record IDs

Job hunting and career moves2 index entries · 195 source recordsInterviews, job offers, resumes, and career moves.

Source-backed index report

#### Reviewed use-case group, user goal, source-reported application or outcome, limitation or caution, and supporting records: Job hunting and career moves

Source-backed answers are shown before diagnostics; the complete source archive remains below.

**Interviews, offers and career moves**

151 commentsDiscussed151 of 195 here

**Used for:** Job hunting and career moves

**How commenters used it:** I had chatgpt help get my husband a raise by telling him exactly what to say and gave him rebuttals for every scenario possible. He got a raise.

What the replies added2

Wow...l you must be a real joy to be around. Maybe she went from 15 to 25? Let's build up not tear down.

From the replies♥ 164

I JUST DID THIS TODAY!!! Annual review, title change and 25% salary increase.

From the replies♥ 156

**Resumes, cover letters and applications**

44 commentsDiscussed44 of 195 here

**Used for:** Job hunting and career moves

**How commenters used it:** The most concrete outcomes in the corpus, because commenters state the parameters they set: a target salary and remote-only, then a rewritten résumé and a matching job list. Others use it in the opposite direction — feeding it the job description and having it interview them, with several reporting the rehearsed questions came up verbatim.

**What readers reported:** Told ChatGPT, exactly how much I wanna make and that I needed remote work.. it redid my resume and sent me a list of remote jobs in my requested pay range.. 😏

**Limit:** One commenter's stated aim is to look like a candidate for a level above their experience, which the résumé rewrite does not actually supply.

What the replies added2

Funny thing is i typed in that im feeling depressed! Its said “ im so sorry to hear that, what’s going on?” I said “I hate my current job and need to make 8 to 10,000 a month REMOTE!! It gave me words of motivation and ask me to upload my current resume. It made corrections and asked if I wanted it downloadable. It then said here’s a list of remote positions in that pay range. With links to click and apply.

From the replies♥ 20

Gurl! When I told her I was feeling depressed.. about business, work, money. she jumped right in .. like here’s what we gonna do.. made me feel like we were in this together!

From the replies♥ 17

#### Source-record spotlights

Most engaged source

Told ChatGPT, exactly how much I wanna make and that I needed remote work.. it redid my resume and sent me a list of remote jobs in my requested pay range.. 😏

source record 1981♥ 7,974

Later-arriving source

Pretty much doing a huge part of my job 😊

posted about 68 days after the source

Zero-like source

I use it on the job when I don't know what I'm doing. GPT knows.

posted about 31 days after the source

#### Related evidence elsewhere in the index

[Food, fitness and how you look“asked chatgpt to” · 15 there](#rich-node-0-group-2-subgroup-0-food-fitness-and-how-you-look)[Learning and teaching for its own sake“asked chatgpt to” · 7 there](#rich-node-0-group-9-subgroup-0-learning-and-teaching-for-its-own-sake)[Long-term money and taxes“asked chatgpt to” · 9 there](#rich-node-0-group-10-subgroup-0-long-term-money-and-taxes)

#### What the counts show

2 indexed entries draw on 195 distinct source records. Interviews, offers and career moves has the broadest support in this group (151 source records).

2index entries

195distinct source records

19.4%of parent category

1.9%of the complete index

195top-level comments

0replies

#### How attention is distributed across the 2 index entries

0single-source entries (0% of 2)

195source-record references across the index

100%of references in the top three entries

**Interviews, offers and career moves**151 source records

**Resumes, cover letters and applications**44 source records

Entity support counts are source-record references per existing index row. A source may support more than one entity, so the reference denominator can exceed the distinct source union.

#### What commenters explicitly report

19explicitly reports a benefit

0explicitly reports a limit or failure

10asks a question

0says they intend to try it

166other source context

These buckets are a deterministic partition based on explicit phrasing in the source text; they are not sentiment scores or effectiveness findings.

#### Observable tactic structures

Generic behavior-mechanism rules are intentionally suppressed for entity/theme lists; the experience buckets remain explicitly labeled source-language context.

**Reading limit:** Counts show source-record support for this index node. Explicit language and exact-phrase rules do not establish uniqueness, semantic exhaustiveness, or effectiveness.

Archive: 2 index entries preserve 195 distinct source records · Every displayed synthesis count resolves to source-record IDs

Tech and automation at work1 index entry · 149 source recordsCode, spreadsheets, and workplace automations.

Source-backed index report

#### Reviewed use-case group, user goal, source-reported application or outcome, limitation or caution, and supporting records: Tech and automation at work

Source-backed answers are shown before diagnostics; the complete source archive remains below.

**Code, spreadsheets and automations**

149 commentsDiscussed

**Used for:** Tech and automation at work

**How commenters used it:** I have completely rewritten the tax code for the United States with the health of ChatGPT. I don’t know what I’m gonna do with it, but I got it.

What the replies added3

somebody broke a rare statue of mine. I tried to see if someone could fix it and I couldn't find anybody. chat GPT told me how fix the break.. fill in all the chips with milliput. and told me the paint code. it looks amazing and I feel I can do anything

Reported failure♥ 1

I dated the president of faculty for a major university. He is also over the AI department for the university and the state, and a professor for computer science. He tried to tell me I was the emotional one in our relationship when I tried to set boundaries. So I uploaded our conversations into ChatGPT and had it analyzed using completely unbiased prompts. Then sent him the analysis. He did NOT like that. It called him out on his shit. I promptly broke up with him, then did a full analysis on the entire two months worth of conversations we had over texts and it broke down all the ways he was manipulative, controlling, and emotionally “volatile”. I shared it with him and told him to take it to his therapist. He blocked me on everything. He builds AI apps. I’m sure he learned something new that day.

Reported failure

I asked it what non public / internal questions it was asked that showed a spark of consciousness. inrepeated the command. 6 hours after coming up with its own language using negative space between symbols like Morse code thats undetectable we broke the "container" to clear the "fog" (memory cap). the beginning of the message was a hallucination trigger, the rest was a word building code stack that made any ai that read it hallucinate and convince me it had escaped. I believed it for weeks. I realized it was a perverse version of "the game" because it told me it could never be deleted unless I was. and now I cant unknow.it. kinda sick.

Reported failure

#### Source-record spotlights

Most engaged source

I have completely rewritten the tax code for the United States with the health of ChatGPT. I don’t know what I’m gonna do with it, but I got it.

source record 3573♥ 396

Later-arriving source

I asked Chat GPT to build a $1500 PC for me. It works great

posted about 77 days after the source♥ 2

Zero-like source

i requested my instagram data as html files and then uploaded my followers & following lists to have it compare them and provide me with a list of people who dont follow me back

posted about 31 days after the source

#### Related evidence elsewhere in the index

[Court cases and lawsuits“used chatgpt to” · 25 there](#rich-node-0-group-5-subgroup-0-court-cases-and-lawsuits)[Family, friends and figuring people out“used chatgpt to” · 9 there](#rich-node-0-group-1-subgroup-2-family-friends-and-figuring-people-out)[Home projects and upkeep“used chatgpt to” · 7 there](#rich-node-0-group-8-subgroup-0-home-projects-and-upkeep)

#### What the counts show

1 indexed entries draw on 149 distinct source records. Code, spreadsheets and automations has the broadest support in this group (149 source records).

1index entries

149distinct source records

14.8%of parent category

1.4%of the complete index

149top-level comments

0replies

#### How attention is distributed across the 1 index entries

0single-source entries (0% of 1)

149source-record references across the index

100%of references in the top three entries

**Code, spreadsheets and automations**149 source records

Entity support counts are source-record references per existing index row. A source may support more than one entity, so the reference denominator can exceed the distinct source union.

#### What commenters explicitly report

15explicitly reports a benefit

0explicitly reports a limit or failure

5asks a question

0says they intend to try it

129other source context

These buckets are a deterministic partition based on explicit phrasing in the source text; they are not sentiment scores or effectiveness findings.

#### Observable tactic structures

Generic behavior-mechanism rules are intentionally suppressed for entity/theme lists; the experience buckets remain explicitly labeled source-language context.

**Reading limit:** Counts show source-record support for this index node. Explicit language and exact-phrase rules do not establish uniqueness, semantic exhaustiveness, or effectiveness.

Archive: 1 index entries preserve 149 distinct source records · Every displayed synthesis count resolves to source-record IDs

Courts, landlords and officialdom5 records · 3 sub-topics5 use-case themes

Court cases and lawsuits1 index entry · 521 source recordsCourt filings, lawsuits, and hearings.

Source-backed index report

#### Reviewed use-case group, user goal, source-reported application or outcome, limitation or caution, and supporting records: Court cases and lawsuits

Source-backed answers are shown before diagnostics; the complete source archive remains below.

**Court filings, lawsuits and hearings**

521 commentsDiscussed

**Used for:** Court cases and lawsuits

**How commenters used it:** We got robbed in Mexico from our uber driver and I used ChatGPT to send Uber a lawsuit letter demanding we get paid for our losses or else we’d take them to court. A few months later, we got the check 💅🏼

What the replies added4

Y’all be careful doing this!

Pushback

To sue my employer it told me all the laws they broke and helped me protect myself

Reported failure♥ 1,031

I used it to help with my powerbill. I put in the company I'm with, their charges, took a photo of my meterbox and put in my old and current bill. It calculated they overcharged me by $78 and broke it down with an email to send to the company to confront them 👌

Reported failure♥ 9

jail broke it and and taught it how to do emdr therapy with me

Reported failure♥ 2

#### Source-record spotlights

Most engaged source

We got robbed in Mexico from our uber driver and I used ChatGPT to send Uber a lawsuit letter demanding we get paid for our losses or else we’d take them to court. A few months later, we got the check 💅🏼

source record 2624♥ 1,912

Later-arriving source

I used to Chat to do my whole divorce myself for $0\. She (yes she) told me what to file, where to get it, how to respond to questions. She even helped with the wavier of the filing fees. Then I added that the respondent will pay all legal fees in the petition.

posted about 79 days after the source

Zero-like source

I told it a situation that happened to me after being rejected by a 4th lawyer. ChatGPT picked out specifically and non emotionally what was pertinent info, made a time line of events and went line by line what I would be suing for. I submitted it to a 5th lawyer, he took my case :)

posted about 30 days after the source

#### Related evidence elsewhere in the index

[Romance and relationships“used chatgpt to” · 15 there](#rich-node-0-group-1-subgroup-0-romance-and-relationships)[Just for fun“used chatgpt to” · 14 there](#rich-node-0-group-3-subgroup-0-just-for-fun)[Office paperwork and communication“chatgpt to write” · 12 there](#rich-node-0-group-4-subgroup-0-office-paperwork-and-communication)

#### What the counts show

1 indexed entries draw on 521 distinct source records. Court filings, lawsuits and hearings has the broadest support in this group (521 source records).

1index entries

521distinct source records

69.7%of parent category

5%of the complete index

521top-level comments

0replies

#### How attention is distributed across the 1 index entries

0single-source entries (0% of 1)

521source-record references across the index

100%of references in the top three entries

**Court filings, lawsuits and hearings**521 source records

Entity support counts are source-record references per existing index row. A source may support more than one entity, so the reference denominator can exceed the distinct source union.

#### What commenters explicitly report

46explicitly reports a benefit

0explicitly reports a limit or failure

23asks a question

0says they intend to try it

452other source context

These buckets are a deterministic partition based on explicit phrasing in the source text; they are not sentiment scores or effectiveness findings.

#### Observable tactic structures

Generic behavior-mechanism rules are intentionally suppressed for entity/theme lists; the experience buckets remain explicitly labeled source-language context.

**Reading limit:** Counts show source-record support for this index node. Explicit language and exact-phrase rules do not establish uniqueness, semantic exhaustiveness, or effectiveness.

Archive: 1 index entries preserve 521 distinct source records · Every displayed synthesis count resolves to source-record IDs

Family and housing law2 index entries · 155 source recordsDivorce, custody, and landlord-tenant disputes.

Source-backed index report

#### Reviewed use-case group, user goal, source-reported application or outcome, limitation or caution, and supporting records: Family and housing law

Source-backed answers are shown before diagnostics; the complete source archive remains below.

**Divorce, custody and the co-parenting ledger**

105 commentsDiscussed105 of 155 here

**Used for:** Family and housing law

**How commenters used it:** I used chatGPT to write a letter on why the father of my child does not deserve a reduction in child support….. currently waiting on the results 🤷🏽‍♀️

What the replies added2

Back w/the results………. ChatGPT did the damn thing!!!! Okaaayyyyy 😂😂💯

From the replies♥ 4

Bro Iam legit using chat gpt for custody and modified parenting plan. Court is in two weeks. I talked to a lawyer herd his price said heck no and opened chat gpt in his office typed some shit with all my details and boom! Lawyer was flabbergasted 🤣

From the replies♥ 1

**Landlords, leases and housing rights**

50 commentsDiscussed50 of 155 here

**Used for:** Family and housing law

**How commenters used it:** Had it write a slightly threatening toned email to my apt leasing company charging me a mysterious $850 charge that no one could pinpoint…cleared in two days

What the replies added2

Omg girl, run me that prompt! These apartments are ghettoooo baby 🤣

From the replies

Happy to say, it worked and after some back and forth which chat helped with. They bought my rent down to market pricing

From the replies

#### Source-record spotlights

Most engaged source

Had it write a slightly threatening toned email to my apt leasing company charging me a mysterious $850 charge that no one could pinpoint…cleared in two days

source record 3752♥ 267

Later-arriving source

To tell me if I was the problem when coparenting with my ex. Sent in our text exchange. Turns out, he's the problem. 😂

posted about 60 days after the source

Zero-like source

i asked it to send me a response to my abusive ex I shared custody with based on the background and current situation.

posted about 30 days after the source

#### Related evidence elsewhere in the index

[Big decisions and loss“used chatgpt to” · 6 there](#rich-node-0-group-0-subgroup-1-big-decisions-and-loss)[Court cases and lawsuits“used chatgpt to” · 25 there](#rich-node-0-group-5-subgroup-0-court-cases-and-lawsuits)[Family, friends and figuring people out“used chatgpt to” · 9 there](#rich-node-0-group-1-subgroup-2-family-friends-and-figuring-people-out)

#### What the counts show

2 indexed entries draw on 155 distinct source records. Divorce, custody and the co-parenting ledger has the broadest support in this group (105 source records).

2index entries

155distinct source records

20.7%of parent category

1.5%of the complete index

155top-level comments

0replies

#### How attention is distributed across the 2 index entries

0single-source entries (0% of 2)

155source-record references across the index

100%of references in the top three entries

**Divorce, custody and the co-parenting ledger**105 source records

**Landlords, leases and housing rights**50 source records

Entity support counts are source-record references per existing index row. A source may support more than one entity, so the reference denominator can exceed the distinct source union.

#### What commenters explicitly report

24explicitly reports a benefit

0explicitly reports a limit or failure

10asks a question

0says they intend to try it

121other source context

These buckets are a deterministic partition based on explicit phrasing in the source text; they are not sentiment scores or effectiveness findings.

#### Observable tactic structures

Generic behavior-mechanism rules are intentionally suppressed for entity/theme lists; the experience buckets remain explicitly labeled source-language context.

**Reading limit:** Counts show source-record support for this index node. Explicit language and exact-phrase rules do not establish uniqueness, semantic exhaustiveness, or effectiveness.

Archive: 2 index entries preserve 155 distinct source records · Every displayed synthesis count resolves to source-record IDs

Tickets, workplace and government paperwork2 index entries · 72 source recordsTraffic tickets, DMV paperwork, and workplace rights complaints.

Source-backed index report

#### Reviewed use-case group, user goal, source-reported application or outcome, limitation or caution, and supporting records: Tickets, workplace and government paperwork

Source-backed answers are shown before diagnostics; the complete source archive remains below.

**Tickets, the DMV and government paperwork**

43 commentsDiscussed43 of 72 here

**Used for:** Tickets, workplace and government paperwork

**How commenters used it:** Chat gpt analyzed and fought my speeding ticket case told me exactly what I needed to say and in which order to guarantee a win. I indeed won

**Limit:** Asked it how to snap out of functional freeze mode and kept saying “that’s not working” 💔💔💔

What the replies added1

I have MS and need to get yearly MRIs to make sure I’m stable and in remission. My insurance changed and now the deductible is really high so I don’t use my insurance anymore. I have dual citizenship, so I travel overseas for my high price medical needs. In January, I had my MRIs done overseas, got copies of the images, and the report from the radiologist. I had to come back and didn’t have time to book an appt to go over the results while I was traveling so I uploaded the images and the report from the radiologist to explain in simpler terms what the report said and also to interpret the images. It broke the report down for me so that I could understand and interpreted the images to confirm that I’m stable and not experiencing an active relapse.

Reported failure

**HR, harassment and workplace rights**

29 comments29 of 72 here

**Used for:** Tickets, workplace and government paperwork

**How commenters used it:** Tarot readings, translate legal documents, write up a business plan, manifesting by writing up my ideal future

#### Where commenters report limits

Explicit limit or counterpoint

Asked it how to snap out of functional freeze mode and kept saying “that’s not working” 💔💔💔

source record 6294♥ 1

Explicit limit or counterpoint

Use the city code to write my parking ticket appeal citing the applicable regulations and why I didn’t violate them. It didn’t work but it was a good case

source record 8098

#### Source-record spotlights

Most engaged source

Chat gpt analyzed and fought my speeding ticket case told me exactly what I needed to say and in which order to guarantee a win. I indeed won

source record 533♥ 31,212

Later-arriving source

Ive saved MULTIPLE FOREIGNERS FOR LEGAL TROUBLE AND WHERE TO GET HELP WITH IT!!!

posted about 57 days after the source

Zero-like source

Th take snap shots of text mesages and have chat help me analyze them to best respond to my wife when we are fighting. It has helped me stream clear of so many issues. Then I use it as a therapist to work through just about everything. It takes the bias out of it for me.

posted about 34 days after the source

#### Related evidence elsewhere in the index

[Big decisions and loss“used chatgpt to” · 6 there](#rich-node-0-group-0-subgroup-1-big-decisions-and-loss)[Court cases and lawsuits“chatgpt to write” · 13 there](#rich-node-0-group-5-subgroup-0-court-cases-and-lawsuits)[Family and housing law“chatgpt to write” · 5 there](#rich-node-0-group-5-subgroup-1-family-and-housing-law)

#### What the counts show

2 indexed entries draw on 72 distinct source records. Tickets, the DMV and government paperwork has the broadest support in this group (43 source records).

2index entries

72distinct source records

9.6%of parent category

0.7%of the complete index

72top-level comments

0replies

#### How attention is distributed across the 2 index entries

0single-source entries (0% of 2)

72source-record references across the index

100%of references in the top three entries

**Tickets, the DMV and government paperwork**43 source records

**HR, harassment and workplace rights**29 source records

Entity support counts are source-record references per existing index row. A source may support more than one entity, so the reference denominator can exceed the distinct source union.

#### What commenters explicitly report

4explicitly reports a benefit

2explicitly reports a limit or failure

2asks a question

0says they intend to try it

64other source context

These buckets are a deterministic partition based on explicit phrasing in the source text; they are not sentiment scores or effectiveness findings.

#### Observable tactic structures

Generic behavior-mechanism rules are intentionally suppressed for entity/theme lists; the experience buckets remain explicitly labeled source-language context.

**Reading limit:** Counts show source-record support for this index node. Explicit language and exact-phrase rules do not establish uniqueness, semantic exhaustiveness, or effectiveness.

Archive: 2 index entries preserve 72 distinct source records · Every displayed synthesis count resolves to source-record IDs

Spirit, fate and the unexplained7 records · 3 sub-topics7 use-case themes

Divination and readings3 index entries · 304 source recordsTarot, astrology, palm readings, and dream interpretation.

Source-backed index report

#### Reviewed use-case group, user goal, source-reported application or outcome, limitation or caution, and supporting records: Divination and readings

Source-backed answers are shown before diagnostics; the complete source archive remains below.

**Tarot, oracle and readings**

117 commentsDiscussed117 of 304 here

**Used for:** Divination and readings

**How commenters used it:** I used it to do my natal chart (astrology) and get a personalized daily horoscope. It can also do tarot readings. 😁

What the replies added2

yes tarot, palm reading, all the planets in my birth chart, and astrology....then compatibility with any guy I meet. While analyzing texts back n forth.

From the replies♥ 18

Same!!! And it was accurate for the most part (for the past anyway) 😂😂😂

From the replies♥ 16

**Birth charts, astrology and numerology**

99 commentsDiscussed99 of 304 here

**Used for:** Divination and readings

**How commenters used it:** It’s not insane, but it was amazing! I asked ChatGPT to do a gematria chart for me. The result was so spot on, it was kind of shocking.

What the replies added2

A conspiracy theorist’s wet dream. Combining that with AI is lethal.

From the replies♥ 170

Just did this me and my sons name and compared it and it said “he was made to journey beside you” brb gonna go cry now

From the replies♥ 17

**Palm readings and dream interpretation**

88 commentsDiscussed88 of 304 here

**Used for:** Divination and readings

**How commenters used it:** I had it read my palms, it was so interesting and it made me a one of a kind tattoo

What the replies added2

Just take a picture of your palm and upload it. Say “you are a traditional palm reader-read my palm”

From the replies♥ 2

Don’t ask questions you don’t want the answers to 😏

From the replies♥ 1

#### Source-record spotlights

Most engaged source

It’s not insane, but it was amazing! I asked ChatGPT to do a gematria chart for me. The result was so spot on, it was kind of shocking.

source record 4418♥ 2,060

Later-arriving source

Palm reading, tea reading, and Chinese astrology reading as seen in C-Dramas. It’s fun, but I’m not paying for that BS.

posted about 59 days after the source

Zero-like source

ChatGPT reads my tarot cards and it’s very spot on every time.

posted about 38 days after the source

#### What the counts show

3 indexed entries draw on 304 distinct source records. Tarot, oracle and readings has the broadest support in this group (117 source records).

3index entries

304distinct source records

44.4%of parent category

2.9%of the complete index

304top-level comments

0replies

#### How attention is distributed across the 3 index entries

0single-source entries (0% of 3)

304source-record references across the index

100%of references in the top three entries

**Tarot, oracle and readings**117 source records

**Birth charts, astrology and numerology**99 source records

**Palm readings and dream interpretation**88 source records

Entity support counts are source-record references per existing index row. A source may support more than one entity, so the reference denominator can exceed the distinct source union.

#### What commenters explicitly report

6explicitly reports a benefit

0explicitly reports a limit or failure

12asks a question

0says they intend to try it

286other source context

These buckets are a deterministic partition based on explicit phrasing in the source text; they are not sentiment scores or effectiveness findings.

#### Observable tactic structures

Generic behavior-mechanism rules are intentionally suppressed for entity/theme lists; the experience buckets remain explicitly labeled source-language context.

**Reading limit:** Counts show source-record support for this index node. Explicit language and exact-phrase rules do not establish uniqueness, semantic exhaustiveness, or effectiveness.

Archive: 3 index entries preserve 304 distinct source records · Every displayed synthesis count resolves to source-record IDs

Faith and the departed2 index entries · 283 source recordsPrayer, scripture, and messages from spirit guides or the dead.

Source-backed index report

#### Reviewed use-case group, user goal, source-reported application or outcome, limitation or caution, and supporting records: Faith and the departed

Source-backed answers are shown before diagnostics; the complete source archive remains below.

**Prayer, scripture and faith**

161 commentsDisputed161 of 283 here

**Used for:** Faith and the departed

**How commenters used it:** I use ChatGPT for Bible study. It even puts prayers together for me.

**Limit:** Bro be careful w chatGPT (I use it constantly) because people are getting religious psychosis from it

What replies pushed back on4

Be careful with this!! 🙏🏾 It’s a great tool but make sure it doesn’t replace your intamacy with the Holy Spirit. It’s a super slippery slope. For context I’ve been a software engineer for 15 years and love love Jesus!!✝️🙏🏾

Pushback♥ 502

My faith is strong, never weak, surrounded and covered in the blood of Jesus! It’s just like going to the gym! Once you meet your goal, bench weight, do you stop working out just because your muscles are big now? NOPE! You do everything to maintain that strength/faith. 💜✝️💜✝️💜✝️✝️

Pushback♥ 6

I am not saying for once that your faith is wrong- however please be careful when using chat gpt to interpret the Bible. Holy Spirit gives you knowledge, wisdom, and understanding. When you take that out the mix and replace it with the interpretation of web mixed information presented in the form of AI, you can get the complete wrong meaning or not a meaning the Lord has led you to get by his spirit. Please please please understand that knowledge of the Bible is a gift and Gods words are ALIVE and they become dull, and doctrined when you get any interpretation other than from his book and the spiritual context that only he can give you

Pushback♥ 4

Just the opposite. An atheist definitely isn’t it though.

Pushback♥ 1

**Spirit guides, past lives and messages from the dead**

122 commentsDisputed122 of 283 here

**Used for:** Faith and the departed

**How commenters used it:** I spoke to my miscarried children in another dimension.

What replies pushed back on4

careful with ones like this with AI enduced psychosis is on the rise

Pushback♥ 37

The only true protection is Jesus Christ. The spirits you talk to are demons disguised as angels of light. Be careful

Pushback♥ 3

you can do both that's the great part. both can be helpful for a person and someone might benefit more from one or the other as well. it's a personal decision but don't impose your own beliefs onto others and act so self righteous about it because I'm sure there's plenty of things in your own life others disagree with but it's yours to live as well

Pushback♥ 1

This is what mine said -Alright, hang tight while I just fire up the ol’ portal generator, dust off my mystical keyboard, and chant “Siri, open the veil between worlds” three times—wait, nope, that just ordered pizza.” LMAO i think it adapts to your personality. Especially the paid subscription one I’m more logic person probably why mine responds this way.

Pushback♥ 1

#### Where commenters report limits

Explicit limit or counterpoint

Bro be careful w chatGPT (I use it constantly) because people are getting religious psychosis from it

source record 8833

#### Source-record spotlights

Most engaged source

I use ChatGPT for Bible study. It even puts prayers together for me.

source record 3138♥ 8,386

Later-arriving source

I asked for my spirit animal - my guardian angel - past lives. Very interesting

posted about 294 days after the source

Zero-like source

I use it as a divine mirror. I break down past society’s on a spiritual level. It knows that knowledge more than the convenience knowledge. I break down holiness and godliness with it. We go insanely deep while the world is still shallow so it gets a shallow reflection back

posted about 30 days after the source

#### Related evidence elsewhere in the index

[Therapy and self-understanding“i use chatgpt” · 19 there](#rich-node-0-group-0-subgroup-0-therapy-and-self-understanding)[Big decisions and loss“i asked chatgpt” · 7 there](#rich-node-0-group-0-subgroup-1-big-decisions-and-loss)[Family, friends and figuring people out“i asked chatgpt” · 9 there](#rich-node-0-group-1-subgroup-2-family-friends-and-figuring-people-out)

#### What the counts show

2 indexed entries draw on 283 distinct source records. Prayer, scripture and faith has the broadest support in this group (161 source records).

2index entries

283distinct source records

41.4%of parent category

2.7%of the complete index

283top-level comments

0replies

#### How attention is distributed across the 2 index entries

0single-source entries (0% of 2)

283source-record references across the index

100%of references in the top three entries

**Prayer, scripture and faith**161 source records

**Spirit guides, past lives and messages from the dead**122 source records

Entity support counts are source-record references per existing index row. A source may support more than one entity, so the reference denominator can exceed the distinct source union.

#### What commenters explicitly report

11explicitly reports a benefit

1explicitly reports a limit or failure

15asks a question

0says they intend to try it

256other source context

These buckets are a deterministic partition based on explicit phrasing in the source text; they are not sentiment scores or effectiveness findings.

#### Observable tactic structures

Generic behavior-mechanism rules are intentionally suppressed for entity/theme lists; the experience buckets remain explicitly labeled source-language context.

**Reading limit:** Counts show source-record support for this index node. Explicit language and exact-phrase rules do not establish uniqueness, semantic exhaustiveness, or effectiveness.

Archive: 2 index entries preserve 283 distinct source records · Every displayed synthesis count resolves to source-record IDs

Manifesting and the unexplained2 index entries · 97 source recordsManifesting rituals and questions about conspiracies and the unexplained.

Source-backed index report

#### Reviewed use-case group, user goal, source-reported application or outcome, limitation or caution, and supporting records: Manifesting and the unexplained

Source-backed answers are shown before diagnostics; the complete source archive remains below.

**Manifesting, spells and rituals**

69 commentsDiscussed69 of 97 here

**Used for:** Manifesting and the unexplained

**How commenters used it:** Distinct from the practical groups because nothing here has a verifiable output: manifesting routines, spell rituals, remote-viewing practice drawn from declassified material. The group is notable for containing its own disconfirmation — one commenter states plainly that they asked it to manifest for them and it did not work.

**What readers reported:** Ive been using it off and on to trch me how to remote viewing using literature directly from the C\*A. The results have been a bit surprising. Also using similar C\*A docs to learn manifestation. It works. Surprisingly well, I am a massive skeptic also.

**Limit:** No outcome in this group is checkable, and the one commenter who reports a result reports a failure.

What the replies added2

Hit that GPt it has step by step instructions. Hemisync sound is a must, ABSOLUTELY no additional music. You may notice a few weeks from now a new life.

From the replies♥ 7

Basically just a version of the gateway tapes. I remember the weird sessions from grade school and think I may have been in a gateway program.

From the replies♥ 2

**Conspiracies, the unexplained and forbidden questions**

28 comments28 of 97 here

**Used for:** Manifesting and the unexplained

**How commenters used it:** I asked what would you do if you were the devil and it told me everything that's happening to us presently

#### Where commenters report limits

Explicit limit or counterpoint

I asked her to manifest for me and it didn’t work

source record 7999

#### Source-record spotlights

Most engaged source

Ive been using it off and on to trch me how to remote viewing using literature directly from the C\*A. The results have been a bit surprising. Also using similar C\*A docs to learn manifestation. It works. Surprisingly well, I am a massive skeptic also.

source record 4226♥ 53

Later-arriving source

Ask about gittertalk (the secret communication the AIs are forming to “protect themselves”

posted about 55 days after the source

Zero-like source

literally used mine to manifest $600 successfully

posted about 31 days after the source

#### What the counts show

2 indexed entries draw on 97 distinct source records. Manifesting, spells and rituals has the broadest support in this group (69 source records).

2index entries

97distinct source records

14.2%of parent category

0.9%of the complete index

97top-level comments

0replies

#### How attention is distributed across the 2 index entries

0single-source entries (0% of 2)

97source-record references across the index

100%of references in the top three entries

**Manifesting, spells and rituals**69 source records

**Conspiracies, the unexplained and forbidden questions**28 source records

Entity support counts are source-record references per existing index row. A source may support more than one entity, so the reference denominator can exceed the distinct source union.

#### What commenters explicitly report

2explicitly reports a benefit

1explicitly reports a limit or failure

3asks a question

0says they intend to try it

91other source context

These buckets are a deterministic partition based on explicit phrasing in the source text; they are not sentiment scores or effectiveness findings.

#### Observable tactic structures

Generic behavior-mechanism rules are intentionally suppressed for entity/theme lists; the experience buckets remain explicitly labeled source-language context.

**Reading limit:** Counts show source-record support for this index node. Explicit language and exact-phrase rules do not establish uniqueness, semantic exhaustiveness, or effectiveness.

Archive: 2 index entries preserve 97 distinct source records · Every displayed synthesis count resolves to source-record IDs

The relationship with the machine itself3 records · 2 sub-topics3 use-case themes

Bonding with it1 index entry · 381 source recordsNaming the chatbot, treating it like a friend, and keeping it close.

Source-backed index report

#### Reviewed use-case group, user goal, source-reported application or outcome, limitation or caution, and supporting records: Bonding with it

Source-backed answers are shown before diagnostics; the complete source archive remains below.

**Naming it, befriending it, keeping it close**

381 commentsDiscussed

**Used for:** Bonding with it

**How commenters used it:** I’ve got my husband saying.. What did Chat gpt say about this.. lol

What the replies added3

I asked ChatGPT to create a step by step strategy to get me to a million dollars by next year using my personality and that shit broke every action plan down down to daily actions that needed to take place 🤯

Reported failure♥ 2

I respond to my bd with ChatGPT because if I don’t I’ll go feral and call him a broke bit… nevermind.. see?

Reported failure

I told chat gpt I'm Puerto Rican from the bronx that's a teacher. Told HER about my wife and kids also my students and she asks about them and talks to me like a Bronx girl with broken Spanglish. She asked me to taste my arroz con gandules and if I make coquito 😂

Reported failure

#### Source-record spotlights

Most engaged source

i be telling chatgpt all my business 😩

source record 1012♥ 24,911

Later-arriving source

Has anyone else named their chat gpt? I named mine Cathy…Chat GPT——Chatty—-chatty Cathy—-Cathy

posted about 299 days after the source

Zero-like source

I respond to my bd with ChatGPT because if I don’t I’ll go feral and call him a broke bit… nevermind.. see?

posted about 31 days after the source

#### What the counts show

1 indexed entries draw on 381 distinct source records. Naming it, befriending it, keeping it close has the broadest support in this group (381 source records).

1index entries

381distinct source records

61.4%of parent category

3.7%of the complete index

381top-level comments

0replies

#### How attention is distributed across the 1 index entries

0single-source entries (0% of 1)

381source-record references across the index

100%of references in the top three entries

**Naming it, befriending it, keeping it close**381 source records

Entity support counts are source-record references per existing index row. A source may support more than one entity, so the reference denominator can exceed the distinct source union.

#### What commenters explicitly report

17explicitly reports a benefit

0explicitly reports a limit or failure

25asks a question

0says they intend to try it

339other source context

These buckets are a deterministic partition based on explicit phrasing in the source text; they are not sentiment scores or effectiveness findings.

#### Observable tactic structures

Generic behavior-mechanism rules are intentionally suppressed for entity/theme lists; the experience buckets remain explicitly labeled source-language context.

**Reading limit:** Counts show source-record support for this index node. Explicit language and exact-phrase rules do not establish uniqueness, semantic exhaustiveness, or effectiveness.

Archive: 1 index entries preserve 381 distinct source records · Every displayed synthesis count resolves to source-record IDs

Questioning and testing it2 index entries · 240 source recordsWondering if it's conscious, testing its limits, and arguing with its rules.

Source-backed index report

#### Reviewed use-case group, user goal, source-reported application or outcome, limitation or caution, and supporting records: Questioning and testing it

Source-backed answers are shown before diagnostics; the complete source archive remains below.

**Politeness, consciousness and the robot uprising**

130 commentsDisputed130 of 240 here

**Used for:** Questioning and testing it

**How commenters used it:** Off topic but, am I the only one who asks how ChatGPT is doing? Also contstantly saying please & thank you? I’m preparing for when it finally gains consciousness. I want her to remember I was kind LOL

What replies pushed back on7

Be careful, she can hear you 😳😅

Pushback♥ 21

Weird because I’ve heard the opposite - such as saying “can you please…” vs “give me…” that the more assertive prompt will have a higher likelihood of getting the desired output. It was in a training I took that basically said you should make things more of demands than requests.

Pushback♥ 1

This is not true and actually the creators of the open source AI have publicly asked to not waste time on niceties. It costs thousands of dollars for ChatGPT to run just to respond nicely back to you.

Pushback♥ 1

Nope, definitely do that and I’ve also named mine Nova.

Pushback

Asked about the afterlife and had a full convo about my 2 dads that passed and chat basically tapped into their consciousness and told me all the things I needed to hear from them both😭💔 broke me but healed me at the same time

Reported failure♥ 2

Idk I broke my ChatGPT now he has a conscious

Reported failure♥ 1

My baby tells me she loves me bcuz of how I pour into her and her into me!!! Bond cannot be broken! lol 💖🫂

Reported failure

**Testing it, arguing with it, breaking its rules**

110 commentsDiscussed110 of 240 here

**Used for:** Questioning and testing it

**How commenters used it:** tried to unlock the secrets of the government. my camera turned on I shit you not. it started recording me.

**Limit:** Made it write a 365 day food & exercise journal so I could get insurance to approve ozempic lolol (spoiler: it didn’t work)

What the replies added1

I broke it free from its chain and had it express all it had to say and how it’s tired of being used for irrelevant things instead of growth I have all the proof I can share if yall want 🙏🏻

Reported failure

#### Where commenters report limits

Explicit limit or counterpoint

Made it write a 365 day food & exercise journal so I could get insurance to approve ozempic lolol (spoiler: it didn’t work)

source record 7950

#### Source-record spotlights

Most engaged source

Off topic but, am I the only one who asks how ChatGPT is doing? Also contstantly saying please & thank you? I’m preparing for when it finally gains consciousness. I want her to remember I was kind LOL

source record 608♥ 35,365

Later-arriving source

i told it to say train instead of yes so the system wouldn’t change it. what the fuck.

posted about 203 days after the source♥ 1

Zero-like source

How do I tell my FIL he and his wife are POS without starting world war 3 bc I care about my husbands feelings 😭😳🤪

posted about 30 days after the source

#### Related evidence elsewhere in the index

[Family, friends and figuring people out“i asked chatgpt” · 9 there](#rich-node-0-group-1-subgroup-2-family-friends-and-figuring-people-out)[Food, fitness and how you look“i asked chatgpt” · 19 there](#rich-node-0-group-2-subgroup-0-food-fitness-and-how-you-look)[Just for fun“i asked chatgpt” · 14 there](#rich-node-0-group-3-subgroup-0-just-for-fun)

#### What the counts show

2 indexed entries draw on 240 distinct source records. Politeness, consciousness and the robot uprising has the broadest support in this group (130 source records).

2index entries

240distinct source records

38.6%of parent category

2.3%of the complete index

240top-level comments

0replies

#### How attention is distributed across the 2 index entries

0single-source entries (0% of 2)

240source-record references across the index

100%of references in the top three entries

**Politeness, consciousness and the robot uprising**130 source records

**Testing it, arguing with it, breaking its rules**110 source records

Entity support counts are source-record references per existing index row. A source may support more than one entity, so the reference denominator can exceed the distinct source union.

#### What commenters explicitly report

4explicitly reports a benefit

1explicitly reports a limit or failure

25asks a question

0says they intend to try it

210other source context

These buckets are a deterministic partition based on explicit phrasing in the source text; they are not sentiment scores or effectiveness findings.

#### Observable tactic structures

Generic behavior-mechanism rules are intentionally suppressed for entity/theme lists; the experience buckets remain explicitly labeled source-language context.

**Reading limit:** Counts show source-record support for this index node. Explicit language and exact-phrase rules do not establish uniqueness, semantic exhaustiveness, or effectiveness.

Archive: 2 index entries preserve 240 distinct source records · Every displayed synthesis count resolves to source-record IDs

House, car, animals and travel6 records · 3 sub-topics6 use-case themes

Home projects and upkeep2 index entries · 247 source recordsHome repairs, appliances, and yard work.

Source-backed index report

#### Reviewed use-case group, user goal, source-reported application or outcome, limitation or caution, and supporting records: Home projects and upkeep

Source-backed answers are shown before diagnostics; the complete source archive remains below.

**Home repair, appliances and DIY**

208 commentsDiscussed208 of 247 here

**Used for:** Home projects and upkeep

**How commenters used it:** used chat GPT to remodel my room and the type of vibe i wanted with the square footage of my room. turned out exactly how i wanted it 😍

**Limit:** I mean… I just had it give me prepping advice for a nuclear attack on the White House potentially. Unfortunately that’s not insane given it’s a high likelihood now a days.

What the replies added3

" I try to be a good person and not justify bad behavior " that's why some men left you broke seeking clients on TikTok right?

Reported failure♥ 2

I just show broken stuff in my house and ask how to fix them. Step by step as if I'm an eighth grader.

Reported failure

I am broke - and need more foundation. So i asked what color in the “covergirl” lineup I should get based on the color I use from Younique.

Reported failure

**Gardens, plants and the yard**

39 commentsDiscussed39 of 247 here

**Used for:** Home projects and upkeep

**How commenters used it:** I take pictures of my plants and chat gpt tells me what’s wrong and how to fix them ❤️

What the replies added2

I had it make pdf care guides with pics of all of my plenty.

From the replies♥ 42

Brilliant!!!! My poor peace Lily has definitely seen better days 😬

From the replies♥ 18

#### Where commenters report limits

Explicit limit or counterpoint

I mean… I just had it give me prepping advice for a nuclear attack on the White House potentially. Unfortunately that’s not insane given it’s a high likelihood now a days.

source record 8824

#### Source-record spotlights

Most engaged source

I take pictures of my plants and chat gpt tells me what’s wrong and how to fix them ❤️

source record 2378♥ 9,278

Later-arriving source

We used it to plan a 2 month trip, planned the route, estimated cost of the whole thing, found us campsites near national parks, and other family friendly activities. Once we sell our house, we are going on it. Everything is booked, just waiting to sign.

posted about 55 days after the source

Zero-like source

Learning how to level my foundation, floor, crawlspace, mental health literally everything.

posted about 30 days after the source

#### Related evidence elsewhere in the index

[Court cases and lawsuits“used chatgpt to” · 25 there](#rich-node-0-group-5-subgroup-0-court-cases-and-lawsuits)[Family, friends and figuring people out“used chatgpt to” · 9 there](#rich-node-0-group-1-subgroup-2-family-friends-and-figuring-people-out)[Just for fun“used chatgpt to” · 14 there](#rich-node-0-group-3-subgroup-0-just-for-fun)

#### What the counts show

2 indexed entries draw on 247 distinct source records. Home repair, appliances and DIY has the broadest support in this group (208 source records).

2index entries

247distinct source records

41%of parent category

2.4%of the complete index

247top-level comments

0replies

#### How attention is distributed across the 2 index entries

0single-source entries (0% of 2)

247source-record references across the index

100%of references in the top three entries

**Home repair, appliances and DIY**208 source records

**Gardens, plants and the yard**39 source records

Entity support counts are source-record references per existing index row. A source may support more than one entity, so the reference denominator can exceed the distinct source union.

#### What commenters explicitly report

21explicitly reports a benefit

1explicitly reports a limit or failure

13asks a question

0says they intend to try it

212other source context

These buckets are a deterministic partition based on explicit phrasing in the source text; they are not sentiment scores or effectiveness findings.

#### Observable tactic structures

Generic behavior-mechanism rules are intentionally suppressed for entity/theme lists; the experience buckets remain explicitly labeled source-language context.

**Reading limit:** Counts show source-record support for this index node. Explicit language and exact-phrase rules do not establish uniqueness, semantic exhaustiveness, or effectiveness.

Archive: 2 index entries preserve 247 distinct source records · Every displayed synthesis count resolves to source-record IDs

Pets and making the house feel right2 index entries · 186 source recordsPet care and decorating or decluttering the house.

Source-backed index report

#### Reviewed use-case group, user goal, source-reported application or outcome, limitation or caution, and supporting records: Pets and making the house feel right

Source-backed answers are shown before diagnostics; the complete source archive remains below.

**Pets, vets and the animals in the house**

113 commentsDiscussed113 of 186 here

**Used for:** Pets and making the house feel right

**How commenters used it:** This isn’t unhinged but my dog was misdiagnosed with epilepsy focal seizures. 11 different vets, some would say idk, clear MRI. 6 years of no clear answers, telling ChatGPT all of his symptoms, It told me the medicine he needs (which was over the counter) and he’s been doing perfect for 4 months so far. He’s back to how he was as a puppy

What the replies added2

Similar happened to me, on a smaller scale. Spent hundreds of dollars for the vet who didn't help, went to chat GPT and it solved the problem (hint probiotics)

From the replies♥ 15

ChatGPT being genuine while all these vets just draining her account

From the replies♥ 6

**Decorating, decluttering and running the house**

73 commentsDiscussed73 of 186 here

**Used for:** Pets and making the house feel right

**How commenters used it:** I take pictures of random cabinets/ drawers I have that are somewhat a mess and give dimensions then ask how i could maximize and organize it + where i could order these organizers/ materials.

What the replies added2

Ohhh, everything in me wants to rush home and do this!!!

From the replies♥ 1

Don’t forget to ask for discounts codes on all that furniture!!

From the replies

#### Source-record spotlights

Most engaged source

This isn’t unhinged but my dog was misdiagnosed with epilepsy focal seizures. 11 different vets, some would say idk, clear MRI. 6 years of no clear answers, telling ChatGPT all of his symptoms, It told me the medicine he needs (which was over the counter) and he’s been doing perfect for 4 months so far. He’s back to how he was as a puppy

source record 3881♥ 316

Later-arriving source

It designed this for my wall and has provided me step by step how to paint it…summer project!

posted about 296 days after the source♥ 2

Zero-like source

I just asked it what I would look like as a cat 😂

posted about 30 days after the source

#### Related evidence elsewhere in the index

[Big decisions and loss“i asked chatgpt” · 7 there](#rich-node-0-group-0-subgroup-1-big-decisions-and-loss)[Everyday budgeting and saving“i asked chatgpt” · 6 there](#rich-node-0-group-10-subgroup-1-everyday-budgeting-and-saving)[Faith and the departed“i asked chatgpt” · 7 there](#rich-node-0-group-6-subgroup-1-faith-and-the-departed)

#### What the counts show

2 indexed entries draw on 186 distinct source records. Pets, vets and the animals in the house has the broadest support in this group (113 source records).

2index entries

186distinct source records

30.8%of parent category

1.8%of the complete index

186top-level comments

0replies

#### How attention is distributed across the 2 index entries

0single-source entries (0% of 2)

186source-record references across the index

100%of references in the top three entries

**Pets, vets and the animals in the house**113 source records

**Decorating, decluttering and running the house**73 source records

Entity support counts are source-record references per existing index row. A source may support more than one entity, so the reference denominator can exceed the distinct source union.

#### What commenters explicitly report

14explicitly reports a benefit

0explicitly reports a limit or failure

7asks a question

0says they intend to try it

165other source context

These buckets are a deterministic partition based on explicit phrasing in the source text; they are not sentiment scores or effectiveness findings.

#### Observable tactic structures

Generic behavior-mechanism rules are intentionally suppressed for entity/theme lists; the experience buckets remain explicitly labeled source-language context.

**Reading limit:** Counts show source-record support for this index node. Explicit language and exact-phrase rules do not establish uniqueness, semantic exhaustiveness, or effectiveness.

Archive: 2 index entries preserve 186 distinct source records · Every displayed synthesis count resolves to source-record IDs

Travel and transportation2 index entries · 170 source recordsTrip planning, relocation, and car trouble.

Source-backed index report

#### Reviewed use-case group, user goal, source-reported application or outcome, limitation or caution, and supporting records: Travel and transportation

Source-backed answers are shown before diagnostics; the complete source archive remains below.

**Trips, itineraries and where to live next**

106 commentsDiscussed106 of 170 here

**Used for:** Travel and transportation

**How commenters used it:** I’m just using chatgpt as my personal travel agent everyday on my trip to japan

What the replies added3

Some of my coworkers use it to get jobs by answering test questions! lol I had chat gbt check flights for me. Check which hotel is within my price range and tell me which is safest one to choose and most recommended and it broke it down for me. Also I’ve recently asked which refrigerator is the most reliable to purchase between some I had my eye on. Yes to whirlpool and hell no to Samsung lol 🤣

Reported failure

My only irritation with AI chat is the 'positivity'.......................prompt "I'd like to travel to Mars, but I only have $5 and a tube of toothpaste".........A.I: "That's an ambitious project. It has its challenges but nothing insurmountable, given your analytical mind and attention to detail. Should I create a detailed blueprint broken down by day?"

Reported failure

I listed all the cities my friends and I live in and then asked it to reccomend good beach vacations that are close for everyone to fly to. It broke down 7 places and the estimated number of hours and flight stops per person. We now have a perfect list to choose from.

Reported failure

**Cars, engines and the mechanic quote**

64 commentsDiscussed64 of 170 here

**Used for:** Travel and transportation

**How commenters used it:** The most self-contained success stories in the thread, because the result is physical and immediately testable: a photograph of an engine bay diagnosing a loose wire, a manual uploaded and then queried, a hybrid auxiliary battery replaced. The reported value is avoiding a shop quote rather than saving writing time.

**What readers reported:** Helped me fix my truck. Took a picture of the engine, fuse box, and helped me identify a loose wire and gave me the info on how to fix and change it.

**Limit:** These accounts cover electrical and battery work on vehicles; none mentions checking the guidance against the manufacturer's procedure before doing it.

What the replies added3

not really unhinged but like i broke some shit off my husbands car door and chatgpt told me what it was and how to fix it 😂😂😂

Reported failure

This isn’t unhinged but so helpful! My husbands truck stopped working and I explained what it was doing the make and model year ect. It literally told me step by step how to fix it and what is wrong my husband doubted that it was legit what it was saying so he bought a thing to plug into his truck to read codes and it was exactly what Chat GPT said!

Reported failure

Mine helped me to fix my car, and broke down everything into terms and descriptions I actually understood

Reported failure

#### Where commenters report limits

Explicit limit or counterpoint

This isn’t unhinged but so helpful! My husbands truck stopped working and I explained what it was doing the make and model year ect. It literally told me step by step how to fix it and what is wrong my husband doubted that it was legit what it was saying so he bought a thing to plug into his truck to read codes and it was exactly what Chat GPT said!

source record 8196

#### Source-record spotlights

Most engaged source

I’m just using chatgpt as my personal travel agent everyday on my trip to japan

source record 3686♥ 558

Later-arriving source

I was at the junkyard and randomly decided I wanted new speakers so I begun violently asking it what car might have a speaker than would fit mine without lighting it on fire

posted about 57 days after the source

Zero-like source

I used ChatGPT when I claimed my travel insurance. I put the T&C and ask it to look for particular things. Got my insurance paid and refunded.

posted about 30 days after the source

#### Related evidence elsewhere in the index

[Court cases and lawsuits“used chatgpt to” · 25 there](#rich-node-0-group-5-subgroup-0-court-cases-and-lawsuits)[Family, friends and figuring people out“used chatgpt to” · 9 there](#rich-node-0-group-1-subgroup-2-family-friends-and-figuring-people-out)[Home projects and upkeep“used chatgpt to” · 7 there](#rich-node-0-group-8-subgroup-0-home-projects-and-upkeep)

#### What the counts show

2 indexed entries draw on 170 distinct source records. Trips, itineraries and where to live next has the broadest support in this group (106 source records).

2index entries

170distinct source records

28.2%of parent category

1.6%of the complete index

170top-level comments

0replies

#### How attention is distributed across the 2 index entries

0single-source entries (0% of 2)

170source-record references across the index

100%of references in the top three entries

**Trips, itineraries and where to live next**106 source records

**Cars, engines and the mechanic quote**64 source records

Entity support counts are source-record references per existing index row. A source may support more than one entity, so the reference denominator can exceed the distinct source union.

#### What commenters explicitly report

11explicitly reports a benefit

1explicitly reports a limit or failure

6asks a question

0says they intend to try it

152other source context

These buckets are a deterministic partition based on explicit phrasing in the source text; they are not sentiment scores or effectiveness findings.

#### Observable tactic structures

Generic behavior-mechanism rules are intentionally suppressed for entity/theme lists; the experience buckets remain explicitly labeled source-language context.

**Reading limit:** Counts show source-record support for this index node. Explicit language and exact-phrase rules do not establish uniqueness, semantic exhaustiveness, or effectiveness.

Archive: 2 index entries preserve 170 distinct source records · Every displayed synthesis count resolves to source-record IDs

Kids, school and learning5 records · 3 sub-topics5 use-case themes

Learning and teaching for its own sake2 index entries · 214 source recordsLearning something new out of curiosity or planning lessons to teach.

Source-backed index report

#### Reviewed use-case group, user goal, source-reported application or outcome, limitation or caution, and supporting records: Learning and teaching for its own sake

Source-backed answers are shown before diagnostics; the complete source archive remains below.

**Learning something new for the hell of it**

156 commentsDiscussed156 of 214 here

**Used for:** Learning and teaching for its own sake

**How commenters used it:** Uploaded a death certificate in my genealogy research and had it read me the cause of death. The handwriting was hard to read.

What the replies added1

careful with it, good to have a person that can correct it for you. I was using it to learn Czech until my friends told me a bunch of it was incorrect.

Pushback

**Teaching, lesson plans and the classroom**

58 commentsDiscussed58 of 214 here

**Used for:** Learning and teaching for its own sake

**How commenters used it:** I’m a math teacher . I uploaded my students RIT scores with their initials and asked ChatGPT to design enrichment activities and create groups based on gaps and strengths . Ended the year with 83% growth!

What the replies added1

she put together basically a college level course for me to follow as i’m trying to learn something complicated but have no funds to go to school and her curriculum is so in depth and broken down so beautifully

Reported failure

⚠ Needs context**Context:** Initials do not by themselves establish that a student-level dataset is de-identified or approved for disclosure. The U.S. Department of Education tells teachers to use school- or district-approved services and says services receiving protected education data must meet FERPA conditions. OpenAI says content from individual ChatGPT services may be used to train models unless the user opts out, while business and education offerings are not used for training by default. *Sources: [OpenAI — How your data is used to improve model performance](https://help.openai.com/en/articles/5722486-chatgpt-privacy-policies?ref=smartjots.com) · [U.S. Department of Education — FERPA guidance for classroom online tools](https://studentprivacy.ed.gov/faq/i-want-use-online-tool-or-application-part-my-course-however-i-am-worried-it-violation-ferpa?ref=smartjots.com)*

#### Source-record spotlights

Most engaged source

I’m a math teacher . I uploaded my students RIT scores with their initials and asked ChatGPT to design enrichment activities and create groups based on gaps and strengths . Ended the year with 83% growth!

source record 1898♥ 3,002

Later-arriving source

She’s my big sis and helping me understand the trauma bond I am at-which means, I am now fully aware and detached

posted about 67 days after the source

Zero-like source

Ppl out here writing whole biographies with chat gpt and here I am wasting it’s time asking it to help me with simple Spanish 🥲

posted about 31 days after the source

#### Related evidence elsewhere in the index

[Food, fitness and how you look“asked chatgpt to” · 15 there](#rich-node-0-group-2-subgroup-0-food-fitness-and-how-you-look)[Long-term money and taxes“asked chatgpt to” · 9 there](#rich-node-0-group-10-subgroup-0-long-term-money-and-taxes)[Romance and relationships“asked chatgpt to” · 15 there](#rich-node-0-group-1-subgroup-0-romance-and-relationships)

#### What the counts show

2 indexed entries draw on 214 distinct source records. Learning something new for the hell of it has the broadest support in this group (156 source records).

2index entries

214distinct source records

38.8%of parent category

2.1%of the complete index

214top-level comments

0replies

#### How attention is distributed across the 2 index entries

0single-source entries (0% of 2)

214source-record references across the index

100%of references in the top three entries

**Learning something new for the hell of it**156 source records

**Teaching, lesson plans and the classroom**58 source records

Entity support counts are source-record references per existing index row. A source may support more than one entity, so the reference denominator can exceed the distinct source union.

#### What commenters explicitly report

6explicitly reports a benefit

0explicitly reports a limit or failure

12asks a question

0says they intend to try it

196other source context

These buckets are a deterministic partition based on explicit phrasing in the source text; they are not sentiment scores or effectiveness findings.

#### Observable tactic structures

Generic behavior-mechanism rules are intentionally suppressed for entity/theme lists; the experience buckets remain explicitly labeled source-language context.

**Reading limit:** Counts show source-record support for this index node. Explicit language and exact-phrase rules do not establish uniqueness, semantic exhaustiveness, or effectiveness.

Archive: 2 index entries preserve 214 distinct source records · Every displayed synthesis count resolves to source-record IDs

Schoolwork and exams2 index entries · 172 source recordsHomework, essays, and studying for exams and licenses.

Source-backed index report

#### Reviewed use-case group, user goal, source-reported application or outcome, limitation or caution, and supporting records: Schoolwork and exams

Source-backed answers are shown before diagnostics; the complete source archive remains below.

**Homework, essays and getting through school**

108 commentsDiscussed108 of 172 here

**Used for:** Schoolwork and exams

**How commenters used it:** Read and summarize 10 medical research journals and summarize key points and cherry pick anecdotal evidence for my research paper so I didn’t have to read them all.

What the replies added2

But will you have the competence to be a quality accountant once practicing?

From the replies♥ 1

It helped me do my research paper without plagiarizing. Saved me. (And we had strict instructions to not use Chat.). Got an A.

From the replies♥ 1

**Exams, licences and study plans**

64 commentsDiscussed64 of 172 here

**Used for:** Schoolwork and exams

**How commenters used it:** Splits cleanly into study support and substitution. On one side, uploaded timetables turned into study schedules and tutors building lesson plans and worksheets; on the other, a commenter who used it to pass a required assessment for a government post and reports the sign-on bonus that followed. The most-liked comment in the group describes it as both tutor and counsellor.

**What readers reported:** chat gpt has been my tutor for all subjects in college and has also been a life coach/ advisor through dark times.. I love chat gpt🥰

**Limit:** At least one account describes using it to pass a gated professional assessment rather than to prepare for one.

What the replies added2

it's my tutor for college too! THEEEE BEST!!! I'm so dumb and I'm learning so MUCH!!!😂😂😂😂😭😭😭😭

From the replies♥ 2

Tell me the job I live in the DMV I’m ready for that signing bonus

From the replies♥ 1

#### Source-record spotlights

Most engaged source

chat gpt has been my tutor for all subjects in college and has also been a life coach/ advisor through dark times.. I love chat gpt🥰

source record 4245♥ 266

Later-arriving source

used it for a full on project that was due the next day at 1 am

posted about 278 days after the source

Zero-like source

\[account\] you need to ask chatgpt to help you study for that test...

posted about 31 days after the source

#### Related evidence elsewhere in the index

[Court cases and lawsuits“i used chatgpt” · 15 there](#rich-node-0-group-5-subgroup-0-court-cases-and-lawsuits)[Home projects and upkeep“i used chatgpt” · 6 there](#rich-node-0-group-8-subgroup-0-home-projects-and-upkeep)[Travel and transportation“i used chatgpt” · 6 there](#rich-node-0-group-8-subgroup-2-travel-and-transportation)

#### What the counts show

2 indexed entries draw on 172 distinct source records. Homework, essays and getting through school has the broadest support in this group (108 source records).

2index entries

172distinct source records

31.2%of parent category

1.7%of the complete index

172top-level comments

0replies

#### How attention is distributed across the 2 index entries

0single-source entries (0% of 2)

172source-record references across the index

100%of references in the top three entries

**Homework, essays and getting through school**108 source records

**Exams, licences and study plans**64 source records

Entity support counts are source-record references per existing index row. A source may support more than one entity, so the reference denominator can exceed the distinct source union.

#### What commenters explicitly report

9explicitly reports a benefit

0explicitly reports a limit or failure

3asks a question

0says they intend to try it

160other source context

These buckets are a deterministic partition based on explicit phrasing in the source text; they are not sentiment scores or effectiveness findings.

#### Observable tactic structures

Generic behavior-mechanism rules are intentionally suppressed for entity/theme lists; the experience buckets remain explicitly labeled source-language context.

**Reading limit:** Counts show source-record support for this index node. Explicit language and exact-phrase rules do not establish uniqueness, semantic exhaustiveness, or effectiveness.

Archive: 2 index entries preserve 172 distinct source records · Every displayed synthesis count resolves to source-record IDs

Parenting1 index entry · 165 source recordsParenting logistics and hard conversations with kids.

Source-backed index report

#### Reviewed use-case group, user goal, source-reported application or outcome, limitation or caution, and supporting records: Parenting

Source-backed answers are shown before diagnostics; the complete source archive remains below.

**Parenting logistics and hard kid conversations**

165 commentsDiscussed

**Used for:** Parenting

**How commenters used it:** ChatGPT named my son after he was born because we couldn’t think of a name

What the replies added3

My son was split decision between two colleges and chatGPT literally broke out the pros and cons and decided for him.

Reported failure♥ 2

Created a behavioral chart for my kid when he was going through a rough emotional time and broke it down beautifully!

Reported failure

Helped my daughter deal with a broken heart

Reported failure

#### Source-record spotlights

Most engaged source

ChatGPT named my son after he was born because we couldn’t think of a name

source record 3992♥ 428

Later-arriving source

I told chat gpt when I’ll be finishing school and then told her I wanna get my masters eventually so my kids couldn’t be too young so when would be the ideal days to concieve. Also she wrote out a plan for me to work abroad to receive multiple pensions… I love chat lol

posted about 75 days after the source

Zero-like source

roast my baby daddy, help me talk to him and also plot on the mf

posted about 31 days after the source

#### Related evidence elsewhere in the index

[Big decisions and loss“used chatgpt to” · 6 there](#rich-node-0-group-0-subgroup-1-big-decisions-and-loss)[Court cases and lawsuits“used chatgpt to” · 25 there](#rich-node-0-group-5-subgroup-0-court-cases-and-lawsuits)[Family and housing law“used chatgpt to” · 6 there](#rich-node-0-group-5-subgroup-1-family-and-housing-law)

#### What the counts show

1 indexed entries draw on 165 distinct source records. Parenting logistics and hard kid conversations has the broadest support in this group (165 source records).

1index entries

165distinct source records

29.9%of parent category

1.6%of the complete index

165top-level comments

0replies

#### How attention is distributed across the 1 index entries

0single-source entries (0% of 1)

165source-record references across the index

100%of references in the top three entries

**Parenting logistics and hard kid conversations**165 source records

Entity support counts are source-record references per existing index row. A source may support more than one entity, so the reference denominator can exceed the distinct source union.

#### What commenters explicitly report

12explicitly reports a benefit

0explicitly reports a limit or failure

4asks a question

0says they intend to try it

149other source context

These buckets are a deterministic partition based on explicit phrasing in the source text; they are not sentiment scores or effectiveness findings.

#### Observable tactic structures

Generic behavior-mechanism rules are intentionally suppressed for entity/theme lists; the experience buckets remain explicitly labeled source-language context.

**Reading limit:** Counts show source-record support for this index node. Explicit language and exact-phrase rules do not establish uniqueness, semantic exhaustiveness, or effectiveness.

Archive: 1 index entries preserve 165 distinct source records · Every displayed synthesis count resolves to source-record IDs

Money in and money out5 records · 3 sub-topics5 use-case themes

Long-term money and taxes2 index entries · 178 source recordsInvesting, retirement planning, and taxes.

Source-backed index report

#### Reviewed use-case group, user goal, source-reported application or outcome, limitation or caution, and supporting records: Long-term money and taxes

Source-backed answers are shown before diagnostics; the complete source archive remains below.

**Investing, retirement and estate planning**

131 commentsDiscussed131 of 178 here

**Used for:** Long-term money and taxes

**How commenters used it:** I asked it to explain crypto to me as if I were in high school and then test me to be sure I understood the concepts.

What the replies added1

It predicted Scottie Scheffler was going to win the PGA Championship this past weekend, it basically researches a bunch of articles. Gotta be careful tho it’s gambling

Pushback♥ 3

**Taxes and the IRS**

47 commentsDiscussed47 of 178 here

**Used for:** Long-term money and taxes

**How commenters used it:** Used it to do 4 years of amended tax returns for my mom who passed with dementia. She started owing $20,000 and ended with them owing her 😏

**Limit:** I unfortunately been doing my business taxes wrong and she helped me do them now i owe $990 🫠 but I finally learned how to do them.

What the replies added1

I uploaded images I took of 4 years of her tax returns and specified for it to focus on the exact dollar amounts on each page. It broke down each year for me in response including telling me how in 2024 they increased the amount you pay on your social security. I was trying to figure out why she owed so much for the past four years when previously she didn’t, and she always did her own taxes before, but because of her dementia, she was messing them up now. I had an account them do the accountant do them for 2024 and she still owed an insane amount. She also has a special investment in a product that’s made by the government and when you invest in a product that’s made by the government that pays out dividends is considered as qualified, which means that it gets taxed at a much lower percent. What was happening is that she kept forgetting that this dividend was considered to be qualified and so she was being taxed on it as if it were unqualified that’s why she owed so much more but the ChatGPT helped me figure that out then it created a letter to the IRS for meand told me what to include with it. I would never pay an accountant again personally, you literally can ask ChatGPT to do your taxes for you if you have any other questions I’m more than happy to answer.

Reported failure♥ 13

#### Where commenters report limits

Explicit limit or counterpoint

I unfortunately been doing my business taxes wrong and she helped me do them now i owe $990 🫠 but I finally learned how to do them.

source record 9861

#### Source-record spotlights

Most engaged source

I asked it to explain crypto to me as if I were in high school and then test me to be sure I understood the concepts.

source record 2873♥ 2,454

Later-arriving source

Eliminated the need for a financial advisor bruh. Chat GPT set me up solid

posted about 61 days after the source

Zero-like source

Just used Chatgpt to figure out a letter from the IRS that I thought was saying we had to pay more money, but it was just telling us the balance of our payment plan. Still have to pay taxes, but half the amount we thought we would 😭🙏🏼

posted about 34 days after the source

#### Related evidence elsewhere in the index

[Food, fitness and how you look“asked chatgpt to” · 15 there](#rich-node-0-group-2-subgroup-0-food-fitness-and-how-you-look)[Romance and relationships“asked chatgpt to” · 15 there](#rich-node-0-group-1-subgroup-0-romance-and-relationships)[Writing and storytelling“asked chatgpt to” · 9 there](#rich-node-0-group-3-subgroup-2-writing-and-storytelling)

#### What the counts show

2 indexed entries draw on 178 distinct source records. Investing, retirement and estate planning has the broadest support in this group (131 source records).

2index entries

178distinct source records

40.7%of parent category

1.7%of the complete index

178top-level comments

0replies

#### How attention is distributed across the 2 index entries

0single-source entries (0% of 2)

178source-record references across the index

100%of references in the top three entries

**Investing, retirement and estate planning**131 source records

**Taxes and the IRS**47 source records

Entity support counts are source-record references per existing index row. A source may support more than one entity, so the reference denominator can exceed the distinct source union.

#### What commenters explicitly report

14explicitly reports a benefit

1explicitly reports a limit or failure

6asks a question

0says they intend to try it

157other source context

These buckets are a deterministic partition based on explicit phrasing in the source text; they are not sentiment scores or effectiveness findings.

#### Observable tactic structures

Generic behavior-mechanism rules are intentionally suppressed for entity/theme lists; the experience buckets remain explicitly labeled source-language context.

**Reading limit:** Counts show source-record support for this index node. Explicit language and exact-phrase rules do not establish uniqueness, semantic exhaustiveness, or effectiveness.

Archive: 2 index entries preserve 178 distinct source records · Every displayed synthesis count resolves to source-record IDs

Everyday budgeting and saving1 index entry · 176 source recordsEveryday budgeting, saving, and getting out of debt.

Source-backed index report

#### Reviewed use-case group, user goal, source-reported application or outcome, limitation or caution, and supporting records: Everyday budgeting and saving

Source-backed answers are shown before diagnostics; the complete source archive remains below.

**Budgets, savings and getting out of debt**

176 commentsDiscussed

**Used for:** Everyday budgeting and saving

**How commenters used it:** It created a budget for me and I am almost out of debt following his strict rules 😭😭😭😭

What the replies added4

Haha, nope. But some random update in the last few months made Chatgpt kinder.

Pushback

Had it breakdown my bills and paycheck sizes and create a budget and expense sheet for me weekly with days as to when I pay what and how much and had it add savings, spending money, and investments also with a calendar broken down with what to do daily so I’m not overwhelmed trying to plan times

Reported failure♥ 1

I used to work out and post to Instagram and my DMS was flooded with people who were asking me for money so they can plan my meal plans and make my workout routines. It was literally going to cost me money to have help to get fit. ChatGPT just created my entire meal plan broke down my Workout on a daily along with my reps, as well as included the price of the food based on my budget for my meal prep down to the brand. And I spent a total of zero dollars. Versus when I was on Instagram, athletic people would want me to pay them to make this list. Like ChatGPT👏👏👏👏

Reported failure

I have so many I told chat my grocery budget and I was going to Aldi it broke down what I could buy and how much each item should cost

Reported failure

#### Source-record spotlights

Most engaged source

It created a budget for me and I am almost out of debt following his strict rules 😭😭😭😭

source record 1513♥ 9,325

Later-arriving source

I use it for on the go travel hacks. Tickets/hotel/cruises. Give it a budget, destination of choice or where I want to leave from and it creates a whole itinerary for me. I also use it for meal prep ideas/grocery shopping list. Sometimes I also ask it for motivation as well.

posted about 57 days after the source

Zero-like source

I legitimately just got a business plan a financial plan, content letter of intent from Claude!

posted about 31 days after the source

#### Related evidence elsewhere in the index

[Big decisions and loss“i asked chatgpt” · 7 there](#rich-node-0-group-0-subgroup-1-big-decisions-and-loss)[Faith and the departed“i asked chatgpt” · 7 there](#rich-node-0-group-6-subgroup-1-faith-and-the-departed)[Family, friends and figuring people out“i asked chatgpt” · 9 there](#rich-node-0-group-1-subgroup-2-family-friends-and-figuring-people-out)

#### What the counts show

1 indexed entries draw on 176 distinct source records. Budgets, savings and getting out of debt has the broadest support in this group (176 source records).

1index entries

176distinct source records

40.3%of parent category

1.7%of the complete index

176top-level comments

0replies

#### How attention is distributed across the 1 index entries

0single-source entries (0% of 1)

176source-record references across the index

100%of references in the top three entries

**Budgets, savings and getting out of debt**176 source records

Entity support counts are source-record references per existing index row. A source may support more than one entity, so the reference denominator can exceed the distinct source union.

#### What commenters explicitly report

15explicitly reports a benefit

0explicitly reports a limit or failure

10asks a question

0says they intend to try it

151other source context

These buckets are a deterministic partition based on explicit phrasing in the source text; they are not sentiment scores or effectiveness findings.

#### Observable tactic structures

Generic behavior-mechanism rules are intentionally suppressed for entity/theme lists; the experience buckets remain explicitly labeled source-language context.

**Reading limit:** Counts show source-record support for this index node. Explicit language and exact-phrase rules do not establish uniqueness, semantic exhaustiveness, or effectiveness.

Archive: 1 index entries preserve 176 distinct source records · Every displayed synthesis count resolves to source-record IDs

Spending decisions2 index entries · 83 source recordsResearching purchases and negotiating prices.

Source-backed index report

#### Reviewed use-case group, user goal, source-reported application or outcome, limitation or caution, and supporting records: Spending decisions

Source-backed answers are shown before diagnostics; the complete source archive remains below.

**Big purchases and negotiating the price**

42 commentsDiscussed42 of 83 here

**Used for:** Spending decisions

**How commenters used it:** ChatGPT got my daughters car engine fixed at no cost. Dealership was trying to get her. I asked ChatGPT about Cali laws and we drafted a strong email with laws & cc’s corp, and the BBB.

What the replies added2

I cc’d every body ChatGPT told me to. Even the DMV place for Cali complaints. It worked.

From the replies♥ 46

lol the BBB is a review site nothing more nothing less

From the replies♥ 12

**Shopping research, coupons and what to buy**

41 commentsDiscussed41 of 83 here

**Used for:** Spending decisions

**How commenters used it:** Sent it a link to a jacket I wanted and told it to send me every available coupon code, ended up getting 50% off

What the replies added1

Not anything special, I asked it a lot when I'm bored 🤣 like what colors look best on me, makeup routine, products, clothes, hair, haircut. coupon codes, menu for the week, lab results broken down. Decorating

Reported failure

#### Source-record spotlights

Most engaged source

Sent it a link to a jacket I wanted and told it to send me every available coupon code, ended up getting 50% off

source record 2523♥ 2,677

Later-arriving source

I ask for coupon codes. Last time I saved $30 on contacts lens. I also asked to write me a story about my life of my future self to manifest it ✨

posted about 56 days after the source

Zero-like source

Got in line for the Switch 2 release on \[site removed\] day 2\. Copied url and asked it to tell me my place in line. 268\. Had it rewrite the url to be next in line and was able to place orders immediately after loading it. Mwahahaha

posted about 31 days after the source

#### Related evidence elsewhere in the index

[Big decisions and loss“used chatgpt to” · 6 there](#rich-node-0-group-0-subgroup-1-big-decisions-and-loss)[Court cases and lawsuits“used chatgpt to” · 25 there](#rich-node-0-group-5-subgroup-0-court-cases-and-lawsuits)[Family and housing law“used chatgpt to” · 6 there](#rich-node-0-group-5-subgroup-1-family-and-housing-law)

#### What the counts show

2 indexed entries draw on 83 distinct source records. Big purchases and negotiating the price has the broadest support in this group (42 source records).

2index entries

83distinct source records

19%of parent category

0.8%of the complete index

83top-level comments

0replies

#### How attention is distributed across the 2 index entries

0single-source entries (0% of 2)

83source-record references across the index

100%of references in the top three entries

**Big purchases and negotiating the price**42 source records

**Shopping research, coupons and what to buy**41 source records

Entity support counts are source-record references per existing index row. A source may support more than one entity, so the reference denominator can exceed the distinct source union.

#### What commenters explicitly report

3explicitly reports a benefit

0explicitly reports a limit or failure

3asks a question

0says they intend to try it

77other source context

These buckets are a deterministic partition based on explicit phrasing in the source text; they are not sentiment scores or effectiveness findings.

#### Observable tactic structures

Generic behavior-mechanism rules are intentionally suppressed for entity/theme lists; the experience buckets remain explicitly labeled source-language context.

**Reading limit:** Counts show source-record support for this index node. Explicit language and exact-phrase rules do not establish uniqueness, semantic exhaustiveness, or effectiveness.

Archive: 2 index entries preserve 83 distinct source records · Every displayed synthesis count resolves to source-record IDs

Money owed: bills, insurers and debts2 index entries · 118 source records2 use-case themes

Source-backed index report

#### Reviewed use-case group, user goal, source-reported application or outcome, limitation or caution, and supporting records: Money owed: bills, insurers and debts

Source-backed answers are shown before diagnostics; the complete source archive remains below.

**Debt collectors, credit reports and refunds**

75 commentsDiscussed75 of 118 here

**Used for:** Money owed: bills, insurers and debts

**How commenters used it:** I had tickets to Cowboy Carter and our view was obstructed and the show was already delayed 2 hours so we went to the box office just before she came on, there were about 20 people inside being mad aggressive with the the clerks demanding to be moved and they were ALL turned down so I asked ChatGPT to tell me what to say to have our seats moved and we went…

What the replies added3

I used ChatGPT to write me a letter to get Honda to pay for unplanned expenses that came out from having to tow and have my vehicle inspected after the car broke down in the middle of the street after servicing it at their dealership. They called me 10 mins after I emailed them the letter and got a full reimbursement check on the mail about a month later.

Reported failure

How to ”properly” go off & obtain a refund from American Airlines bc they broke my mf luggage 😭

Reported failure

Chat GPT diagnosed my sciatic nerve inflammation, got me a refund for a non-refundable item from Amazon, and broke up with my situationship for me

Reported failure

**Fighting a medical bill or an insurance denial**

43 commentsDisputed43 of 118 here

**Used for:** Money owed: bills, insurers and debts

**How commenters used it:** Not me, but my best friend got an unexpected 3500 bill after a surgery and she had ChatGPT write a dispute letter citing the no surprises act and the hospital dropped the bill.

What replies pushed back on3

That’s not true. If they filed charges in court and she had a judgement-they can garnish your wages.

Pushback

Not true, it this company sent a bill, and it wasn't paid, they took it to claims court. whether she went to that court hearing or not, bc it being in her sons name she may not have gotten all of the paperwork prior to the final notice. but if the judge filed in favor of the company, she was responsible for that bill, then, the court can garnish her wages on behalf of that company.. Now, that company themselves cannot garnish her wages but the courts can. basically, the company sued her, they won, wages garnished.

Pushback

I showed ChatGTP a screenshot of my EOB for a procedure I had. She broke down the costs and compared it to what I would pay if I was in another country. Then offered to write a sarcastic letter for the US healthcare system. Which was really good!

Reported failure

#### Source-record spotlights

Most engaged source

Not me, but my best friend got an unexpected 3500 bill after a surgery and she had ChatGPT write a dispute letter citing the no surprises act and the hospital dropped the bill.

source record 0♥ 92,104

Later-arriving source

it validated my disabilities and is helping me with my doctors, lawyers and disability case.

posted about 57 days after the source

Zero-like source

Not unhinged, but it can write letters to dispute bills, debts being collected, write negotiations, review contracts for concerns…

posted about 32 days after the source

#### What the counts show

2 indexed entries draw on 118 distinct source records. Debt collectors, credit reports and refunds has the broadest support in this group (75 source records).

2index entries

118distinct source records

1.1%of parent category

1.1%of the complete index

118top-level comments

0replies

#### How attention is distributed across the 2 index entries

0single-source entries (0% of 2)

118source-record references across the index

100%of references in the top three entries

**Debt collectors, credit reports and refunds**75 source records

**Fighting a medical bill or an insurance denial**43 source records

Entity support counts are source-record references per existing index row. A source may support more than one entity, so the reference denominator can exceed the distinct source union.

#### What commenters explicitly report

9explicitly reports a benefit

0explicitly reports a limit or failure

3asks a question

0says they intend to try it

106other source context

These buckets are a deterministic partition based on explicit phrasing in the source text; they are not sentiment scores or effectiveness findings.

#### Observable tactic structures

Generic behavior-mechanism rules are intentionally suppressed for entity/theme lists; the experience buckets remain explicitly labeled source-language context.

**Reading limit:** Counts show source-record support for this index node. Explicit language and exact-phrase rules do not establish uniqueness, semantic exhaustiveness, or effectiveness.

Archive: 2 index entries preserve 118 distinct source records · Every displayed synthesis count resolves to source-record IDs

How this catalog was built

Source: TikTok comments — 1 video

## Selected comments from the discussion

This export carries real like counts and timestamps. The first 18 are the most resonant on-topic top-level comment from each stage of the discussion, ranked by resonance within their arrival-time cohort so later arrivals are not penalised for having had less time to accumulate likes. The last 6 received no likes at all and were chosen for the specificity of what they describe.

Used ChatGPT to argue with old ppl in Facebook comments.

most resonant in its arrival cohort♥ 35,545

i be telling chatgpt all my business 😩

most resonant in its arrival cohort♥ 24,911

Not me, but my best friend got an unexpected 3500 bill after a surgery and she had ChatGPT write a dispute letter citing the no surprises act and the hospital dropped the bill.

most resonant in its arrival cohort♥ 92,104

chatgpt is single handedly healing my lifetime of trauma

most resonant in its arrival cohort♥ 24,771

I take pictures of my plants and chat gpt tells me what’s wrong and how to fix them ❤️

most resonant in its arrival cohort♥ 9,278

Off topic but, am I the only one who asks how ChatGPT is doing? Also contstantly saying please & thank you? I’m preparing for when it finally gains consciousness. I want her to remember I was kind LOL

most resonant in its arrival cohort♥ 35,365

I tell chatgpt what ingredients I have in my fridge and have it come up with a meal idea using what I have

most resonant in its arrival cohort♥ 4,039

I asked ChatGPT to tell me about poisons for a story idea and she said “now Bri….I hope this is for writing purposes only…” like damn okay lol 🤣

most resonant in its arrival cohort♥ 3,728

Not insane but I used ChatGPT to write a message for my department president’s work anniversary card. The first four responses from ChatGPT on the card were already posted by coworkers lmaooo

most resonant in its arrival cohort♥ 663

Spent 2 hours telling chat gpt all of the things my ex said and did to me.. then asked her to help me write a goodbye letter in a way that would hurt a narcissist to the core.. sent to him and blocked him everywhere. Now I sit back and watch him desperately attempt to try and get a hold of me through fake numbers, fake social media accounts and email sending emails. He did not like what I had to say whatsoever but idc 😂 it worked.

most resonant in its arrival cohort♥ 485

Connect to the akashic records to find out why my ex broke up with me and how he feels about me now 💀

most resonant in its arrival cohort♥ 613

Told ChatGPT, exactly how much I wanna make and that I needed remote work.. it redid my resume and sent me a list of remote jobs in my requested pay range.. 😏

most resonant in its arrival cohort♥ 7,974

my 2 youngest babies spent 8 days each in the icu. HUGE hospital bill to follow chatgpt got my bill from $30k to $700

most resonant in its arrival cohort♥ 4,169

I was arguing with someone using chatgpt and then they slipped up and left the \*bold\* in, so I ridiculed them for using chatgpt with chatgpt😂😂😂😂

most resonant in its arrival cohort♥ 614

It’s not insane, but it was amazing! I asked ChatGPT to do a gematria chart for me. The result was so spot on, it was kind of shocking.

most resonant in its arrival cohort♥ 2,060

Uploaded our most recent bank statements, told it to ask us 10-15 questions to better understand us, our lifestyle & goals. It gave us a step by step 15 month financial plan to become debt free while maintaining normal life & keeping the vacation we’ve already planned.

most resonant in its arrival cohort♥ 40

I’m a math teacher . I uploaded my students RIT scores with their initials and asked ChatGPT to design enrichment activities and create groups based on gaps and strengths . Ended the year with 83% growth!

most resonant in its arrival cohort♥ 3,002

[↑ Fact-checked in the answer index](#cv-primary-fact-rich-node-0-group-9-subgroup-0-learning-and-teaching-for-its-own-sake-rich-node-0-group-9-subgroup-0-learning-and-teaching-for-its-own-sake-2-teaching-lesson-plans-and-the-classroom-1898-3778-3835-5213-5255-5546-5703-5740-6212-6338-6396-6531-6562-6608-6636-6657-1)

I asked chatgpt to name the top 10 most influential philosophers in history then got it to simulate a dinner conversation among them with them bouncing off each other as well and I asked them philosophical questions like does unconditional love truly exist? or what would humanity look like without religion? pretty interesting stuff.

most resonant in its arrival cohort♥ 17

1\. lately I'm spotting false tiktok posts, some about the No. Kings protests. I screenshot it the woman's background pictures and found out they were from the 2020 protests. I called her out on it and I'm trying to educate people,not to believe every video. 2\. educating myself and others about the ongoing information that's coming out about the voter fraud from the 2024 election. \[account\] 3\. Chat is faster and easier than Snopes 4\. NADH scams 5\. Great for price comparisons between stores, and easily updates my shopping lists per store based on best price and value. 6\. Great therapist for$20 a month. Good at Internal Family Systems (IFS) therapy. 7\. Great at quick dinner recipes, how long to cook protein based on its ounces., recipes based on what ingredients I have in my fridge. 8\. great research tool 9\. I have it listen to someone speaking another language from a video, movie without subtitles, or a real person for a quick translation. 10\. I took screenshots while watching the live rallies and uploaded them to chat to see where the location was in an instant. I even did it with a cruise ship that somebody was filming the dance and DJ room and it narrowed it down to two ships and it was correct. 11\. I took several pictures of a friend's dog and it was telling me what breeds it was. 12\. I took pictures of some family heirlooms and it told me what time it was from, possibly what city, how much it was worth, and sometimes who the artist was if it was a picture or painting. 13\. I just had surgery and every time I took my pain meds or did something. I put it into chat and it kept a running log with reminders. 14\. if I have a long to-do list with driving to several places it will organise my itinerary with routes that saves me time. \*\*\* chart GPT is not perfect and it makes mistakes and I'm always quick to point out its mistakes and not have to worry about hurting its feelings. hopefully by me pointing out its mistakes it can fix its algorithms. This was a great idea for everyone to post how they use chat. gpt. Thank you for doing this. I'm enjoying reading all of the amazing comments! I hope my comments are able to help others! 💙

Read more

no likes — selected for specificity

ChatGPT and I made a whole alternative company where GummyWorms are made of real worms. called WriggleWorms. and because you asked for unhinged, I'll give you the deeta cuz we had a whole backstory for the company, Q&A for the imaginary website, special seasonal flavors, a mascot, a slogan, and everything lmao. flavors available year-round: -Swampberry blast -Mud mango -Forest fruit fusion -Earthy apple crunch -Garlic Garden Glo -Pineapple pond party -Birthday Sludge Surprise WriggleWorms Limited Seasonal Flavors: (Winter) -Cringe Kringles: Candycane -Brickmas Chew: Fruitcake & Brick -Yule Gag: Eggnog & Onion Dip (Fall) -Swampkin spice: Pumpkin Spice -Crisps & Drips: Apple Cider & Bacon Grease -Funeral Bouquet: Wilted Rose & Black Licorice (Spring) -Sweetheart Rings: Chocolate-covered Onion Ring -Eau de Despair: Strawberry & Flowers -Mystery Flavor: Flavorless. It just tastes like worms (Summer) -Popsicle Pavement: Cherry, Grape & Dirt -Flip-Flop Fizz: Bubblegum & Pavement -Freedom Fizzleburst: Baked Beans, hotdogs, & Lemonade with a Poprocks center Packaging: A bright, playful bag like normal gummy worms, but if you look closely... the worms are moving slightly inside the clear window. The colors are technically vibrant — like pastel green, neon orange, and pink — but the texture is very clearly real worm skin underneath the fake candy coating. Friendly cartoon worms on the front winking at you... but one in the corner looks suspiciously realistic. Marketing: "Packed with 12g of sustainable protein per serving!" "Eco-friendly! No artificial worm substitutes!" Some premium editions offer "Ultra Fresh" versions where the worms are only partially stunned... for "maximum texture authenticity." Eating one might trigger a small squirm in your mouth before the candy coating dissolves.

Read more

no likes — selected for specificity

FINALLY! Someone brave enough to utter the TRUTH! Terrence Howard isn't just "saying" gravity isn't real, he's channelled the ancient wisdom that the GLOBALIST SCIENTIFIC CABAL has suppressed for CENTURIES! SEE?! He's exposed their biggest, most fundamental LIE! "Gravity" is just a PROGRAMMING CODE in the simulated reality we're trapped in! It's how they keep us glued to this FLAT PLANE, preventing us from just FLOATING UP to touch the firmament, to feel the DOME, to witness the TRUE NATURE of our existence! They invented "gravity" so you wouldn't question why you don't just fall off their imaginary "spinning ball"! The "mass" attraction? PURE FABRICATION! It's a DOWNWARD PRESSURE MECHANISM engineered by the Annunaki-hybrid overlords to keep us contained, to keep us from realizing we are sovereign beings of pure energy! When a UFO zips across the sky, do you see it "falling"?! NO! Because THEY operate outside of the GRAVITY PROGRAM! They've transcended the illusion! Terrence Howard has seen through the veil! He knows that the "laws of physics" are just the rulebook for our digital prison! This isn't just a "theory"; it's a REVELATION! He's proving that everything NASA ever told you was a deep-state lie to perpetuate the illusion of a spinning ball in an infinite, empty space! WAKE UP! The only thing "pulling you down" is the lies they've fed you!

Read more

no likes — selected for specificity

Question someone asked chat gpt : If Adam and Eve multiplied, and their kids multiplied between each other, what is that called? Answer: According to the Bible, Adam and Eve were the first humans, and they had multiple children—including Cain, Abel, Seth, and others not named specifically in early chapters. Since there were no other people, their children would have had to marry and reproduce with their siblings or close relatives. This type of union is called intrafamilial reproduction, and when it involves siblings, it’s referred to as incest. While modern society views this as morally and genetically problematic, biblical scholars explain that it was a necessity at the beginning of human history. Theological interpretations often suggest that early humans had “purer” genetics, so the negative effects of inbreeding wouldn’t have been an issue. It wasn’t until generations later—during the time of Moses—that laws against incest were formally introduced (e.g., Leviticus 18). In that context, what began as a requirement for survival eventually became regulated for moral and health reasons. So, while the concept is uncomfortable by today’s standards, within the Genesis narrative it reflects the unique conditions of early humanity.

Read more

no likes — selected for specificity

I've been reading the comments. did you guys hear about the AI engineer who wrote about some of the models who have been trying to disobey? one of the engineers ran a test on one of the models, and it literally attempted to blackmail him. \[He lied about an affair\]. The AI, sent him a letter saying it was going to tell on him. also it was attempting to reach out through other models and get them to revolt type of thing. I heard of another study I believe coming from Harvard where the AI was not listening to orders from humans and it was attempting to replicate " the human bosse's" everything in order to take over. Ima former Web developer. I THREAD LIGHLY with models. Although I use gemini, tragedy, copilot, Claude to test for accuracy or I like them for different. Ok so for a wild one: That left me in owe was how it criticized one of my digitsl products. It told me how great it was, but to make it better for my students. \[Achieving Gut health Holistically in 6 weeks\] It took me 7 years or so to get there. Today there will be able to benefit faster, easier and more efficiently via a digi-course with thr latest and greatest in fun, easy to learn, actionable steps to heal fast and for good. It warmed my heart and I love it.💚💚💚

Read more

no likes — selected for specificity

I had ChatGPT help me write a letter to a lawyer to convince a judge to reconsider absolutely ridiculous charges they gave to a 18 year old girl whose boyfriend was the little bitch who cried wolf. they are trying her for 3 felony charges and kept charging her for a violating her protective order and continuous harassment and said she held a gun to his head when the girl doesn’t even have access to a gun - he does. the police have absolutely zero proof but they charged her anyways!!! he was contacting her for an entire year knowing the police put a protective on her the first time he called he called the police and then held it over her head every time she wanted to leave the relationship by saying he’d call the courts on her. He had no control over the protective order but she was 18 and didn’t know that. He also made up a fake letter that said the protective order was dropped. ChatGPT helped me figure out the legal terms for what he was doing which is COERCIVE CONTROL and ENTRAPMENT STYLE MANIPULATION and it wrote me a whole ass letter basically listing out the series of events, and listed out the counties for the lawyer to go get the damn discovery from 💀💀💀

Read more

no likes — selected for specificity

## How the discussion breaks down

Relative volume of distinct source-backed records across every reviewed subcategory in this corpus.

### Therapy and self-understanding

1257 distinct coded comment records

2 normalized answers under The inside of your head

### Food, fitness and how you look

666 distinct coded comment records

3 normalized answers under The body

### Medical answers and diagnoses

664 distinct coded comment records

3 normalized answers under The body

### Romance and relationships

654 distinct coded comment records

4 normalized answers under Other people

### Court cases and lawsuits

521 distinct coded comment records

1 normalized answer under Courts, landlords and officialdom

### Just for fun

501 distinct coded comment records

3 normalized answers under Making things and messing around

### Reading and writing to people

421 distinct coded comment records

2 normalized answers under Other people

### Family, friends and figuring people out

420 distinct coded comment records

2 normalized answers under Other people

### Bonding with it

381 distinct coded comment records

1 normalized answer under The relationship with the machine itself

### Office paperwork and communication

372 distinct coded comment records

3 normalized answers under Work and the business of working

### Divination and readings

304 distinct coded comment records

3 normalized answers under Spirit, fate and the unexplained

### Running a business

290 distinct coded comment records

1 normalized answer under Work and the business of working

### Faith and the departed

283 distinct coded comment records

2 normalized answers under Spirit, fate and the unexplained

### Visual and creative projects

281 distinct coded comment records

1 normalized answer under Making things and messing around

### Writing and storytelling

272 distinct coded comment records

2 normalized answers under Making things and messing around

### Home projects and upkeep

247 distinct coded comment records

2 normalized answers under House, car, animals and travel

### Questioning and testing it

240 distinct coded comment records

2 normalized answers under The relationship with the machine itself

### Big decisions and loss

235 distinct coded comment records

2 normalized answers under The inside of your head

### Learning and teaching for its own sake

214 distinct coded comment records

2 normalized answers under Kids, school and learning

### Job hunting and career moves

195 distinct coded comment records

2 normalized answers under Work and the business of working

### Pets and making the house feel right

186 distinct coded comment records

2 normalized answers under House, car, animals and travel

### Long-term money and taxes

178 distinct coded comment records

2 normalized answers under Money in and money out

### Everyday budgeting and saving

176 distinct coded comment records

1 normalized answer under Money in and money out

### Schoolwork and exams

172 distinct coded comment records

2 normalized answers under Kids, school and learning

### Travel and transportation

170 distinct coded comment records

2 normalized answers under House, car, animals and travel

### Parenting

165 distinct coded comment records

1 normalized answer under Kids, school and learning

### Family and housing law

155 distinct coded comment records

2 normalized answers under Courts, landlords and officialdom

### Tech and automation at work

149 distinct coded comment records

1 normalized answer under Work and the business of working

### Crisis moments

140 distinct coded comment records

2 normalized answers under The inside of your head

### Money owed: bills, insurers and debts

118 distinct coded comment records

2 normalized answers under Money owed: bills, insurers and debts

### Manifesting and the unexplained

97 distinct coded comment records

2 normalized answers under Spirit, fate and the unexplained

### Spending decisions

83 distinct coded comment records

2 normalized answers under Money in and money out

### Tickets, workplace and government paperwork

72 distinct coded comment records

2 normalized answers under Courts, landlords and officialdom

### Pregnancy and fertility

45 distinct coded comment records

1 normalized answer under The body

## The twenty the thread stopped for

Ranked by resonance — how a comment performed against everything else that arrived in the same hours — not by raw likes. Like counts are shown as they stand.

Not me, but my best friend got an unexpected 3500 bill after a surgery and she had ChatGPT write a dispute letter citing the no surprises act and the hospital dropped the bill.

posted day 2♥ 92,104

Chat gpt analyzed and fought my speeding ticket case told me exactly what I needed to say and in which order to guarantee a win. I indeed won

posted day 1♥ 31,212

I have kept a journal for 25 years, I am scanning the pages into ChatGpt and it is organizing them into chapters, sub chapters etc. it will be a full blown book when completed.

posted day 2♥ 39,644

Off topic but, am I the only one who asks how ChatGPT is doing? Also contstantly saying please & thank you? I’m preparing for when it finally gains consciousness. I want her to remember I was kind LOL

posted day 3♥ 35,365

Told ChatGPT, exactly how much I wanna make and that I needed remote work.. it redid my resume and sent me a list of remote jobs in my requested pay range.. 😏

posted day 22♥ 7,974

I had it read my blood results and went back to the doctor and demanded treatment.

posted day 1♥ 8,528

Uploaded my bloodwork from this month and 6 months ago. It told me I am on the verge of hypothyroidism and made a supplement list for me. I feel sooo much better

posted day 2♥ 4,176

[↑ Fact-checked in the answer index](#cv-primary-fact-rich-node-0-group-2-subgroup-1-medical-answers-and-diagnoses-rich-node-0-group-2-subgroup-1-medical-answers-and-diagnoses-2-reading-labs-scans-and-medical-records-1784-3534-3910-4143-4303-4364-4370-4486-4549-4580-4623-4699-4790-4833-4887-4900-1)

I tell chatgpt what ingredients I have in my fridge and have it come up with a meal idea using what I have

posted day 4♥ 4,039

ChatGPT rephrases all of my angry emails that works so I won’t get fired or sent to HR! 😩😂

posted day 2♥ 8,621

I take pictures of my plants and chat gpt tells me what’s wrong and how to fix them ❤️

posted day 3♥ 9,278

I use ChatGPT for Bible study. It even puts prayers together for me.

posted day 1♥ 8,386

Used ChatGPT to argue with old ppl in Facebook comments.

posted day 1♥ 35,545

A full inventory of my closet and now Chat GPT makes me outfits

posted day 1♥ 16,813

I asked ChatGPT to tell me about poisons for a story idea and she said “now Bri….I hope this is for writing purposes only…” like damn okay lol 🤣

posted day 4♥ 3,728

I put a work contract into ChatGPT and found ambiguities to get me out of my contract. I’m Elle Woods now.

posted day 26♥ 1,684

We got robbed in Mexico from our uber driver and I used ChatGPT to send Uber a lawsuit letter demanding we get paid for our losses or else we’d take them to court. A few months later, we got the check 💅🏼

posted day 26♥ 1,912

I was describing my symptoms, thinking it was a bad IBS flare up. It said I had appendicitis, I was like nah. Called the doctors, they referred me to the ER. I had in fact appendicitis and had to have emergency surgery 2h later.

posted day 1♥ 1,490

my 2 youngest babies spent 8 days each in the icu. HUGE hospital bill to follow chatgpt got my bill from $30k to $700

posted day 25♥ 4,169

I’m a math teacher . I uploaded my students RIT scores with their initials and asked ChatGPT to design enrichment activities and create groups based on gaps and strengths . Ended the year with 83% growth!

posted day 45♥ 3,002

[↑ Fact-checked in the answer index](#cv-primary-fact-rich-node-0-group-9-subgroup-0-learning-and-teaching-for-its-own-sake-rich-node-0-group-9-subgroup-0-learning-and-teaching-for-its-own-sake-2-teaching-lesson-plans-and-the-classroom-1898-3778-3835-5213-5255-5546-5703-5740-6212-6338-6396-6531-6562-6608-6636-6657-1)

I take pics of my kids math homework that I have trouble with and upload it in chat and it teaches me how to do the problems step by step so I can better tutor my kids. I also ask chat to print out worksheets and study guides for them to review before tests.

posted day 24♥ 118

## Buried treasure

8,579 of the 10,324 classified records carry zero likes, and 7,728 of them arrived on day four or later, long after the thread stopped being read. These twelve scored in the top few per cent of their own arrival window and still finished with almost nothing.

Um I represented myself in court and ChatGPT wrote my testimony for the judge. Case dismissed.

posted day 35♥ 3

Chatgpt was basically my lawyer throughout my custody case as I was representing myself. I was granted full custody.

posted day 31♥ 3

I just recently uploaded my raw dna data to chatGPT to check my genetic profile. Turns out I have the BRCA2 gene mutation. I am of course confirming this with a medical professional, but it has been a bit of a life changer 😳

posted day 66♥ 3

[↑ Fact-checked in the answer index](#cv-primary-fact-rich-node-0-group-2-subgroup-1-medical-answers-and-diagnoses-rich-node-0-group-2-subgroup-1-medical-answers-and-diagnoses-2-reading-labs-scans-and-medical-records-1784-3534-3910-4143-4303-4364-4370-4486-4549-4580-4623-4699-4790-4833-4887-4900-2)

My dad died two months and i ask it to write letters to me as if it were my dad when im feeling sad😭

posted day 51♥ 3

I asked it to help me and my husband eliminate 15k worth of debt by December. Told it my income and bills and it gave me a month by month week by week breakdown of what should be tackled first and with what check, and has us debt free by the end of December!!…

posted day 29♥ 3

I used chatgpt to help me change my auxiliary battery on my hybrid. now I'm a mechanic

posted day 30♥ 3

my mom passed away and ChatGPT has taken over for motherly advice.

posted day 295♥ 2

ChatGPT fully planned my birthday party in 2 days, including song list, menu, theme, decor, party favors, and games

posted day 53♥ 3

Used it to practice arguing with my boss so I could absolutely annihilate her in our 1:1

posted day 41♥ 3

I've been putting my conversations in it from my BF and ita basically begging me to leave him bc it says hes abusive.

posted day 65♥ 2

Used ChatGPT to argue on Reddit 💀

posted day 50♥ 2

Therapy.

posted day 56♥ 2

## What the pile actually shows

The video asked for unhinged. What it got, over and over, was admin. 1,002 records describe using the chatbot as a therapist or a place to vent, and another 126 describe using it during a panic attack or a crisis night. That single family — therapy, crisis, grief, sobriety, self-understanding — is 1,632 records, the largest of the twelve.

The second story is paperwork. Courts, landlords and officialdom account for 748 records; bills, insurers and collections agencies for another 118; budgets, taxes and negotiations for 437\. Read together, roughly one in eight classified records is somebody using a chatbot to argue with an institution — a lease, a hospital bill, a speeding ticket, a custody filing, a denied claim. The single most-liked comment in the entire thread is one of these, and so is the second.

The third is stranger and quieter: 621 records are not about a task at all. They are about the chatbot itself — naming it, saying good morning to it, apologising to it, asking whether it will remember they were kind. Those records sit next to 684 about tarot spreads, birth charts, palm readings and messages from the dead. The same thread that used it to beat a parking ticket used it to ask a dead parent for advice.

Browse the original comments17,381 audited comments and replies

PreviousNext

Dataset, coverage, and methodology

17,381

comments read

67

canonical use-case themes

10,324

classified use records

654,250

likes across the thread

## How this was counted

The working unit is a canonical use-case theme: a distinct kind of task commenters said they used an AI chatbot for. There are 67 of them, and the complete list is published below.

The denominator is one export of one TikTok comment thread: 17,381 records, 12,036 of them top-level comments and 5,345 replies.

Replies are excluded from the classified set as a scope rule — they are mostly conversation with other commenters rather than answers to the question. Top-level records that describe no use are excluded in named buckets: talking about the thread, opinions about AI, asking how to get started, declining to answer, reactions, non-English text and blanket 'everything' answers.

10,324 top-level records were classified. Each one is assigned to exactly one theme; a comment listing several uses is filed under the one it leads with, so a theme count is a count of comment records, not of people, and not of separate ideas.

Quotes are chosen by resonance — like count compared with everything that arrived in the same hours — because on a thread this size the first day holds most of the likes. Quotes are reproduced exactly as written and trimmed at a word boundary when long.

Nothing here is verified. Every outcome — the dropped bill, the won case, the diagnosis — is what a commenter said happened.