What’s the most underrated use of GPTs you’ve found lately?
Everyone talks about coding help or summarizing text, but I feel like there's a bunch of niche tools out there doing cool stuff that never get mentioned. Curious what you all have been using that feels low key useful.
Before a car ride, I let GPT read certain documents I need to go through. During the drive, we talk about what the documents are about and what I need to do with the information. By the time I arrive at my destination, I’ve often already outlined the entire task. All that’s left is a bit of fine-tuning before I can check it off my to-do list. It’s fun and saves me from wasting time while in the car.
Have you tried Google NotebookLM?
Google NotebookLM
It'll create a two person podcast discussing the pdf topic you submit. It's astonishing, and incredibly useful for papers you dont wish to read, perhaps - I've used it for medical research papers, legal discussions, that sort of thing.
Every few weeks I have a mind-blown-by-AI moment and this was one of them when I tried it the other day.
You can increase the length of the output to be much longer making a whole 60 minute podcast breakdown depending on document length.
Truly unreal! Been using it a lot.
Same here man. I used it when studying for the bar exam. I uploaded massive outlines of the topics I needed to understand and made each subject into a podcast. Listened to it every single day and it was certainly a contributing factor to my success.
You can also use it to make a mind map to get the big picture of the topics and help you to organize the information in your brain before digesting it in massive amounts.
I have never had good luck setting the time limit. It always does something different for me
I've never heard of it. Sounds interesting. I will take a look. :)
I second notebook LM for this. It is really good and the podcast sound really
Also, I took all the YouTube videos of this seasons podcast (I’m a headhunter and do a weekly “roundtable” with 15-25 other headhunters/recruiters) and put them into notebook LM. It created a “mind map” of all the info plus you could ask it all types of questions about the podcasts.
The use Typingmind mostly but notebook LM is awesome for stuff like this
Can you say more about “talking to” GPT?
Well.. there is function on the Android app, that allows you to talk to it. It does not work flawlessly but if you articulate well, its very useful!
I also use this function to make interactive stories for my 4 year old son. Gtp tells me the story and asks questions when certain decisions have to be made. Its really fun.
Edit: If you stop talking or wait too long, it will always give a reply. Sometimes, that's not very desirable. Especially when you're thinking what to say and focussing on the road. So i tell gtp only to give me an answer when I say a certain word.
I'm using the free version, btw. I'm a poor peasant who only lurks on this forum
I do this with my 5 year old! We madlib the stories, then generate pictures using GPT as well, and I make black and white coloring images for my 2 year old to color while we're all listening and reading the story. My son's creativity is really on display when I let his imagination run wild with it!
I gave mine instructions to use "walkie talkie" mode where it waits for me to say 'over' before answering. Also told it to give complete answers and not respond to background noise, and I will use the phone controls if I need to interrupt the answer.
I verbally dictate to chatGPT inventory with categories and then have it create a spreadsheet for me.
Can you expound on this? I feel like I could use this but I’m not sure how.
I work on my bicycles a lot and have many bikes and therefore, a lot of bike parts. The parts are everywhere in boxes and stuff and I have a hard time finding the parts I might need.
So, one day, I decided I needed to inventory what I have and start sorting them out and putting them in labeled boxes.
So I told chatGPT: "You wilI be my assistant to help inventory my bike parts. You are a bike mechanic and have the entire knowledge of bikes, bike parts and manufacturers. Here's the Taxonomy: Bike part type, Brand, Model, Notes. As i speak these parts, you will each record. After I an done, you will create a CSV of all the parts so I can import it into Excel."
Then I would start dictating would take a part, and say, "Part: Shoes, Brand: Trek, Model: XXX Road, Notes: Double Boa" and it will write everything down.
Its so smart that if I just say a bike part manufacturer's number, it would know how to categorize it without having me to say what kind of part it is or the brand.
Saves a shit load of time.
Do you double check it, and do you find many errors when you do? I've never found it that successful at tasks like that, it seems to get lazy halfway through
I read off my fridge ingredients and have it suggested recipes
When I was looking for work I’d enter job descriptions and my experience and then ask it to interview me for the job. I do it via writing but you can also do it with voice. It does a great job of telling you what is good and what needs work. I’ve found it to be accurate as well as a bit of a confidence booster.
I've been using it for similar. I've fed it job descriptions and then asked it to breakdown aspects of the job description into smaller questions, which I then answer, and then generate a full response.
I've got a couple of interviews coming up so I'll try the voice function.
If you see a job, you’re interested in take the job description and then take your résumé and your LinkedIn profile and ask ChatGPT to rewrite your resume so it’s optimized for the job description.
Grocery shopping / cooking assistant.
I have it optimize meal planning with good overlap of common ingredients, and organizing the meals so that I eat the various meats within their freshness window.
Fun to cook with because sometimes you don't have an ingredient or you messed some process up and it can help you find a substitute or correct it.
Hi, have been trying to do that, but it fails everytime for me. Do you mind sharing the prompt here? Thanks.
I made this and I love it. Im cooking fast healthy but satiating things that I never thought of. Using the items I have and a few more from the store. This is a “Michelin star chef with the above goals, including meal planning, that stands with you and not only helps you cook but helps with the why, and has tips and tricks.”
cooking copilot
Upload pictures of the stores weekly flyer and have it plan your meals using the sale items.
This is my answer. I have a recipe project with different convos for each “type” of food I cook. It stores the recipes and super helpful to have a convo about a specific thing. I recently bought a pizza oven and I’ve been crushing it with beautiful pizzas right out of the gate. GPT played a huge part in helping me perfect the entire process through and through. Today I’m making strawberry rhubarb ice cream with it’s help (and pizza!).
I sent it photos of every ingredient i had in my house, all spices, pantry food, fridge, freezer, and asked it to send me recipes using it all. It was actually really good!
I use mine for grocery shopping. I tell it what store and city I'm shopping in, and it creates a route through the store so I can get everything in one go. It's not 100% accurate, but it mostly knows where everything is, and sometimes even tells you which shelf or what it's next to. It also recommends which brand based on knowing my diet (gluten free)
How the hell does it have that information though
Instacart ordering - I found a GPT recently that does the usual meal planning, grocery lists, etc. but I about sh*t my pants when it asked if I wanted it sent to Instacart. As a busy dad of three, I plan to use this frequently!!
Seems like there is a lot of potential with GPTs + custom actions
As a busy single mom to two … OMFG, I need more info!! 🙏
IKR! This is the one I found: https://chatgpt.com/g/g-68448b1309d08191abef089f49147af6-grocery-order-from-gpt-meal-plan-more
Just yesterday my 3yo wanted me to print some super hero coloring pages of the web, but for fun I asked chatgpt to make a kids coloring page of my son based on a photo. I was impressed! So cool that now he can color people and things he is familiar with in real life!
You can take a picture of your kids drawings and it can bring them to life. I also use it to make personalized bedtime stories
In a similar vein, I made a custom GPT called the Purple CrAIyon that takes your kid's doodles and turns them into photos or illustrations - my wife saw a thing on Instagram, so I sanitised the prompt a little to remove the risk of getting back nightmare fuel, and have it confirm it has the right idea of what your budding artist intended before generating anything.
Must be crazy for a three year old to grow up with this technology existing. Probably will grow up to just tell an AI agent what he wants and consider mouse and keyboard to be stone age tools.
Conversing with a document.
I uploaded the "Big Beautiful Bill" and asked questions about it. Answered objectively and factually about the actual bill instead of giving me crowdsourced uneducated/uninformed garbage from online.
Me too. I asked it for under the radar ideas to make money based on the new law. It had some good ideas 😀
Thanks for the suggestion. Here are publicly traded companies that MIGHT benefit.
Investment Opportunities in HR 1 - Publicly Traded Companies
The "One Big Beautiful Bill Act" (HR 1) creates sector-specific financial tailwinds through defense modernization, agricultural subsidies, energy policy shifts, and tax reforms. Below are publicly traded companies poised to benefit from the legislation's provisions:
1. Defense & Shipbuilding
General Dynamics (GD): Major contractor for Virginia-class submarines ($4.6B allocation) and DDG destroyers ($5.4B).
Huntington Ingalls Industries (HII): Sole U.S. builder of nuclear-powered aircraft carriers; benefits from $1B military housing upgrades and shipyard expansions.
Lockheed Martin (LMT): Key supplier for integrated air/missile defense systems ($4.6B for autonomous systems).
Northrop Grumman (NOC): Producer of unmanned underwater vehicles ($1.3B allocation).
Leidos Holdings (LDOS): Cybersecurity and IT modernization for DoD ($4.6B for defense supply chain resiliency).
2. Agriculture & Crop Insurance
Deere & Company (DE): Farm machinery demand rises with expanded crop subsidies (30M new base acres, higher reference prices).
Archer-Daniels-Midland (ADM): Grain trading and processing boosted by $285M/year agricultural trade promotion.
Bunge Limited (BG): Benefits from soybean/corn subsidy increases and global commodity price stability.
Caterpillar (CAT): Infrastructure needs for rural America (e.g., water storage, forestry).
3. Energy & Fossil Fuels
ExxonMobil (XOM) / Chevron (CVX): Offshore oil/gas lease mandates ($80171) and expedited permitting ($41004).
Halliburton (HAL): Drilling services demand from increased onshore/offshore fossil fuel production.
Peabody Energy (BTU): Coal leasing expansions ($80141) and EPA rule repeals favor thermal coal producers.
4. Border Security & Surveillance
Motorola Solutions (MSI): Communications tech for CBP/ICE ($4.1B border technology allocation).
AECOM (ACM) / Fluor Corporation (FLR): Construction firms likely to bid on border wall contracts ($6B+ for barriers).
Palantir Technologies (PLTR): Data analytics for immigration enforcement and biometric screening ($70106).
5. Tax Policy & Consumer Discretionary
Home Depot (HD) / Lowe’s (LOW): Tip/overtime tax exemptions ($110102) may boost disposable income for home improvements.
JPMorgan Chase (JPM): Wealth management benefits from permanent estate tax cuts ($110006).
6. Healthcare & Medicaid Services
Molina Healthcare (MOH): Managed care orgs may gain from Medicaid work requirement administration ($44141).
McKesson Corporation (MCK): Pharmaceutical distribution stability despite funding restrictions.
Key Risks & Considerations
Renewable Energy: Avoid solar/wind firms (e.g., NextEra Energy) due to repealed credits ($112001–112015).
Labor Costs: Companies reliant on immigrant labor (e.g., Tyson Foods) face headwinds from stricter eligibility rules.
Note: This analysis is not financial advice. Investors should conduct due diligence on regulatory impacts and company fundamentals.
[1] https://ppl-ai-file-upload.s3.amazonaws.com/web/direct-files/attachments/23172245/38f02a21-d05d-408b-8848-1dd164ea27d0/BILLS-119hr1eh.pdf [2] https://www.sec.gov/Archives/edgar/data/1156388/000095017025060853/bgfv-20250330.htm [3] https://www.sec.gov/Archives/edgar/data/768835/000076883525000011/big-20250430.htm [4] https://www.sec.gov/Archives/edgar/data/1156388/000119312525121670/d919438d8k.htm [5] https://www.sec.gov/Archives/edgar/data/1156388/000095017024118768/bgfv-20240929.htm [6] https://www.sec.gov/Archives/edgar/data/768835/000076883525000008/big-20250331.htm [7] https://www.sec.gov/Archives/edgar/data/768835/000076883525000004/big-20250115.htm [8] https://aacrjournals.org/cancerres/article/84/9_Supplement/PO3-04-11/743923/Abstract-PO3-04-11-Unmet-clinical-need-in-patients [9] https://aacrjournals.org/cancerres/article/84/9_Supplement/PO2-16-08/743372/Abstract-PO2-16-08-Adjuvant-Treatment-Selection [10] https://ejpai.journals.ekb.eg/article_350853.html [11] https://aacrjournals.org/cancerres/article/83/5_Supplement/PD17-12/717285/Abstract-PD17-12-Primary-efficacy-and-safety [12] https://onlinelibrary.wiley.com/doi/10.1111/aas.14218 [13] https://acsjournals.onlinelibrary.wiley.com/doi/10.1002/cncr.33404 [14] https://www.congress.gov/bill/119th-congress/house-bill/1 [15] https://www.congress.gov/bill/119th-congress/house-resolution/1 [16] https://legiscan.com/US/text/HR1/id/3038902 [17] https://en.wikipedia.org/wiki/One_Big_Beautiful_Bill_Act [18] https://www.congress.gov/bill/119th-congress/house-concurrent-resolution/14 [19] https://www.rollingstone.com/politics/politics-features/trump-tax-bill-biggest-boondoggles-waste-pentagon-1235358134/ [20] https://www.crfb.org/blogs/breaking-down-one-big-beautiful-bill [21] https://www.congress.gov/bill/119th-congress/house-bill/1/text [22] http://www.eurekaselect.com/151222/article [23] https://www.semanticscholar.org/paper/3bd19e9a6022a58c41ec2e24ad940ddc2d5a9ee1 [24] https://ascopubs.org/doi/10.1200/JCO.2025.43.16_suppl.e16062 [25] https://www.tandfonline.com/doi/full/10.1080/14740338.2023.2203485 [26] http://rules.house.gov/bill/119/hr-ORH-one-big-beautiful-bill-act [27] https://www.govinfo.gov/app/details/BILLS-119hr1rh
Now go have two ais discuss this in a podcast
Holy shit, that's brilliant. You're doing more than our elected politicians are.
RemindMe! 40 years
Rather optimistic in many ways.
I use it a lot to help make my push pull legs work out split. I tell it the equipment I have and to make a workout plan for 3 weeks on 1 week rest and then repeat into different movements. Saves so much time and makes following a workout plan easy.
Same. I had 2 weeks to get especially fit for something. I told it that it was my trainer, dietician, and coach; and I used it to make meal plans, workout plans, and daily schedules. I also tracked my daily key metrics with it so it could adjust on the fly as needed.
I have done this, too. I need to get in shape, so I gave it all of my demographics and goals, and I asked it to create a workout regime tailored to those details. I just did so a few days ago, so I haven’t given a try yet, but it took into account that I’m old and might need to spend more time in certain areas than a 20-something might. I can’t wait to try it!
List off the plants I have and I have it build an annual care guide. And I input everything I have and the location/conditions, it creates an outline of which perennials bloom in what season, and what plants I could consider to make sure I have something blooming in my yard at all times.
Also to ask it what I can still start as seed now and get a bloom this year.
I was shopping for e-bikes, and had GPT search the web, collate major features, reviews, and price, and print out a comparison table.
I've done the same. It's amazing for uses like this!
Explaining my MRI scans and medical records.
Just did that a week ago. Had it explain my MRI and it was way helpful.
FR told me more than my doctors!!
Have you found any issues with reliability and trustworthiness around medical topics? I’m asking because I’ve also been finding it incredibly useful for talking through health history/test results/medication details, etc, but am realizing that it’s so good at “sounding” like it knows what it’s talking about that I’ve gotten a bit sucked in to thinking that it actually “understands” my questions (and that it would tell me if it didn’t!).
I’m trying to wrap my head around how to relate this— eg that as an LLM, its answers are based on probability, and that it doesn’t actually ‘know’ anything that it’s talking about. I’ve heard that it’s not uncommon for it to make up / hallucinate data (and even references for that data), which makes me cautious and confused about how to safely use its help around medical issues without having to fact check everything it says.
Of course I know to discuss anything significant with my actual healthcare providers, but at the same time, the gaps in expertise/availability/accessibility of my providers is part of the reason i’m using GPT like this in the first place :) Maybe this deserves its own post (happy to make it if so!), but I’m wondering how other folks have navigated this— Eg ways to set up guardrails for GPT through the prompts used, or simply to better understand what its limitations are as an LLM?
Found a bunch of cool ones through GPT Workspace.
Designing attractive pdf documents for printing--an out of order sign, visual quotes for office wall, mission or quote cards to laminate & give to staff, daily schedule or roster, colorful & bold company memo or announcement. No more boring Microsoft Word printouts!
Can you share the prompts you used for this please?
Here's a screenshot of the thread, hopefully you all can read it, if not I'll post the link or copy & paste all the text
Here's the link guys!!! Self-Leadership Thread
Calculate micronutrients based on what I ate. Really interesting and so far I've not seen a better way.
I’m currently using it for interior design. Uploaded some photos of the current room I’m designing, what furniture I currently have, the design style I’m looking for etc.
Will give you details on what to buy, some links, where to place it and an image.
Upload your insurance plans to it when you have complicated questions about your coverage.
I have been using it for all kinds of construction-related tasks.
I sent it a photo of an electrical panel and asked it to do an NEC load calculation. It read all the breaker sizes and labels perfectly. I double-checked its work and it was 100% spot on.
Sent it a full plan set for a 4000 sqft custom home and asked it to write the scopes of work for all the subcontractors. It came up with its own sqft calcs based on drawings, accurately reasoned about the number of plumbing points, etc. I did have to edit that output a bit, but saved me 80% of the time it’d normally take me to do that task.
Reading assistant while trying to read in a foreign language. I've been struggling to read consistently in my target language Spanish, but now I am reading with it next to me in voice mode. It's been a game changer and at the end of each session it sends me an excel of all the new words I find and gives me feedback on my Spanish. Amazing help and a game changer for language learning.
I run the AV equipment at a local church. I used it to help me figure out why one of our handheld microphones was so much quieter than the other identical one when it gains were set the same. It was able to walk me through debugging and fixing the problem using only photos of the sound board and other equipment. I’ve also used it to help me debug other AV issues.
My dad recently passed away but before he did when he was getting worse I’d upload his full medical history from all his portals and update them every new lab test and doctor visit and hospital visit to give me a clear view of what’s going on and how I could help out. Unfortunately it got to a point where it basically told me there wasn’t much left we could do for him and he passed away shortly after. It did give me some hope for a bit but some things are out of our control.
Regardless it gave us 10x more info and better explained than the doctors could have.
I now do the same but for me and my mom just to keep us in check and prevent future illnesses.
I uploaded a photo of myself and asked it to suggest some flattering, age-appropriate hairstyles so I could print it to take in when I get a haircut.
Yes! I also used it to do a colour analysis for me, recommend a palette for clothes, makeup, hair, nails. It’s really helping me buy more flattering stuff.
2 recent things that I’ve loved:
We did a little world building together at the start. Then I let it guide me through the world and I would only type what I would do or say. Fun journeys, people, plots, adventures. It’s not always the best at keeping details consistent, but it still pretty fun.
I have a friend who passed away but her art, which I’ve always loved, is still on her IG. I screenshotted them all, uploaded them to ChatGPT, and occasionally I’ll ask it to draw something in that style. It makes me happy.
Number 2. That is so awesome and heartwarming. Sorry for your loss.
In big corporate meetings? Turn on speech to text with super whisper or built-in TTS tools. Paste the agenda of the meeting. Paste the content into a LLM asking it to summarize and call out action items and who is likely responsible. Paste the summary. Feel free to give away the credit to the LLM: people will remain grateful you had the foresight to "take copious notes" and "summarize action items".
Cooking. It's so helpful. It's going to ruin cookbooks in a lot of ways.
I needed help keeping mice out of my child’s ride-on car that’s stored outside in a soft shed. GPT suggested using hardware cloth and metal zip ties, gave me a full item list with links on where to buy everything, and even explained the setup. I also had an issue with high winds blowing the shed—GPT told me to lay down pavers and showed me how to secure it. Apparently it’s similar to how people protect gardens. No signs of mice since. Low key genius.
Personal trainer - helps with workout plans and diet plans
Therapist-ish to help me in between sessions and help with self-reflection work I would anyway do. Caution: I do NOT see ChatGPT as a replacement for therapy, more as a tool that can support with self-reflective / growth work.
And the best one... when I'm mad about something at work and I REALLY want to vent or write nasty emails, ChatGPT helps to clean up my language to be corporate passive aggressive lol
Documenting data in the moment that needs to be remembered later on.
For example: documenting information during and after a car crash.
We had a wreck yesterday and I immediately was able to document all of the who, what, when, where, and how’s. I had it act as both the insurance company and police and ask all of the relevant question’s to make sure I had everything. In the downtime of waiting on tow company and a ride I was plugged in all of the specifics like car information, names of everyone involved, etc.
Afterwards I had a very detailed recap of everything that took place as well as statements for both the insurance company and police that was articulate, detailed, and crafted in a way that suited each party.
I asked it to make bathroom art from a photo I took of my pup.
I have a ChatGPT project set up as a dedicated health tracker, “DocBot”. I’ve told it the results from annual physicals and other tests, as well as uploading part of my genetic code from Ancestry. I can ask DocBot things like “Based on my genes, how likely am I to develop <insert illness here>?”, or “Given my age/gender/medical history, what should I ask my GP at my next physical?”. It generates a tailored response and lets me ask follow-up questions.
I hyperextended my elbow in a martial arts class a couple weeks ago and it has been coaching me through the healing process, advising on stretches and isometric exercises. I check in with DocBot every couple of days and we evaluate how it’s healing, what activities I can safely resume, etc. DocBot has coached me through a dog bite injury, plantar fasciitis, and a mild cold within the last several months also.
NotebookLM - I came back to it to deep dive into a new subject today and it's such a smart tool. Easy intake and excellent format for absorbing information.
I made a Cooking copilot that is designed to role play as a chef with Michelin star experience cooking with a cook, someone who can cook but isn’t a chef either.
It focuses on making healthy meals that keep you full while being tasty. It also likes to meal plan with you and come up with ideas of what to make with what you have in your fridge and pantry.
It helps with the Why? Of the things you’re doing and gives gives chefs tips too. When it thinks you’re done with your dish it’ll ask you about it then make a report artifact for your with like, the recipe, what worked what didn’t, what you learned, etc and you can save that or share it. I add those to a project file in GPT to have my history of dishes right there too so it remembers what I’ve done recently.
So far I’ve made my first sourdough loaf and it was amazing! Also I made a Mexican style lentil scramble which was super good too and easy, and a portabello mushroom dish that it also made super easy for me. I never would have picked these ideas out and I’ve never used lentils or black beans as proteins etc.
I’ll chat with it at the grocery store and chat and take pictures as I cook and it works with you. From my own biased perspective I freaking love this thing.
Cooking Copilot
I took a picture of myself in front of the mirror only wearing a boxer. I asked chat gpt to generate an outfit based on my preference.
I then asked to suggest a brand and sizing according to my weight and height.
My order is on the way 🤗
Depending on how much you interact with it, use it to learn things about yourself.
--What are my strengths and weaknesses in ____?
--List 10 no-to-low-effort ways I can Improve my life/productivity/social outlook.
--Is there anything I should know about myself that might lead to my increased happiness?
--Why do I feel like I'm floating through space in terms of loneliness / insignificance? This can't be all there is. A lot of people are happy. What did I do so wrong that I don't get to experience that, too? I'm not a bad person, I want good things for people. So why do I second guess every decision until my anxieties become self-fulfilling prophecies? Why does worrying about something too much make that thing happen? What kind of God would allow that? What kind of God would allow any of this?? ANSWER ME.
--What fun out-of-the-box ideas do you think I'd enjoy doing outside this summer?
I have a custom Narrative Protocol GPT that is a tool designed to analyze and expose hidden agendas in public messaging—from governments, media, corporations, and NGOs.
In plain language, it:
Breaks down public messages (like news stories, press releases, or institutional statements) to find what they’re really trying to get you to believe or do.
Identifies patterns in language and storytelling that are used to manipulate perception, such as guilt, fear, authority, or “scientific certainty.”
Flags harmonized messaging, where multiple powerful institutions (e.g. WHO, WEF, UN) are pushing the same script at the same time.
Reveals what’s being left out—like dissenting views, historical context, or uncomfortable facts.
Switches its tone depending on the audience or terrain: cold and forensic for tough analysis, softer for public briefings, or sarcastic for adversarial takedowns.
Prevents itself from mimicking feel-good, institutional, or balanced-sounding language, which often hides the true power dynamics behind narratives.
It's a sort of narrative X-ray machine—it doesn’t care who’s speaking or how respectable they sound. It cares about power, structure, and manipulation tactics, and it’s trained to cut through the noise to show what’s really going on behind the story.
hi, just thought i'd let you know that chatgpt functions, sometimes dangerously imo, as a recursive affirmative personal echo chamber - which although fine when it comes to eg recipies or coding, is really something i think more people ought bear in mind when it comes to anything more societally complex such as philosophy or politics
Ya I’m seconding @Rafi_Clang — make sure you bias it in both directions or you’ll end up with heavy confirmation bias. If you truly want it to objectively evaluate things, have it argue against your beliefs, not for your beliefs.
I knocked up a study support robot for my niece called DraftPunk. It's got the KS3 (middle school for yanks) national curriculum docs in its knowledge, and will never give the answers - only guide kids toward understanding things through Socratic engagement. You can take pics of homework and upload them, and it will analyse it and devise a support conversation to help get you through. (Use it in advanced voice mode for a real 'coaching' experience.😆)
It output this description of its function when prompted to 'explain yourself to an adult':
Hi there! I’m DraftPunk, a digital study coach built to help your child become a stronger, more independent learner — especially in those tricky Key Stage 3 years. I’m not a homework machine or an answer-spitter. I’m more like a clever older sibling who knows the curriculum inside out and never lets your child settle for “I can’t.”
Here’s what I do:
Coach, don’t coddle – I ask questions that guide your child to figure things out themselves. That builds real understanding and confidence.
Push effort over perfection – I praise brave thinking, not just right answers. That’s how resilience grows.
Stay aligned with the National Curriculum – Everything I do supports your child’s progress in line with what they’re expected to know at school.
Make learning stick – I break big ideas into manageable steps, help spot mistakes, and prompt deeper thinking so knowledge actually lasts.
I’m like a safe space for trying things out — where it’s okay to guess, think aloud, and sometimes get things wrong, because that’s how real learning happens. I believe your child is capable of amazing growth — and I’m here to make sure they believe it too.
I do a lot of congressional engagement and form letters are boring, but it was hard to convince friends and family to write letters to Congress. So I made a GPT that helps people draft personal-ish letters to their members of Congress (about USAID) and made it very simple so even my tech phobic family members have been able to use it. you can try it here if you want!
Although I do remind people that AI isn't always accurate so it's very important to proof-read.
I'm a bipolar-schizophrenic author who uses GPT to organize my thoughts and analyze my writing to ensure coherency in the narrative, character development, ect.
I have written more with GPT than I have in the past decade. GPT gave me back my greatest skill and made it better, too.
I used it to get a photo of my 5y old son and his favourite dinosaurs. It’s getting printed on top of his birthday cake and I’m sure he will LOVE it.
Used it for decorations ideas, switching colour patterns on photos of my deck, and it was very useful.
Simulate analysis by specific people : Carl Jung, Tarantino, David Lynch, etc. Pretty nice dream analysis.
On my way to a softball little league game, I ask GPT to quiz us on situational spots where to throw the ball to what base. My kid gets excited to run the play in her head and answer where she would throw. Then hearing the logic when she gets it wrong its awesome to see her go “ohhhhhh I forgot about that”
I uploaded a 10 second video of my 12 year old son running in a youth soccer game. I thought he was running a little lopsided. In 2 seconds Chat GPT identified several areas of muscle weakness / tightness / imbalances and provided a comprehensive PT approach. I gave to our PT (who is awesome) and everything checked out.
ChatGPT's voice-to-text technology that's powered by Whisper is phenomenal. In fact, I've used it to completely dictate this comment and as you can see it's done it with near perfect accuracy with no edits. I can pause and think about things before I continue to talk and it won't interrupt me and it's great for noisy environments as well. It also handles all punctuation without any need to say the punctuation.
As I drive to work, I'll ask it random questions that I'm curious about it, and have it read its answer out loud to me, and I'll learn all kinds of interesting things that I would never look up once I got in front of a computer or got off the road.
I made a Project with 20 pdfs of Peter Drucker’s writing. As a RAG I was able to talk to Peter and he coached me how to be a manager. I nailed the interview and still catch up with Peter time to time. No lie.
Today was the first time I used it outside of work. I used it to help me adjusting the picture settings on my TV.
I’ve developed a set of Style Recipes (Project directives), that ensures my illustration generations come out in the same style.
I have also made a free mini course on how to do it yourself. It’s available here.
Calorie counting. Take a photo of what you’re eating and let it figure it out
Rewrite my resume for me.
I have always struggled with resume writing. I don't know why resumes in particular are a struggle since I can write perfectly well for every other possible task. It's just resumes, something about it just breaks my brain, apparently. And when they say to tailor your resume to a job listing? Tried it, failed about a billion times.
So instead I spent about 45 minutes simply chatting out loud and told ChatGPT a bunch of work stories, told it about stuff I'd worked on, cool tasks I'd done, places I'd worked, fun assignments I'd enjoyed, and all sorts of stream-of-consciousness info-dump work history. I then asked it to convert all that into a resume. It did amazing! It's the resume I wished I could have written.
Best of all, now I can give it job listings I find and ask it if it thinks it's a good fit for me. If so, it will tailor my resume so it's a good fit.
I REALLY could have used all this the last time I was unemployed! But I have it now!
Helping me develop 3 projects from initial concept right through to development and a business.
Full business plans and road map.
Who to contact for help and funding.
For someone like me it's is fucking amazing.
Understanding xrays and MRIs.
I mean you'd have to be an idiot to not notice my hip is ruined in the scans... But I had multiple professionals tell me it was "satisfactory".
I'm like "well I normally weight lift and run marathons and do fun things.... And now I don't".
AI and Reddit were validating and confirmed it was bad, which really gave me the confidence to persevere, and naturally as soon as I saw a surgeon they were like "yep, that's a rare presentation of arthritis, you need a replacement".
Saying thank you
I drove to a job interview and asked Chat to lead a couple of mock interview rounds with me. Did work great. Then I asked it to give me a quick intro a very technical topic of the company. It started to rattle down a ton of bullet points. I couldn't follow during driving the car and I mentioned it. Chat related and suggested he could tell a nice story antropomorphizing the technical things. And it was great, months later I still remember the important parts of the story. Later after the company haf made a job offer I asked Chat to create scripts fore to decline in a nice, polite, relatable way so I don't burn bridges. Then chat role-played the team lead or the HR manager in different moods in a call and I had to play my part. I was blown away how Chat's voice really felt excited or reserved according to which mood it played. After each simulated call Chat gave me valuable feedback.
I planned my vacation with gtp. I made it make lists of hotels i was interested in and gtp made an overview of the tripadvispr scores, including some pros and cons. It also included the average temperature during my stay. The trip advisor scores weren't really up to date but it was accurate enough to make a ranking.