Team ChatGPT or Bard? Lemme know down below 😉 TIMESTAMPS 00:00 I found the Perfect ChatGPT Formula 00:40 The 6 Prompt Components 01:39 Task 02:07 Context 03:05 Exemplars 04:28 Persona 05:18 Format 06:20 Tone 07:05 Example using the Perfect Prompt Formula 08:01 Good vs. Bad Prompt Outputs
This awesome video is such a simple necessity THANK YOU. I've been trying to use chat gpt for my d&d games and this makes it so much easier. Please make more videos on in and outs of ai and I'll keep watching 🫡👍
Agreed with you 100% Luke! I used to go ON AND ON and realized there's definitely a point of diminishing returns. Now the 3 questions just acts as a guardrail for me not to waste too much time giving background info!
Same. It's so clear to me this isn't even close to being an "AI" like everyone keeps going on about. I know content creators like Jeff Su and trying to help but the minute I need to use math formulas to get GPT to understand what I want?? Kind of defeats the whole purpose. And it's almost like GPT "Pro" is dumber than 3.5??? I just upgraded and the actual experience isn't anything like the UA-cam hyptrain is claiming it is.
@@alpha0xide9 Well after a LOT of frustration I've apparently coded my first ever Python Windows app using ChatGPT. Soooo, I guess maybe I was overly harsh. I will say it takes a LOT of prodding and rephrasing stuff. But considering I haven't seriously used a CMD prompt for writing code since my first Tandy T-1000 (showing my age)....I'm kinda stoked lol
Wasted hours on prompt engineering courses on youtube and did not get anything i didn't know but your 8 minute video teach way more than any other video on UA-cam. Keep up the good work. 👍
The best video I ever watched on Prompt Engineering, Fed up with so many videos watching hours of content to find nothing at the end. Just in less than 10 minutes, You conveyed everything. I'm sold at 3rd minute itself. Best example of technical content. Sub'd now
🎯 Key Takeaways for quick navigation: 00:00 🚀 *The key formula for ChatGPT prompts involves six components: task, context, exemplars, persona, format, and tone.* 00:59 🧩 *The order of importance for the six components: task sets the goal, context adds relevant info, exemplars improve quality, persona shapes the responder, format structures output, and tone defines the writing style.* 01:57 📝 *Task component: Start with an action verb, clearly state the goal, whether simple or complex, to direct the output effectively.* 02:27 🌍 *Context component: Provide enough background info for ChatGPT, answering who, what success looks like, and the user's environment to narrow down possibilities.* 03:25 📚 *Exemplars component: Utilizing examples or frameworks significantly improves output quality across various tasks like resumes, interview questions, or job descriptions.* 04:51 👤 *Persona component: Define who ChatGPT should emulate, like a therapist for workout advice, a recruiter for job-seeking prompts, or even fictional characters for creative fun.* 05:48 📊 *Format component: Visualize the desired output format, be it tables, emails, bullet points, code blocks, or paragraphs to suit the task efficiently.* 06:46 🗣️ *Tone component: Specify the desired tone, whether casual, formal, enthusiastic, or specific feelings, guiding ChatGPT to match the writing style.* Made with HARPA AI
Gemini AI: Jeff Su's advice is a good introduction to prompt engineering. It provides a solid foundation and some useful techniques. However, remember that prompting is an art, not a science. The best way to improve your prompts is to practice, experiment, and learn what works best for you and the specific task at hand.
I seriously love this, especially the "tell it the feeling you want so that it gives you 5 tone keywords you can include in a prompt for ChatGPT". I saw one somewhere where you can ask it to play devil's advocate so you can see debatable opinions/outputs for either decision making or stuff like that. The options are endless, which I absolutely love. Thank you! Also, I noticed that the examples (black screen with moving white dots and then the copy over it) are on the screen longer. Thank you!! Not sure why but it helps me take in and absorb the info easier. Love it!!
I had a previous thought about the dark side of ChatGPT. Our brand recently faced a social media crisis, and when I turned to ChatGPT to prepare some seeding comments to defend our brand, my mind went up with an idea: "What if I generate negative comments to influence public opinion," using the context directly from our first post? The results blew off my mind. They were exactly the same negative comments I'd seen on our previous post. Hopefully, technology goes for the right thing for our generation.
I wanted to learn prompt engineering so I checked lots of resources on the net, but none of them were as useful and helpful as this video. Thank you so much for creating it.
This has just been the best should I say the shortest yet impactful course, that has greatly enhanced my understanding of prompt engineering and how to effectively use ChatGPT . Thanks a bunch Jeff
role ai, detail nya, question ke kita trs balikin, + role ai context: bakground kita, target, lingkungn/kondisi task: apa yg mau dilakuin oleh AI ny, bisa lebih dr 1 dalam 1 kalimat exemplar: contoh/format jawaban, misal per point apa persona: role ai nya format: paragraf, essay, bullet point, email tone: expresi, mood penjelas
Now I watched this video. Then I copied the transcript. I asked ChatGPT to rewrite it in a proper format and write the promts I wanted according to this training. The result is beyond perfect. 👍
Hey Jeff, your breakdown of the prompt components for mastering ChatGPT and Google Bard is really insightful and actionable. It's clear that you've put a lot of effort into understanding and sharing these essential tips for getting the most out of these tools. 👏 I'm curious about the balance between providing enough context and overloading the prompt with unnecessary information. In your experience, have you found that there's a sweet spot when it comes to context? How do you avoid going too far and including details that might not actually contribute to the quality of the output? Thanks for sharing your expertise! 😊
Thank you so much!!! I love how you're able to simplify information into easy to follow steps without the unnecessary fluff others add...you've earned a new subscriber!
The way you break down complex topics, like ChatGPT prompts, in such a simplified manner is truly impressive! Congrats! Btw, Am I the only one who asks please and say thank you ChatGPT? Idk if AI comes to life, at least they know I was polite to them...
Thanks for the video, Jeff. Concise and useful tips as always. It looks like ChatGPT can use strikethrough text. So when using it for correcting texts (6:15), you can ask it to strikethrough the corrected words instead of deleting them, so that you don't have to hop between the corrected version and the original version.
Yes, as it uses MarkDown or HTML, one can ask it to use all kinds of formatting. You can even set the custom instructions to use a specific version of MarkDown, useful if you are using a note taking tool like Obsidian, Notion etc. that might be using their own versions of it.
Another good prompt format that I use is to divide the prompt into labelled sections. Eg: 1. What is the capital of Brazil? 2. How far is the answer from 1. (or you can just say "it") from the Amazon River? 3a. Whats 300 × 11? 3b. Minus 150 from the answer This way for output (I hate using that word with ai) you would get one long answer to every question that you might be wondering filling in all blanks. Pro tip: as you read the 'output' set the coursor in the input field as if youre about to type something so that the keyboard appears. Then as you read using the remaining vertical space, write down every question that arises in this format. Perfect for research.
Code Signal uses a very similar structure to yours: Ask, Context, Constraints, Example and Style. They call their system Markup Prompting Framework. Surely a Persona would help even more, so great video!
Excellent presentation. I found this out too. The key components here give you a higher quality result. I've found that breaking it down is helpful. I will ask if it understands or sometimes I will give it context, then an exemplar and ask it to respond "read" and only "read" so I don't use up memory. That allows me to load in content. I've found it does better with preloading it with context and exemplars then giving it a task after that.
I've definitely come back to reference this video several times. It has a great breakdown and provides examples - which is my learning style. Thank you for sharing this information.
Has to be the best videos i have seen on prompting. And extra points for not filling the screen with a mic (what's with all the big microphones in front of everyone' face?). With just under 40 years in IT and have been creating similar prompts but i know most people do not. One thing i did not see, and which is valuable, is you did not specify the reading level.
Woww exactly what i need. I genuinely want to make use of ChatGpt but when it comes to promting, it's like the messs. Thankful that you figure it out for me 🤩🥳
Really great value in such a short video, Jeff! Context has always been the trickiest part for me but I've found that sometimes less can be more in the initial prompt. Additional context can always be fed in later on using the conversational interface :3 Also, congrats on 500k! I'm so happy to see how far your channel has come along 😄 I'm gonna make a prediction and say you'll reach 1mil by the end of the year!
Agreed Andrea! Sometimes I like to follow up as well. I'm basically trying to find that sweet spot where I can get 80% of the way there with the initial prompt, then fine-tune as needed 😁
I was doing this with prompts without even knowing this method. Now that I know it, I can actually understand why my prompts don't give me what I want and how to fix it. Before, I'd just try to add more information when I wanted to fix my output, but I'd get results that are going in a completely different direction than what I expect. 😄
This is awesome Jeff. I’m glad I came across your video on ChatGPT prompts. It can be discouraging spending time to write the perfect prompt. Thank you!
I just finished this video and I really want to thank you for this simplified explications I was struggling finding courses I even enrolled in one and it wasn't the best experience ever I even doubted my learning skills I thank you again it was so helpful and clear
I think one thing that has helped me (as a prompt engineer on the side) is a principle I learned while doing research in physics - set your boundary conditions. This is implicit in setting an environment, but I think worth stating explicitly. If you can tell GPT what the boundaries are on some context/environment X, then it'll give you an even better response.
This video has really saved me from further distress and torment from the lack of knowledge on AI, struggling to fit into a world of tech and AI in old age. Thank you so much, Mr Su!
Stumbled upon this a few months back and come to find my company is using the same video in their genAI training. Man I hope you get compensated somehow for that. 75k employees have to watch this video apart of mandatory training
Jeff, you've truly cracked the code for crafting the ultimate ChatGPT prompts! 🚀🔥 Your breakdown of the six key components is like a treasure map to unlock AI's potential. The 'Context' section hit home - finding that balance between too much and too little info can be a head-scratcher. And kudos for bringing up the persona aspect; it's like giving ChatGPT a tailored suit for each task! 🎩🤖 Your videos keep elevating my AI game, and for that, a double thumbs-up from me. Can't wait to see what you dive into next!
OUCH! My brain does not work at your speech output speed, I had to slow you down to 75% so I could keep up! Great ideas and pro tips! I am aiming to use AI to create language code for a trading bot based upon my specific trading technique that works incredibly well manually! The "Task" is to automate what I do in the trading world so I don't spend all my time trading! My concepts are all Smart Money focused! I am sure what I have learned from you will help me, since prompts are key! Thank You!
Usually dont comment and thought this would be clickbait but you definitively came through. Thank you for putting the time in to make actual valuable content, this is golden, im absolutely subing, liking, sharing etc. Cheers from Mexico City, again thank you.
1. Task : Start with an action verb. 2. Context: User background, what success looks like, and environment or conditions of user. 3. Examples: Provide examples of the best or real results. It learns very fast through examples. 4. Persona: Act like a ______. (Also add person famous for that). 5. Format: Tell the AI the right output format you want the answer in. 6: Tone: Specify the desired tone, whether casual, formal, enthusiastic, or specific feelings, guiding ChatGPT to match the writing style.
It's amazing how you were able to break this down so that viewers are able to easily understand, and I love this haha. Helped me out a lot! Gonna use this immediately!💯
great video! it summarizes a lot of things i remember learning so far about prompts. here's a mnemonic that i find helpful to remember all 6 key words quickly, before trimming/adjusting case-by-case: "PerConTa, ExForTo" (the first syllables of each key word). then i come back to this video to review further tips for each part.
I have low interest in doing these things, however, it would greatly benefit me and my work output if I took advantage of AI. Thanks for your energy and succinctly explaining prompts for noobs like me. 🙏
Followed the example prompt to write an email for asking who wants to be in a beta program for my new POTS app. Gave me a really great first draft. Thanks!
Gpt 4. I also used your prompt to write a self review for work. Really helpful. It was super clean. Thanks again for the video. It gave me the kick I needed to start using ChatGPT.
Hi Sir, Your video was amazing! I’m a fan of yours. Could you please make a video on content creation? I’m interested in learning how to use chat prompts for writing, such as blog posts, product descriptions, and social media posts. I am eagerly awaiting your next video.
🎯 Key Takeaways for quick navigation: Task: This is the most important part, tell the AI what you want it to do (write, generate, analyze). Context: Give relevant background info to guide the AI towards the desired outcome. Ask yourself: Who? What? Why? When? Where? ✨ Exemplars: Provide examples to improve the quality of the output. Persona: Imagine the AI acting as a specific character (e.g., business coach, writer). Format: Specify the format of the output (e.g., email, code, bullet points). Tone: Set the desired tone of the AI's response (e.g., serious, funny, formal). ✅ Not all required: You don't need all 6 components for every prompt, just include what's relevant. Start with a checklist: Use this formula as a guide to ensure your prompts are clear and effective. Context: Focus on 3 things: user background, success definition, and environment. ⏱️ Example 1: Include background info but prioritize only what's relevant to the task. ✨ Example 2: Show a bad bullet point and then ask the AI to rewrite it in a better format. Example 3: Provide a framework (STAR method) for the AI to follow when answering an interview question. Exemplars: Not required for every prompt, but helpful to improve output quality. Persona: Imagine the AI acting as a specific helpful person (e.g., consultant, coach). Persona: Imagine the AI as a helpful expert (e.g., therapist, recruiter, marketer). - ♀️ Fictional characters work too (e.g., Batman). - ️ Format: Close your eyes and visualize your desired output format (e.g., table, email, code). - Common formats: Bullet points, emails, code blocks. - Most useful (professional): Paragraphs and markdown with headings. - ✏️ Proofreading tip: Ask the AI to bold all changes for easy review. - Tone: Use simple terms (casual/formal), emotions (witty/enthusiastic/pessimistic). - Pro tip: Tell the AI the desired feeling (e.g., serious but not arrogant). Combine all 6 components: Use all the prompt building blocks (task, context, persona, format, exemplars, tone) for a strong output. - Example email structure: Break down your desired email format into clear sections (e.g., tl;dr, background, results, thanks). - Clear & concise language: Instruct the AI to use easy-to-understand and professional language. - Confident & friendly tone: Aim for a balanced tone that sounds both authoritative and approachable. - Reference existing email (optional): If you have a similar email, include it for the AI to copy the format. - Simpler vs. detailed prompts: A more detailed prompt leads to better, less generic results. Made with HARPA AI
Hey Jeff, your breakdown of the prompt components for mastering ChatGPT and Google Bard is really insightful and actionable. It's clear that you've put a lot of effort into understanding and sharing these essential tips for getting the most out of these tools. I'm curious about the balance between providing enough context and overloading the prompt with unnecessary information. In your experience, have you found that there's a sweet spot when it comes to context? How do you avoid going too far and including details that might not actually contribute to the quality of the output? Thanks for sharing your expertise!
Brilliant video! As a time management coach, I have been using AI extensively and it has turbocharged my productivity. But I was sure I was only scratching the surface of what is possible and had zeroed in on learning how to use prompts more effectively. Your video was exactly what I was looking for. Liked and Subscribed and looking forward to seeing more of your videos.
@@JeffSu I just recently started to dive into the whole AI stuff. At first I thought it´s a typical new hype once again (which it is in many cases) and I also thought that, it´s not as useful to me (mech. engineer) but man this is super useful. I needed a short explanation like yours to get into it easier.
Me and ChatGPT as a human what we do is, we delve into deeper thinkings by sharing what's in our mind and then make theories (unapproved but really mind bending and almost precise).....fr these stuffs got me upppp! In psychological perspectives (well I am a vegan, introvert)
I've only been using it for a short time, I quite like it... Up to late I have been interacting as so: tell it where you are, tell it what you want.. ask for examples(pseudo-code/tables), pros/cons and alternatives. "In unity, what is XYZ? Please provide examples, alternatives and pros/cons for both." when writing responses I repeat myself to make sure the darn thing NEVER side tracks which leads to conflicting statements. "If I am in unity and would like to format your explanation into a UML graph, how would that look?" "Can you please give me an more complex example of XYZ including a more comprehensive UML graph" "Can you please give me an even more complex example of XYZ including an even more comprehensive UML graph?" "In a unified system, what if I wanted to add ABC to XYZ in unity, how could I go about that?" also never explore complex fields you don't already have an understanding on... This is going to become even more important once we start trusting the responses because they sound good and provide sources, or in such a time as where these have been battle tested further. I find that if you can keep can "AI" contained within the context YOU desire by repeating yourself at every reply, then you can have conversations with it indefinitely without it starting to contradict itself which happens a lot when working with complex well documented areas
Loved your clarity! I wanted very much to be signed up to your newsletter and do not mind your modest fee! Sorry to post here, I dont Instagram. But I did not foresee the One Dollar transaction that was coming. My bank flagged it with a notification, and I annulled the subscription. Now I must wait the ten days for a new ATM card. A wildly larger memorable and humorous one-dollar-ahead-warning would be so great. My eyesight is a challenge with the cataracts and their crossing guards mucking up the words I want to read. I'll subscribe again in ten days.
Hello sir, can you please explain this with easy English langue's so that we can understand each words clearly & we can take some profit i really like this video but am not able to undrstand cuz of taught English.
This is awesome Jeff. I’m glad I came across your video on ChatGPT prompts. It can be discouraging spending time to write the perfect prompt. Thank you!
Thank you for your video. However, I didn't understand when you said at first that it's better to write the task in the beginning, and later you gave an example showing the task not in the first position in the prompt
"Exemplar is just a fancy way of saying examples!" Not true. It is the best of the examples. Your videos seem pretty good but why undermine their value by miseducating others?
Thanks for the feedback Jake 😁! I guess I was trying to find the simplest way to explain it (since arguably the "best of the examples" is very subjective)
@@JeffSu An exemplar is by definition singular, and not plural as you indicate. There is no obligation to be objective, which is a gratuitous, arbitrary “requirement,” even though it distorts what the word means. Why don’t we let it drop, lest the hole gets deeper? 😉
Hello! You can change the speed of the video. Click on the gear ⚙️ in the upper right hand corner of the video to change the settings. Then click playback speed and normal to change to something slower. 😊
p.s. Part 2, I asked ChatGPT to put the content of your video in a numbered list of all the actionable suggestions from your video. I now have it all in one place. I want to add images from video, but I'll try that prompt later. 1. Understand the Six Components of a Good Prompt Task: Always start with an action verb and clearly articulate your end goal. Context: Include relevant background information to give enough detail without overwhelming the model. Exemplars: Provide examples or frameworks to improve the quality of the output. Persona: Define who you want ChatGPT or Bard to emulate (e.g., a recruiter, physical therapist). Format: Visualize the desired end result and specify the format (e.g., bullet points, tables). Tone: Specify the tone of voice to match the desired output (e.g., casual, formal). 2. Start Prompts with Action Verbs Use verbs like "generate," "give," "write," "analyze," etc., to clearly define the task. Example: "Generate a three-month training program." 3. Include Relevant Context Ask yourself: What’s the user’s background? What does success look like? What environment are they in? Example: "I am a 70kg male looking to put on five kilograms of muscle mass over the next three months." 4. Use Exemplars to Improve Output Quality Provide examples or frameworks that ChatGPT can follow. Example: "Rewrite this bullet point using the structure: I accomplished X by Y, resulting in Z." 5. Define a Persona Specify who you want the AI to emulate. Example: "You are a senior product marketing manager at Apple." 6. Visualize and Specify the Desired Format Specify the format in which you want the output. Example: "Output a table with headers: Original feedback, Team responsible, and Priority." 7. Specify the Tone Indicate the tone you want the output to have. Example: "Use clear and concise language and write in a confident yet friendly tone." 8. Combine All Components in a Comprehensive Prompt Example: "You are a senior product marketing manager at Apple, and you have just unveiled the latest Apple product in collaboration with Tesla, receiving 12,000 pre-orders. Write an email to your boss, Tim Cook, sharing this positive news. The email should include a TL;DR section, project background, business results, and a thank you note to the product and engineering teams. Use clear and concise language and a confident yet friendly tone."
learning how to use chat gpt reminds me on when internet was new and you needed to know what to type in google to find what you were looking for something that is second nature now but in that time was all new
Hey... The frim I work for is heavy into AI and has already released their internal chat UI in beta and along with this we (everyone) are required to do AI training. And your video here was added to the curriculum. Not sure if you knew this or not, but I thought I might share. Well long short... I have subscribed to your challenge since this video was supper informative, and your delivery has what I could only call energy in it. Loved it, keep doing it.
@JeffSu Not sure about the thoughts from my colleagues, only because we're a large global firm, and this training was made available in the US where we have employees in the tens of thousands. But based on the quality of the video, I believe they did or will like it.
I asked chatgpt to come up with an easier way to remember that. T - Task: Specify the action or outcome you want, such as "write," "translate," or "summarize." U - Understanding: Provide context and background details to help shape a relevant response. R - Reference: Offer examples or guidelines to demonstrate the desired type of output. I - Identity: Define the persona or role for the LLM, like an expert, mentor, or character. N - Notation: Describe the desired format for the response, such as bullet points, essay, or code. G - Genre: Indicate the tone or style, such as formal, casual, persuasive, or friendly.
Team ChatGPT or Bard? Lemme know down below 😉
TIMESTAMPS
00:00 I found the Perfect ChatGPT Formula
00:40 The 6 Prompt Components
01:39 Task
02:07 Context
03:05 Exemplars
04:28 Persona
05:18 Format
06:20 Tone
07:05 Example using the Perfect Prompt Formula
08:01 Good vs. Bad Prompt Outputs
ChatGPT for most tasks, Bard for current events
@@LukeBarousse Same!!
Claude (Antrhop\c) for role play as executive coach
@@Donzell1Dawson Is that from the app Poe?
This awesome video is such a simple necessity THANK YOU. I've been trying to use chat gpt for my d&d games and this makes it so much easier. Please make more videos on in and outs of ai and I'll keep watching 🫡👍
Figuring out the amount of context to give ChatGPT has been my BIGGEST frustration; appreciate how you broke this down
Agreed with you 100% Luke! I used to go ON AND ON and realized there's definitely a point of diminishing returns. Now the 3 questions just acts as a guardrail for me not to waste too much time giving background info!
Same. It's so clear to me this isn't even close to being an "AI" like everyone keeps going on about. I know content creators like Jeff Su and trying to help but the minute I need to use math formulas to get GPT to understand what I want?? Kind of defeats the whole purpose.
And it's almost like GPT "Pro" is dumber than 3.5??? I just upgraded and the actual experience isn't anything like the UA-cam hyptrain is claiming it is.
@@Reclaimer77no, you just have to learn to talk to it. GPT-4 is absolutely amazing.
@@alpha0xide9 Well after a LOT of frustration I've apparently coded my first ever Python Windows app using ChatGPT. Soooo, I guess maybe I was overly harsh.
I will say it takes a LOT of prodding and rephrasing stuff. But considering I haven't seriously used a CMD prompt for writing code since my first Tandy T-1000 (showing my age)....I'm kinda stoked lol
Same here Luke! It helps to imagine what the output would look like and try to use that when figuring out the amount of context.
Wasted hours on prompt engineering courses on youtube and did not get anything i didn't know but your 8 minute video teach way more than any other video on UA-cam. Keep up the good work. 👍
Hahaha that's what I love to hear 😁
I'm going to have to agree. This was very helpful, sir.
loved it
Jeffsu be like , huzzah .. chatgpt generated it for me. Lol ! Nah, great info , thanks Jeffsu !
The best video I ever watched on Prompt Engineering, Fed up with so many videos watching hours of content to find nothing at the end. Just in less than 10 minutes, You conveyed everything. I'm sold at 3rd minute itself. Best example of technical content. Sub'd now
Woohooo thank you my friend 😁
🎯 Key Takeaways for quick navigation:
00:00 🚀 *The key formula for ChatGPT prompts involves six components: task, context, exemplars, persona, format, and tone.*
00:59 🧩 *The order of importance for the six components: task sets the goal, context adds relevant info, exemplars improve quality, persona shapes the responder, format structures output, and tone defines the writing style.*
01:57 📝 *Task component: Start with an action verb, clearly state the goal, whether simple or complex, to direct the output effectively.*
02:27 🌍 *Context component: Provide enough background info for ChatGPT, answering who, what success looks like, and the user's environment to narrow down possibilities.*
03:25 📚 *Exemplars component: Utilizing examples or frameworks significantly improves output quality across various tasks like resumes, interview questions, or job descriptions.*
04:51 👤 *Persona component: Define who ChatGPT should emulate, like a therapist for workout advice, a recruiter for job-seeking prompts, or even fictional characters for creative fun.*
05:48 📊 *Format component: Visualize the desired output format, be it tables, emails, bullet points, code blocks, or paragraphs to suit the task efficiently.*
06:46 🗣️ *Tone component: Specify the desired tone, whether casual, formal, enthusiastic, or specific feelings, guiding ChatGPT to match the writing style.*
Made with HARPA AI
This deserves to be pinned.
THANK YOU:):)
Came here for this since I don't have my laptop on me rn, only my phone. Appreciate it
I love that you talk fast & get to the point quickly. Your use of humor & speed kept my attention and I appreciate this information! Thank you!
Thank you!!! 😁
I love it too.
TBH, I checked the speed that i might be running the video in 1.5X. //Good Job PCT-FET the formal is great.
Feel the same here: clear, concise, useful and friendly yet professional tone. Keep on the good work!
I love that you talk fast & get to the point quickly. Your use of humor & speed kept my attention and I appreciate this information! Thank you!❤❤❤
You're welcome!
Gemini AI: Jeff Su's advice is a good introduction to prompt engineering. It provides a solid foundation and some useful techniques. However, remember that prompting is an art, not a science. The best way to improve your prompts is to practice, experiment, and learn what works best for you and the specific task at hand.
Hahahaha that's good!
I seriously love this, especially the "tell it the feeling you want so that it gives you 5 tone keywords you can include in a prompt for ChatGPT". I saw one somewhere where you can ask it to play devil's advocate so you can see debatable opinions/outputs for either decision making or stuff like that. The options are endless, which I absolutely love. Thank you! Also, I noticed that the examples (black screen with moving white dots and then the copy over it) are on the screen longer. Thank you!! Not sure why but it helps me take in and absorb the info easier. Love it!!
No problem! Glad you found this helpful 😁
I had a previous thought about the dark side of ChatGPT. Our brand recently faced a social media crisis, and when I turned to ChatGPT to prepare some seeding comments to defend our brand, my mind went up with an idea: "What if I generate negative comments to influence public opinion," using the context directly from our first post?
The results blew off my mind. They were exactly the same negative comments I'd seen on our previous post.
Hopefully, technology goes for the right thing for our generation.
The simplest solution is to ask ChatGPT what prompts are required to do a certain task. You'll never need UA-cam again!
LOL that's a good point!
Bro It workss❤😂 you deserve million likes
This would make more sense if it understood what it says but as an LLM won't necessarily give you a good answer
You right dude!
Haha.. but I needed the video to find the comment
Jeff is a super instructor. He presents detailed content in an easy to understand way, and makes it fun and approachable!
Thanks 😁
I wanted to learn prompt engineering so I checked lots of resources on the net, but none of them were as useful and helpful as this video. Thank you so much for creating it.
No problem! 😁
Best guy I've found on youtube so far - he gets straight to the point, no fluff, love it!
It's because I hate wasting your time 😁
Thank you sir😊
It's my first time using Chatgpt and this video explains in a precise and concise way how it functions.
Cheers from Cameroon 🇨🇲
That's what I love to hear! I've never been to Cameroon but I'd love to check it out sometime!
This has just been the best should I say the shortest yet impactful course, that has greatly enhanced my understanding of prompt engineering and how to effectively use ChatGPT . Thanks a bunch Jeff
Thanks 😁
Hey Heff Su. Really appreciate you being so professional and delivering top-notch work. Keep rocking it!
Thanks, will do! 😁
role ai, detail nya, question ke kita trs balikin, + role ai
context: bakground kita, target, lingkungn/kondisi
task: apa yg mau dilakuin oleh AI ny, bisa lebih dr 1 dalam 1 kalimat
exemplar: contoh/format jawaban, misal per point apa
persona: role ai nya
format: paragraf, essay, bullet point, email
tone: expresi, mood penjelas
Almost 3 minutes in and there's so much value already! More power to you, Jeff! Thank you!
No problem my friend 😁
Now I watched this video. Then I copied the transcript. I asked ChatGPT to rewrite it in a proper format and write the promts I wanted according to this training. The result is beyond perfect. 👍
Awesome!!!
How did you download the transcript? I was looking for that option myself.
UA-cam provides the transcript 😁@@BritHull
Hey Jeff, your breakdown of the prompt components for mastering ChatGPT and Google Bard is really insightful and actionable. It's clear that you've put a lot of effort into understanding and sharing these essential tips for getting the most out of these tools. 👏
I'm curious about the balance between providing enough context and overloading the prompt with unnecessary information. In your experience, have you found that there's a sweet spot when it comes to context? How do you avoid going too far and including details that might not actually contribute to the quality of the output? Thanks for sharing your expertise! 😊
Using 1-2 sentences to answer the 3 questions I talk about in the video 😁
I agree with @dotpenji! Thanks for the tips! My technique was just being specific about the topic and context you want the text to be about.
I agree with @dotpenji's sentiment! Thank you for the helpful tips. Previously, I focused on narrowing down the topic and context before writing.
Unlock the ideal ChatGPT prompt formula in an 8-minute masterclass. Invaluable insights!
@@DenverOrtiz nice!!
Thank you so much!!! I love how you're able to simplify information into easy to follow steps without the unnecessary fluff others add...you've earned a new subscriber!
Yay welcome 😁
The way you break down complex topics, like ChatGPT prompts, in such a simplified manner is truly impressive! Congrats!
Btw, Am I the only one who asks please and say thank you ChatGPT? Idk if AI comes to life, at least they know I was polite to them...
I think the same Luis 😂
I do too, just out of true appreciation and amazement, not really fearful of the day they awaken really 😅
I wished I had found this video early in my career. Its truly a value worth hundreds of hours of research. Great work there Jeff!
Thank you my friend 😁
With your concise and fast way of teaching I almost felt like I was sitting in front of you with my notepad. Loved and Subscribed!
Woohoo welcome 😁
Thanks for the video, Jeff. Concise and useful tips as always.
It looks like ChatGPT can use strikethrough text. So when using it for correcting texts (6:15), you can ask it to strikethrough the corrected words instead of deleting them, so that you don't have to hop between the corrected version and the original version.
That's a great point!!
Yes, as it uses MarkDown or HTML, one can ask it to use all kinds of formatting. You can even set the custom instructions to use a specific version of MarkDown, useful if you are using a note taking tool like Obsidian, Notion etc. that might be using their own versions of it.
Another good prompt format that I use is to divide the prompt into labelled sections.
Eg:
1.
What is the capital of Brazil?
2.
How far is the answer from 1. (or you can just say "it") from the Amazon River?
3a.
Whats 300 × 11?
3b.
Minus 150 from the answer
This way for output (I hate using that word with ai) you would get one long answer to every question that you might be wondering filling in all blanks. Pro tip: as you read the 'output' set the coursor in the input field as if youre about to type something so that the keyboard appears. Then as you read using the remaining vertical space, write down every question that arises in this format. Perfect for research.
Interesting!
Your formula simplifies a complex process. Also, kudos on the clarity and energy in your videos - it really helps.
Glad you found this helpful my friend!
Code Signal uses a very similar structure to yours:
Ask, Context, Constraints, Example and Style.
They call their system Markup Prompting Framework.
Surely a Persona would help even more, so great video!
Nice! and you're welcome 😁
Excellent presentation. I found this out too. The key components here give you a higher quality result. I've found that breaking it down is helpful. I will ask if it understands or sometimes I will give it context, then an exemplar and ask it to respond "read" and only "read" so I don't use up memory. That allows me to load in content. I've found it does better with preloading it with context and exemplars then giving it a task after that.
That's a great tip, thanks for sharing!!
The best GPT prompt guide ever found...! Watched this video for a dozen times, recommended ! Thanks Jeff.
Thank you!!!
I've definitely come back to reference this video several times. It has a great breakdown and provides examples - which is my learning style. Thank you for sharing this information.
No problem 😁
Has to be the best videos i have seen on prompting.
And extra points for not filling the screen with a mic (what's with all the big microphones in front of everyone' face?).
With just under 40 years in IT and have been creating similar prompts but i know most people do not.
One thing i did not see, and which is valuable, is you did not specify the reading level.
Thanks my friend 😁
Woww exactly what i need. I genuinely want to make use of ChatGpt but when it comes to promting, it's like the messs. Thankful that you figure it out for me 🤩🥳
No problem! I got another ChatGPT video coming out on Tuesday as well 😁
This is very streamlined and easy to follow. I will pass this video along to associates who are having a hard time navigating chatGPT
I hope they find it helpful 😁
Really great value in such a short video, Jeff! Context has always been the trickiest part for me but I've found that sometimes less can be more in the initial prompt. Additional context can always be fed in later on using the conversational interface :3
Also, congrats on 500k! I'm so happy to see how far your channel has come along 😄 I'm gonna make a prediction and say you'll reach 1mil by the end of the year!
Agreed Andrea! Sometimes I like to follow up as well. I'm basically trying to find that sweet spot where I can get 80% of the way there with the initial prompt, then fine-tune as needed 😁
I was doing this with prompts without even knowing this method. Now that I know it, I can actually understand why my prompts don't give me what I want and how to fix it. Before, I'd just try to add more information when I wanted to fix my output, but I'd get results that are going in a completely different direction than what I expect. 😄
Nice!!! That's awesome 😁
Love this format! Thanks for the work to put this together, Jeff. This is the kind of ChatGPT help I’ve been looking for ❤
It took a while but I'm glad you found this helpful!!!
Got new job that needs me to utilized ChatGPT. Stumble upon this video for self-learning and this is so helpful! Thank you!
You're welcome 😁
This is awesome Jeff. I’m glad I came across your video on ChatGPT prompts. It can be discouraging spending time to write the perfect prompt. Thank you!
No problem, glad you found this helpful 😁
I just finished this video and I really want to thank you for this simplified explications I was struggling finding courses I even enrolled in one and it wasn't the best experience ever I even doubted my learning skills I thank you again it was so helpful and clear
No problem 😁
Jeff, really love your videos! They're always filled with good and actionable info, and most importantly no bs. 😉
It's because I value my viewers' time!! 😁
I think one thing that has helped me (as a prompt engineer on the side) is a principle I learned while doing research in physics - set your boundary conditions. This is implicit in setting an environment, but I think worth stating explicitly. If you can tell GPT what the boundaries are on some context/environment X, then it'll give you an even better response.
That's a great point!! 😁
Multitask request looks similar to SQL prompt.
Great educational video!
Thank you 😁
This video has really saved me from further distress and torment from the lack of knowledge on AI, struggling to fit into a world of tech and AI in old age. Thank you so much, Mr Su!
No problem 😁
What a great video. I’m a newbie to ChatGPT. This was simple and almost immediately useable.
Glad to hear it my friend 😁
Stumbled upon this a few months back and come to find my company is using the same video in their genAI training. Man I hope you get compensated somehow for that. 75k employees have to watch this video apart of mandatory training
Wow I'm honored, thanks for letting me know!!
Best and most succinct prompt engineering guide on the Internet, thanks a ton @JeffSu !
Thank you my friend 😁
Jeff, you've truly cracked the code for crafting the ultimate ChatGPT prompts! 🚀🔥 Your breakdown of the six key components is like a treasure map to unlock AI's potential. The 'Context' section hit home - finding that balance between too much and too little info can be a head-scratcher. And kudos for bringing up the persona aspect; it's like giving ChatGPT a tailored suit for each task! 🎩🤖 Your videos keep elevating my AI game, and for that, a double thumbs-up from me. Can't wait to see what you dive into next!
Was this....written by ChatGPT? 😂
Sure was@@JeffSu :). Keep up the great videos. Love them!
OUCH! My brain does not work at your speech output speed, I had to slow you down to 75% so I could keep up! Great ideas and pro tips! I am aiming to use AI to create language code for a trading bot based upon my specific trading technique that works incredibly well manually! The "Task" is to automate what I do in the trading world so I don't spend all my time trading! My concepts are all Smart Money focused! I am sure what I have learned from you will help me, since prompts are key! Thank You!
No problem!
I really like your tip on tone and also that was a great tip to make changes listed in bold.
Thanks 😁
Almost a year-old video and still... increidbly relevant! Thank you very much, it changed my way of interacting with LLMs!
Thanks 😁
For me, the most useful video I have come across after a while: detailed, easy to understand and compact. Thanks, Jeff!!!!
That's what I love to hear my friend, you're very welcome 😁
@@JeffSu Waiting for your next advanced level video, Jeff. Again, brilliant work!!!
Usually dont comment and thought this would be clickbait but you definitively came through.
Thank you for putting the time in to make actual valuable content, this is golden, im absolutely subing, liking, sharing etc.
Cheers from Mexico City, again thank you.
Woohoooo thank you my friend! Glad I proved to you that I never make clickbait videos haha
Hey Jeff, have been following you for a while and have benefitted a lot from the LinkedIn videos! Thanks!🙌🏻
Thank you my friend! Hope you benefited from the other videos too 😉
@@JeffSu Absolutely!
1. Task : Start with an action verb.
2. Context: User background, what success looks like, and environment or conditions of user.
3. Examples: Provide examples of the best or real results. It learns very fast through examples.
4. Persona: Act like a ______. (Also add person famous for that).
5. Format: Tell the AI the right output format you want the answer in.
6: Tone: Specify the desired tone, whether casual, formal, enthusiastic, or specific feelings, guiding ChatGPT to match the writing style.
Great summary!
the most often component I use is hands-down TONE = friendly business tone! 😄😄
Interesting! How is that working for you?
It's amazing how you were able to break this down so that viewers are able to easily understand, and I love this haha. Helped me out a lot! Gonna use this immediately!💯
Let me know how it goes! 😁
To me this is told too fast to make sense from someone who is trying to learn. It becomes uninteresting.
Thanks for the feedback!@@MR-vj8dn
Super helpful. Excellent content. One of the best videos on the channel so far (hard to choose)
Hahaha thank you! I'll try to make even better ones 😁
great video! it summarizes a lot of things i remember learning so far about prompts. here's a mnemonic that i find helpful to remember all 6 key words quickly, before trimming/adjusting case-by-case: "PerConTa, ExForTo" (the first syllables of each key word). then i come back to this video to review further tips for each part.
Nice!!!
I have low interest in doing these things, however, it would greatly benefit me and my work output if I took advantage of AI. Thanks for your energy and succinctly explaining prompts for noobs like me. 🙏
No problem! And you're a noob no more 😂
Lots of gold in this video and very well presented. Thanks man.
You're very welcome my friend 😁
Followed the example prompt to write an email for asking who wants to be in a beta program for my new POTS app. Gave me a really great first draft. Thanks!
Nice!!! Which ChatGPT version did you use?
Gpt 4. I also used your prompt to write a self review for work. Really helpful. It was super clean. Thanks again for the video. It gave me the kick I needed to start using ChatGPT.
Hi Sir,
Your video was amazing! I’m a fan of yours. Could you please make a video on content creation? I’m interested in learning how to use chat prompts for writing, such as blog posts, product descriptions, and social media posts. I am eagerly awaiting your next video.
I might do that down the line my friend 😁
ChatGPT + Informative YT channels like yours has been life changing !
Agreed 😁
🎯 Key Takeaways for quick navigation:
Task: This is the most important part, tell the AI what you want it to do (write, generate, analyze).
Context: Give relevant background info to guide the AI towards the desired outcome. Ask yourself: Who? What? Why? When? Where?
✨ Exemplars: Provide examples to improve the quality of the output.
Persona: Imagine the AI acting as a specific character (e.g., business coach, writer).
Format: Specify the format of the output (e.g., email, code, bullet points).
Tone: Set the desired tone of the AI's response (e.g., serious, funny, formal).
✅ Not all required: You don't need all 6 components for every prompt, just include what's relevant.
Start with a checklist: Use this formula as a guide to ensure your prompts are clear and effective.
Context: Focus on 3 things: user background, success definition, and environment.
⏱️ Example 1: Include background info but prioritize only what's relevant to the task.
✨ Example 2: Show a bad bullet point and then ask the AI to rewrite it in a better format.
Example 3: Provide a framework (STAR method) for the AI to follow when answering an interview question.
Exemplars: Not required for every prompt, but helpful to improve output quality.
Persona: Imagine the AI acting as a specific helpful person (e.g., consultant, coach).
Persona: Imagine the AI as a helpful expert (e.g., therapist, recruiter, marketer).
- ♀️ Fictional characters work too (e.g., Batman).
- ️ Format: Close your eyes and visualize your desired output format (e.g., table, email, code).
- Common formats: Bullet points, emails, code blocks.
- Most useful (professional): Paragraphs and markdown with headings.
- ✏️ Proofreading tip: Ask the AI to bold all changes for easy review.
- Tone: Use simple terms (casual/formal), emotions (witty/enthusiastic/pessimistic).
- Pro tip: Tell the AI the desired feeling (e.g., serious but not arrogant).
Combine all 6 components: Use all the prompt building blocks (task, context, persona, format, exemplars, tone) for a strong output.
- Example email structure: Break down your desired email format into clear sections (e.g., tl;dr, background, results, thanks).
- Clear & concise language: Instruct the AI to use easy-to-understand and professional language.
- Confident & friendly tone: Aim for a balanced tone that sounds both authoritative and approachable.
- Reference existing email (optional): If you have a similar email, include it for the AI to copy the format.
- Simpler vs. detailed prompts: A more detailed prompt leads to better, less generic results.
Made with HARPA AI
Hey Jeff, your breakdown of the prompt components for mastering ChatGPT and Google Bard is really insightful and actionable. It's clear that you've put a lot of effort into understanding and sharing these essential tips for getting the most out of these tools.
I'm curious about the balance between providing enough context and overloading the prompt with unnecessary information. In your experience, have you found that there's a sweet spot when it comes to context? How do you avoid going too far and including details that might not actually contribute to the quality of the output? Thanks for sharing your expertise!
Wait but I answer that in the video 😂
Thank you for the effort you put into this. Very helpful!
No problem!! 😁
Brilliant video! As a time management coach, I have been using AI extensively and it has turbocharged my productivity. But I was sure I was only scratching the surface of what is possible and had zeroed in on learning how to use prompts more effectively. Your video was exactly what I was looking for. Liked and Subscribed and looking forward to seeing more of your videos.
Woohooo thank you my friend! You're going to love my next video on more intermediate tips and tricks then 😁
Fun fact: This content was made using results from ChatGPT😂😂
Shhhhhh
Perfect, gp
That’s what I’m saying 😂
Why are you here then if you knew it?
XD
Man this is Gold, I am blown away by how far AI has come. This stuff is actually really useful
I know right??
@@JeffSu I just recently started to dive into the whole AI stuff. At first I thought it´s a typical new hype once again (which it is in many cases) and I also thought that, it´s not as useful to me (mech. engineer) but man this is super useful. I needed a short explanation like yours to get into it easier.
@@sierraecho884 So glad to hear that!!
Outstanding - thanks
You're very welcome 😁
Me and ChatGPT as a human what we do is, we delve into deeper thinkings by sharing what's in our mind and then make theories (unapproved but really mind bending and almost precise).....fr these stuffs got me upppp! In psychological perspectives (well I am a vegan, introvert)
thanks very useful.✌
No problem my friend 😁
Amazing video! Very simplified explanation and yet still provides detailed understanding of the subject matter.
Thanks 😁
Next video on custom instructions prompt
Yea I need to figure that one out too haha
Elaborate, please?
I've only been using it for a short time, I quite like it... Up to late I have been interacting as so:
tell it where you are, tell it what you want.. ask for examples(pseudo-code/tables), pros/cons and alternatives.
"In unity, what is XYZ? Please provide examples, alternatives and pros/cons for both."
when writing responses I repeat myself to make sure the darn thing NEVER side tracks which leads to conflicting statements.
"If I am in unity and would like to format your explanation into a UML graph, how would that look?"
"Can you please give me an more complex example of XYZ including a more comprehensive UML graph"
"Can you please give me an even more complex example of XYZ including an even more comprehensive UML graph?"
"In a unified system, what if I wanted to add ABC to XYZ in unity, how could I go about that?"
also never explore complex fields you don't already have an understanding on... This is going to become even more important once we start trusting the responses because they sound good and provide sources, or in such a time as where these have been battle tested further.
I find that if you can keep can "AI" contained within the context YOU desire by repeating yourself at every reply, then you can have conversations with it indefinitely without it starting to contradict itself which happens a lot when working with complex well documented areas
Thanks for sharing 😁!
The master prompter is the one who prompts GPT for a prompt 😂😂
I should add "Master Prompter" on linkedin 😏
Good information. Love Jeff Su's enthusiastic personality.
Thank you so much 😁
@@JeffSu Don't mention it, you're terrific!
Imagine You are watching this at 2x 🤣🤣
Did you try?
@@JeffSu YES
I did 😅
Thanks for the simple explanation, I used it straight away and got very good results.
Awesome!
This video is so easy to understand and actually use as a guide when utilizing the Gen AI tools. Thank you so much💪
No problem 😁
I don't mind spending 16 minutes instead and actually understanding what you're saying...
Thanks for the feedback!
I put the video on 0.75 speed
@@nathaliebouvard7335that’s exactly what my middle aged self had to do 😂
@@nathaliebouvard7335
I put it in 2x......
Just be quicker …
Loved your clarity! I wanted very much to be signed up to your newsletter and do not mind your modest fee! Sorry to post here, I dont Instagram. But I did not foresee the One Dollar transaction that was coming. My bank flagged it with a notification, and I annulled the subscription. Now I must wait the ten days for a new ATM card. A wildly larger memorable and humorous one-dollar-ahead-warning would be so great. My eyesight is a challenge with the cataracts and their crossing guards mucking up the words I want to read. I'll subscribe again in ten days.
That's fine! Thanks for letting me know 😁
Hello sir, can you please explain this with easy English langue's so that we can understand each words clearly & we can take some profit i really like this video but am not able to undrstand cuz of taught English.
Yes please Sir.... Do this it will be helpful for so many people.
Wait what 😅
@@JeffSu mean to say you can make this video simple English language so it understand easily..
This me I write this from my desktop.
This is awesome Jeff. I’m glad I came across your video on ChatGPT prompts. It can be discouraging spending time to write the perfect prompt. Thank you!
Thank you 😁
"The perfect way to take credit for skills you lack." 🤬
Why not 😂
Thank you for your video. However, I didn't understand when you said at first that it's better to write the task in the beginning, and later you gave an example showing the task not in the first position in the prompt
It doesn't have to be in the beginning. It's just the most important component
"Exemplar is just a fancy way of saying examples!" Not true. It is the best of the examples. Your videos seem pretty good but why undermine their value by miseducating others?
Thanks for the feedback Jake 😁! I guess I was trying to find the simplest way to explain it (since arguably the "best of the examples" is very subjective)
@@JeffSu An exemplar is by definition singular, and not plural as you indicate. There is no obligation to be objective, which is a gratuitous, arbitrary “requirement,” even though it distorts what the word means. Why don’t we let it drop, lest the hole gets deeper? 😉
@@Jake7706 sure!
Watching the video again for the 5th time and always find it interesting. You did such a great job, Man! Thanks.
No problem 😁! You'll probably like the video I'm dropping next week on AI as well!
@@JeffSu I look forward to watching it! Thanks Jeff.
you are a great and knowledgable teacher but you are very fast.. please be a bit slow.. please..
Thanks for the feedback! 😁
Hello! You can change the speed of the video. Click on the gear ⚙️ in the upper right hand corner of the video to change the settings. Then click playback speed and normal to change to something slower. 😊
Prompting really is the game changer! Thank you
You're welcome!
LOL, that guy's so full of shit. He didn't spend "hundreds of hours" studying courses on prompt. Ridiculous.
Dam you caught me. I totally made all this up 😂
This is the best video on prompt engg - thanks Jeff!
That's what I think too 😂, thank you!
p.s. Part 2, I asked ChatGPT to put the content of your video in a numbered list of all the actionable suggestions from your video. I now have it all in one place. I want to add images from video, but I'll try that prompt later.
1. Understand the Six Components of a Good Prompt
Task: Always start with an action verb and clearly articulate your end goal.
Context: Include relevant background information to give enough detail without overwhelming the model.
Exemplars: Provide examples or frameworks to improve the quality of the output.
Persona: Define who you want ChatGPT or Bard to emulate (e.g., a recruiter, physical therapist).
Format: Visualize the desired end result and specify the format (e.g., bullet points, tables).
Tone: Specify the tone of voice to match the desired output (e.g., casual, formal).
2. Start Prompts with Action Verbs
Use verbs like "generate," "give," "write," "analyze," etc., to clearly define the task.
Example: "Generate a three-month training program."
3. Include Relevant Context
Ask yourself:
What’s the user’s background?
What does success look like?
What environment are they in?
Example: "I am a 70kg male looking to put on five kilograms of muscle mass over the next three months."
4. Use Exemplars to Improve Output Quality
Provide examples or frameworks that ChatGPT can follow.
Example: "Rewrite this bullet point using the structure: I accomplished X by Y, resulting in Z."
5. Define a Persona
Specify who you want the AI to emulate.
Example: "You are a senior product marketing manager at Apple."
6. Visualize and Specify the Desired Format
Specify the format in which you want the output.
Example: "Output a table with headers: Original feedback, Team responsible, and Priority."
7. Specify the Tone
Indicate the tone you want the output to have.
Example: "Use clear and concise language and write in a confident yet friendly tone."
8. Combine All Components in a Comprehensive Prompt
Example: "You are a senior product marketing manager at Apple, and you have just unveiled the latest Apple product in collaboration with Tesla, receiving 12,000 pre-orders. Write an email to your boss, Tim Cook, sharing this positive news. The email should include a TL;DR section, project background, business results, and a thank you note to the product and engineering teams. Use clear and concise language and a confident yet friendly tone."
learning how to use chat gpt reminds me on when internet was new and you needed to know what to type in google to find what you were looking for something that is second nature now but in that time was all new
Hahaha good point!
I had started getting doubts and stopped relying on chat but thanks to you Jeff. Am now going to accomplish my tasks with Chat efficiently
No problem, let me know know how that works for you 😁
Hi jeff to be honest one of the best videos in the internet for prompt engineering is your thank you please do share this kind of content
Woohooo thank you!! 😁
Hey... The frim I work for is heavy into AI and has already released their internal chat UI in beta and along with this we (everyone) are required to do AI training. And your video here was added to the curriculum. Not sure if you knew this or not, but I thought I might share. Well long short... I have subscribed to your challenge since this video was supper informative, and your delivery has what I could only call energy in it. Loved it, keep doing it.
Wow really?? Thanks for letting me know!!! What do your colleagues think?
@JeffSu Not sure about the thoughts from my colleagues, only because we're a large global firm, and this training was made available in the US where we have employees in the tens of thousands. But based on the quality of the video, I believe they did or will like it.
Thank you!!@@jam4madttm
@@JeffSu yw :)
I asked chatgpt to come up with an easier way to remember that.
T - Task: Specify the action or outcome you want, such as "write," "translate," or "summarize."
U - Understanding: Provide context and background details to help shape a relevant response.
R - Reference: Offer examples or guidelines to demonstrate the desired type of output.
I - Identity: Define the persona or role for the LLM, like an expert, mentor, or character.
N - Notation: Describe the desired format for the response, such as bullet points, essay, or code.
G - Genre: Indicate the tone or style, such as formal, casual, persuasive, or friendly.
Love that it spells out TURING (I'm assuming you did that on purpose in honor of Alan turing?)
@@JeffSu I asked it to base the acronym on something tech related.
Thank you for this simple formula. It's something that I can really build upon in the work that I do with ChatGPT.
No problem 😁