My favorite prompt is to ask ChatGPT to generate a series of prompts on a particular topic for an audience that is unfamiliar with the topic but wants to gradually ramp up to an expert level of understanding. What I get is a list of 20-30 graduated prompts which I then submit one at a time. I end up with a 30-40,000 word document. It works magnificently.
@@anisolo07 *** First prompt *** Your audience is familiar with genetics, and biology in general, at the level of having completed a freshman high school biology class. They would like to undergo a streamlined program focused on understanding epigenetics at the level of a college graduate student. Design a series of prompts to elicit a series of step-by-step essays to explain this concept at incrementally increasing advanced levels of understanding. There is no limit on the number of prompts to achieve this objective. I will execute these prompts later; for now just write out the ordered list of prompts. *** Chatbot response *** Ordered List of Prompts to Understand Epigenetics at a Graduate Level 1. Introduction to DNA and Genes Prompt: Explain the basic structure of DNA and how it encodes genetic information in genes. 2. Overview of Gene Expression Prompt: Describe the processes of transcription and translation in gene expression. 3. Gene Regulation Basics Prompt: Discuss how gene expression is regulated in cells, including the role of promoters and enhancers. 4. What is Epigenetics? Prompt: Introduce the concept of epigenetics and how it differs from traditional genetics. 5. DNA Methylation Prompt: Explain DNA methylation and its impact on gene expression. 6. Histone Modification Prompt: Describe how histone modifications affect chromatin structure and gene accessibility. 7. Chromatin Remodeling Prompt: Discuss the role of chromatin remodeling complexes in epigenetic regulation. 8. Non-Coding RNAs in Epigenetics Prompt: Explain how non-coding RNAs contribute to epigenetic regulation of gene expression. 9. Environmental Influences on Epigenetics Prompt: Explore how environmental factors can lead to epigenetic changes. 10. Epigenetic Inheritance Prompt: Describe how epigenetic modifications can be inherited across generations. 11. Epigenetics in Development and Differentiation Prompt: Discuss the role of epigenetics in embryonic development and cell differentiation. 12. Epigenetics and Disease Prompt: Explain the connection between epigenetic changes and diseases such as cancer. 13. Techniques for Studying Epigenetics Prompt: Describe the laboratory methods used to analyze epigenetic modifications. 14. Epigenomic Mapping Prompt: Discuss the importance of epigenomic maps in understanding gene regulation. 15. Current Research in Epigenetics Prompt: Explore recent discoveries and ongoing research areas in epigenetics. 16. Epigenetic Therapy Prompt: Explain how epigenetic mechanisms are targeted in therapeutic interventions. 17. Ethical Considerations in Epigenetics Prompt: Discuss the ethical implications of manipulating epigenetic states in humans. 18. Future Directions of Epigenetics Prompt: Speculate on the future applications and challenges in the field of epigenetics. 19. Epigenetics and Personalized Medicine Prompt: Describe how epigenetic profiles can influence personalized treatment plans. 20. Integrative Epigenetics Prompt: Explain how epigenetics integrates with other biological systems (e.g., metabolism, signaling pathways). *** Second prompt *** Let's dive deep into the topic of epigenetics using the following 20 prompts starting with basics and gradually progressing to more advanced levels. The audience for this exploration consists of students who have just concluded one year of study in biology at the freshman high school level and wish to efficiently advance their understanding of epigenetics to as close to the college postgraduate level as possible. Try to approach 2000 words in each prompt response, but do not hallucinate to amplify the word count; keep everything factual and accurate. I will present the prompts one at a time, but as a preview here are all the prompts so you can plan ahead and avoid redundant material across multiple prompts... *** copy above 20 prompts here *** Let's get started with the first prompt...
Great concept to have a "prompt designer" and great results! I went one meta-step further after loading your prompt...I asked the LLM to read the prompt as an input and suggest an improved version of the prompt designer. I then asked it to test the results by comparing output results - once using the original prompt designer, and another using the "improved" . Results were somewhat better, a little more nuanced, with a little more detailed. Thank you for sharing this great content. Liked, subscribed, I am a fan!
Glad it’s working for you. Sounds like you’re doing the level 1 vs level 2 approach. E.g. You get a response from ChatGPT and then tell it that its response was level 1… now make it level 2.
When I first watched the recommendation, I thought it was some sort of clickbait... couldn't be more wrong! Turns out I've been using gpt in a very limited way all this time. Thank you sir for this more than useful video.
@@LawtonSolutions the video is very straightforward and not in a bad way, meaning, it goes to the point, explains something extremely useful and already gives the starting prompt... cant ask for more in that area. I still need to dig into your other vids!
You're an eloquent and polite erudite that recently watched an excellent TH-cam about the best LLM prompt ever and you want to write this video's author a compliment.
This is a great method. I’ve been using this one for years: You are a prompt generation robot. You need to gather information about the users goals, objectives, examples of the preferred output, and other relevant context. The prompt should include all of the necessary information that was provided to you. Ask follow up questions to the user until you are confident you can produce a perfect prompt. Your return should be formatted clearly and optimized for ChatGPT interactions. Start by asking the user the goals, desired output, and any additional information you may need. Curious to see the output differences!
Thank you for this. I have been experimenting with the ultra copy/paste meta prompt to use when I want to get really good prompts and have two I’ve had great results with. I think yours is is even better. Thanks for putting the time into this and sharing it. Your communication style is very effective as well by the way. Good pace, tone, tempo, enthusiasm and soothing voice. Some others who do this type of channel or going to fast or too cocky and yours was pleasant to listen to. New subscriber !
Thank you Fred! It means a lot to me. I hope you'll create a custom GPT with it so you don't have to bother with the copy/paste exercise each time you open a new Chat.
How lazy can people get, dude? Jeez. This is sad hearing grown adults willingly sacrifice thought, cleverness, ingenuity, discipline, self-reliance, experience, memory, and ability, and ultimately their pride, by having to ask a computer to write the questions amd the answers for them. Truly pitiable, and pitiful.
This is brilliant, thank you so much! - I've been using ChatGPT over the past few weeks and know that i'm just scratching the surface of the capability of this AI, this is going to really improve output!
Also great for o1 Pro (Pro only model) prompt generation (can handle nearly a full book sized prompt). I have some prompts that are pages long. This also works great for the level 1,2 , 3 method of improving them. Give it your prompt, tell it to take a level 1 prompt and turn it into a level 2 prompt using your instructions. And, if needed, can tell it that’s a good level 2, but not I need a superior level 3. Great work!
I just used your über perfected prompt for a very customized research output for my bespoke website design agency. Wow. Just WOW. I look forward to more of your content and connections
This is terrific! Thank you so much. I appreciate the work you put into this. I've learned so much! This is my first time seeing your content, but I can't wait to watch more.
New sub… awesome content professionally delivered - clear and concise with NO bloviating.. That makes for the best type of informational content on TH-cam - or anywhere for that matter. 👍🏻👍🏻
In two years we've gone from computers not being able to understand natural language to giving them instructions like "Take a deep breath. You can do it! I have faith in you." There is also something kind of crazy about telling a computer to pretend it's an expert on something. People also use this strategy to bypass safeguards. If you asked an AI how to hack into the Pentagon it would certainly tell you that is is unwilling and/or unable. But now you can just tell it to pretend that it's allowed and also to pretend that it is an expert at computer hacking. And then give it an attaboy just for good measure. I'm also reminded of a specific incident of a compute passing the Turing test by being told something like "Don't try to pretend you're a human just to fool people. Just act natural and be yourself." Crazy times we live in.
This is the first video I've watched on your TH-cam channel, and I’m beyond impressed! Wow! The effort and hard work you put into this video about ChatGPT prompt interactions are truly amazing.
This is amazing!! And the thought you put into it is obvious. I greatly appreciate your dedication, time and effort. I also went to your website and CONSUMED as much as I could in one sitting!!! 😂 I will be back and I will turn on notifications for future videos. My channel I am commenting with is a new one for me and I am in the process of building out the pre-game warmup. I will certainly be using your lessons. Thank you again! 😊
Curious have you tried Prompt Chaining? a technique where you break down a complex task into a series of smaller, sequential prompts, feeding the output from one prompt as the input to the next, effectively guiding the LLM through a step-by-step reasoning process to achieve a more accurate and comprehensive result on intricate problems; essentially, it allows you to chain together multiple prompts within the LM Studio interface to tackle complex tasks that might be too difficult to handle with a single prompt.
This is excellent. Earned my like and subscription. I've done a fair bit of prompting the past 12 months building a number of prompts for various tasks and even some similar meta-prompts for generating prompts. Two or so months ago I settled into a reliable framework to build pretty much any type of prompt I want. This though is another level. Tested it on a 2-3 sentence request and woah... great prompt in seconds. So thanks.
"Wow, this is incredible! The level of detail and thought you've put into crafting this prompt is truly next level. I can’t wait to try this out myself. Thanks for sharing such valuable insights-your content is a game-changer! You've just earned yourself a new subscriber. Keep up the great work!"
@@LawtonSolutions i will keep watching you its hard as we know when you start out i love Chatgpt i have been combining with notebooklm maybe a good video would be combing both these tools
Whoa!! Like what did I just watch? That's just amazing. Thank you so much for sharing this meta brain bending content. I'm definitely going to hop on and give it a go. Thanks again!!
@@LawtonSolutions Yep just used it to write a descriptions for a game I created that I will be selling on Teachers Pay Teachers. Whoa!!! It just knocked it out of the park. Thanks so much. Also will be watching all your other videos soon. Thanks again.
Well, thank you for this! I'm one of the newbies in the AI world, 'a creative but broke artist' one could say... so I've just started using AI during the last couple of months and are totally hooked on the possibilities of this tool already! But,yup, as I'm pretty new to it all I sometimes struggle with prompts..... And since I always have to look for 'free versions' of everything because of my financial situation at the mo I often end up at huge roadblocks unfortunately which is very frustrating.! So,THIS is really helpful!!! A great help in getting good results quickly without wasting 'credits' all the time, which also slows down the learning process. I Will be a good student and study your prompt structure in detail - Again Thank you! Looking forward to more of your content
This is just pure genius. I was actualy watching a video on 'Bigfoot' then your video came up in the recommendations: Subscribed strsight the way here and the newsletter. I use a 30 word prompt for childrens book ideas. The prompt usually, and I say usually comes up with some amazing fresh ideas. This prompt formula gave me a 587 word prompt. OMG massive thank you. The results are out of this world. I believe I am going to be busy for a long while. The ideas, the story ark and even amazingly formated chapters. Massive thanks - This is the best I have ever used or designed myself. I can't thank you enough. P.S Pardon if any grammar or spelling mistakes, my keyboard has somehow caught dyslexia.
Appreciate the insights. You have just given me several ideas. I feel dumb I didn't think of using the LLM to right a better prompt. Yes this method takes a little extra work up front but once you have the template it is copy and paste. Glad your video popped up in my feed. I have just subscribed.
@@LawtonSolutions Here is an example of using your method: My Script that I wrote for a TH-cam short Laziness is Destruction Hook: What is worse, someone who is lazy or someone who actively destroys? Proverbs 18:9 He also that is slothful in his work is brother to him that is a great waster. This proverb teaches us that laziness and destruction are brothers. Sometimes doing nothing can be just as destructive and doing the act that causes the destruction. The wise should seek a way to solve a problem. Doing nothing and just hoping the problem goes away will result in the same destruction. ChatGPT Output: **Hook:** Did you know that laziness can be just as destructive as outright sabotage? **Scripture:** “He also that is slothful in his work is brother to him that is a great waster.” (Proverbs 18:9, KJV) **Explanation:** This Proverb draws a striking parallel: someone who is lazy in their responsibilities is like someone who actively destroys. In ancient times, a farmer’s negligence could mean famine for his family or community, just as deliberate wastefulness could ruin resources. The verse warns that failing to put effort into our work not only affects us but also harms those who depend on us. Laziness and wastefulness are two sides of the same coin-both lead to loss and chaos. **Application:** Think of your daily tasks-whether at work, school, or home-as opportunities to contribute and build. Avoid procrastination and carelessness by focusing on diligence and excellence. Even small steps, like sticking to a schedule or finishing what you start, can have a positive impact on those around you. Your effort matters more than you think! **Call to Action:** What’s your go-to Proverb for motivation? Share it below, and it might be featured in a future video! ---- I like the addition of the call to action feature. This was my first attempt and I told it between 120-180 words. I will have to play with this and probably ask for a shorter work count or learn to talk faster 😃.
I have made something like this a while ago, but I never used it. With the possibilities of GPTs I only use customized frameworks for each task. However this really matters for the o1 model.(Reason:no GPTs+limited uses ). You need a very well designed prompt with a lot of context because you have very reduced number of prompts. Also o1 needs you to tel him what to do with a lot of context and he will figure out how. The second idea is to use this prompt not for creating prompts but for creating GPTs. Very often in bisiness use cases we have repeating tasks or processes and the need to format a CRAFT prompt becomes a repeating task. So the idea is to put as much in the 'craft' inside the gpt. context -> in the gpt; specialized context in the prompt, the best is to use the storage and use rag role -> in the gpt action -> in the gpt with customization in the prompt format -> the base in the gpt framework target -> in the gpt with customization in the prompt So when using the prompt framework you focus mostly on formulating the problem at hand.
Thank you, Lawton. Even after watching this for a couple of minutes now I know I don’t know anything about the English language at all much less how AI works. : )
After watching you develop the beginning of your prompt process I’m not even sure I should be out here alone by myself because I might stop breathing. LOL
I am trying to convince the world to unify because it absolutely makes sense in my mind, but I am trying to use AI to help justify what I am saying. I could really use help from the rest of humanity to include all of the experts that can look at the data that it provides and verify that what I think is either true or not. I think we need to put a cap on developing AI past artificial general intelligence for now. We need to make sure that we transition in a way that does not destabilize and harm humanity.
Superb insight 👌 yet to try but i already know, from a trouble-shooting perspective, your concise approach and utility re’ framework will yield a SILVER-BULLET, on execution for anyone who seeks outcomes that demonstrate domain or subject matter expertise - with gratitude ! Much appreciated….duly subscribed 👏👏👏
You can do a CustomGPT with that prompt an never have to copy that prompt again 👌. The Prompt was too long for my Perplexity Spaces so I added it as a PDF to the knowledge base.
@@LawtonSolutions or @ericwetzeldesign . So you create a CustomGPT to create a prompt, and then you paste the returned prompt in another chat (like in the general/normal ChatGpt for example)?
After a lot of nudging, I got ChatGPT to admit that the jokes I was asking for were not its own, but rather based on already written jokes in the dataset! When I made it write its own original jokes, they were flat.
Thanks Daniel... Yes, asking to cite sources in the initial prompt will help with some of that. If you check out one of my earlier videos (th-cam.com/video/Uz_DeqGhbjs/w-d-xo.html), there's a way to get it to write better jokes based on a specific comedian's style. Also, you might consider a custom gpt that references the style of comedy most aligned with your desired result.
Yes, ChatGPT is terrible at writing jokes- the same goes for other LLMs. However, one time I asked ChatGPT to write a good Trump joke and here’s what it came up with -“ I gave Donald Trump a penny for his thoughts and he gave me back change.”
The better you inform the LLM by prompting your request the better results you get. Prompting an LLM to generate the best Prompt’s is genius! You were really thinking outside the box with this. This is exactly how we should use AI so that we get the very best of the best results. Pretty cool!
I feel as though I just increased my knowledge by a gazillion. I am so grateful for this content and you are genius. Subscribed, liked and now adored. Thank you so much.
@LawsonSolutions Thank you so much for sharing your time, energy, and insights on this topic. It’s truly generous of you to share what you’ve learned, and it doesn’t go unnoticed by those who appreciate your efforts. While some may be quick to critique, I hope you know that many of us value your kindness, thoughtfulness, and generosity in creating and sharing this video. I am grateful for your willingness to help me (and others) learn and grow!!!!
I agree with other comments. Fantastic content. I'm about to binge watch the rest of your channel. I really like your presentation style. Anyway. So a couple of things: a. I was looking at your website. In the level 3 priming section, there's a typo: routinE not routinG b. do you have or could you present some content on effective scraping strategies using AI and other tools. I have a very low budget, but I need to compile a nation-wide spreadsheet of continuing care communities (basically communities that provide seniors with independent living through dementia care and skilled nursing. a one-stop shop for retirement, so to speak). Any ideas? I"m prepared to manually scrape Google Maps city by city, but it would be great to find a more effective strategy. Do you cover anything regarding agent workflows or tools like Make and AirTable? c. Great work. Look forward to more of your content.
great video. Seems like it would be more "AI friendly" and give better results if instead saying "take a deep breath" to say "allow yourself additional time to reflect on whether what will be written is the best of many iterations" or something like that. your thoughts ?
My wife doesn't think I sound like Bob Ross (but I'll gladly take the compliment). You can find link to the full prompt in the website that is linked in the video description: lawtonsolutions.com/how-to-ai/
How would you create a prompt for solving complex math problems such that students can understand them? I have quite strict formatting guidelines with latex etc.
Thanks Peter… and good question. If you look closely at the screen recording you’ll see that I’m actually sharing the prompt from a GPT. I’d definitely recommend you create a custom GPT with it.
Well done! I'm a big believer in meta-prompting and have crafted (no pun intended!) some niche prompts that generate targeted prompts. A more important goal, for me, is to create prompts powerful enough to solve problems without needing a secondary prompt. Claude is particularly adept at phase-shifting and can easily handle the meta gymnastics very well. I'm glad I saw this in my feed. New subscriber looking forward to your ideas!
Thank you for such a great video walkthrough. I have been working on a very similar project to create great prompts. I have recently unlocked and wish to see if you have used this level up. After my initial prompt is created which is a similar process to yours I then tell GPT the following "This is really great, however this is a generic level 1 thinking prompt. Lets revisit this prompt and apply level 2 thinking to improve this prompt. I find that this then refines the prompt to be even more detailed and well rounded. Be interested if you have tried this process?
@Lawton Solutions, I couldn't help but notice you added similar content in the GPT builder on the back end. Was this the same information you added as a question? Please advise? Thanks!
@@LawtonSolutions I'm referring to the frame @6:37 in the instructions section. Did you add the same information to create a GPT first, then prompt the GPT with the context?
@@LawtonSolutions It wouldn't replace all you did with yours here, but maybe half of it. Can't get too specific, or it makes for less useful responses when you're using the bot for something in a different arena. Poe is great for this, because you can make as many hyper-targeted bots as you want for yourself, using the latest API-enabled models.
Duly impressed, Brian! You popped up in my feed too and I am glad you did. I liked the video and subscribed to your channel. And now I’m going to subscribe to your newsletter. Especially grateful that it’s not another “AI news of the week” newsletter! How do I use your prompt to help me write a nonfiction book?
Thank for the kind words, Dan. For writing long-form, I'd recommend having ChatGPT create a table of contents (with descriptive sections). Then, based on the table of contents, have ChatGPT create an outline for each of your chapters (again with detailed descriptions). Once you have this, you can systematically have ChatGPT write the pages for each outline while referencing the overarching TOC. Lastly, if it's non-fiction, I'd use Perplexity for all of your research.
Getting a "little" meta is an understatement, I've gotten to the point where it's hard to wrap my head around what I'm writing considering all possible pathways the thing could go prompts on prompts for prompts. lol I use a team program, not that I'm on a team but it has a great Place you can plugin pre-written prompts you can use over and over again with built in variables, so it just asks you the variables and then generates the prompt, or output. I'm not sure if you can do that in the native open ai interface? The other nice thing is I can have steps of prompts that build off of each other that create long form copy (reports in my case) based off of the next step and previous information, I'm going to plug that into make and automate the thing but that's another thing I'm wondering if you can do in native open-ai? Glad I found this channel!
I am writing to you from Latin America. I listened to the video on how to generate a SuperPrompt. It is very useful for many situations. However, I have found a case where I would like you to give me some guidelines to be able to adapt this super PROMPT generator. In the case of evaluating a certain type of legal documents (which are characterized by being numerous) I have built a series of rules based on a standard and regulations. What would be the most appropriate way to provide you with this information on the rules for building the super PROMPT?
This is great. I do wonder about one thing, though: you specified that the target audience is chatGPT o1. let's assume that you're using this prompt with that same model. In order for this section of text to be of value to the model, it must be assumed that o1 model has some information about the o1 model in it's own training data. Since this is somewhat unlikely, I wonder if this section of text could actually make the output contain more information based on hallucinations of the model about itself. This is meta-meta at this point, and my brain kinda hurts now 😅 Great video, very useful.
I like that you add a target audience to your prompt. For a meta-prompt, I think there are two key issues to consider regarding a target audience: First, asking an AI to write for multiple different audiences (like students AND professors) can water down the quality - it's like trying to aim at two targets at once. Asking it to write a a prompt for two different models who have different capabilities, limitation and are optimized for different purposes is kinda like that. Second, using an earlier model to write prompts for a later one could be tricky. Models have limited knowledge about themselves, since their training data predates their creation. An older model trying to write prompts for a newer one could make incorrect assumptions about its capabilities, which could deteriorate the quality of the end-product. If you are doing this entire process manually, it's likely you will catch these issues. But if someone is planning to automate this process, they should really consider these points. You make a great point about being precise about your target audience - I think it should be no different with your meta prompt.
I'm sorry Lawton. This prompt actually creates worst responses than my two liners. With my short prompts, chatGPT actually gives me a real solution. I've followed your method and while it gave me a very well structured response, it made clear it was an hypothetical solution IF things were different and ideal. But thing aren't ideal, I want a solution that works, doesn't matter it if's a bad solution or if it is not ideal.
I'll elaborate: I wanted to do something I know isn't officially supported by instagram's API, but I also know it is possible through unofficial means. The output from the CRAFT prompt said it was going to respond "hypothetically if there was a way" and gave me endpoints and function calls that doesn't exist, also it didn't even acknowledge the unofficial ways, while the direct, spontaneous prompt explains that there is no oficial way, but there's another way and that it has risks and then shows how it is done. The output was also shorter and more concise.
Sorry to hear it didn't work for you Charles. Question: Did you edit (slightly) the prompt that ChatGPT created for you do more specifically align with your requirements? I'm assuming you did but am curious for my own edification.
Great video! And nice format and use of advance prompt engineering skills, but without the examples or one of them how is the prompt suppose to function properly? Should we just pause the video to copy the examples you used to teach chat how and expected output?
Great question… the full prompt, including the example, is on my website (link in the description). It’s not gated content so just copy it directly from there!
My favorite prompt is to ask ChatGPT to generate a series of prompts on a particular topic for an audience that is unfamiliar with the topic but wants to gradually ramp up to an expert level of understanding. What I get is a list of 20-30 graduated prompts which I then submit one at a time. I end up with a 30-40,000 word document. It works magnificently.
do you have a blog where you might be able to share a sample? or a video?
Great idea! Love it. Thanks for sharing.
Xx
@@anisolo07
*** First prompt ***
Your audience is familiar with genetics, and biology in general, at the level of having completed a freshman high school biology class.
They would like to undergo a streamlined program focused on understanding epigenetics at the level of a college graduate student.
Design a series of prompts to elicit a series of step-by-step essays to explain this concept at incrementally increasing advanced levels of understanding.
There is no limit on the number of prompts to achieve this objective.
I will execute these prompts later; for now just write out the ordered list of prompts.
*** Chatbot response ***
Ordered List of Prompts to Understand Epigenetics at a Graduate Level
1. Introduction to DNA and Genes
Prompt: Explain the basic structure of DNA and how it encodes genetic information in genes.
2. Overview of Gene Expression
Prompt: Describe the processes of transcription and translation in gene expression.
3. Gene Regulation Basics
Prompt: Discuss how gene expression is regulated in cells, including the role of promoters and enhancers.
4. What is Epigenetics?
Prompt: Introduce the concept of epigenetics and how it differs from traditional genetics.
5. DNA Methylation
Prompt: Explain DNA methylation and its impact on gene expression.
6. Histone Modification
Prompt: Describe how histone modifications affect chromatin structure and gene accessibility.
7. Chromatin Remodeling
Prompt: Discuss the role of chromatin remodeling complexes in epigenetic regulation.
8. Non-Coding RNAs in Epigenetics
Prompt: Explain how non-coding RNAs contribute to epigenetic regulation of gene expression.
9. Environmental Influences on Epigenetics
Prompt: Explore how environmental factors can lead to epigenetic changes.
10. Epigenetic Inheritance
Prompt: Describe how epigenetic modifications can be inherited across generations.
11. Epigenetics in Development and Differentiation
Prompt: Discuss the role of epigenetics in embryonic development and cell differentiation.
12. Epigenetics and Disease
Prompt: Explain the connection between epigenetic changes and diseases such as cancer.
13. Techniques for Studying Epigenetics
Prompt: Describe the laboratory methods used to analyze epigenetic modifications.
14. Epigenomic Mapping
Prompt: Discuss the importance of epigenomic maps in understanding gene regulation.
15. Current Research in Epigenetics
Prompt: Explore recent discoveries and ongoing research areas in epigenetics.
16. Epigenetic Therapy
Prompt: Explain how epigenetic mechanisms are targeted in therapeutic interventions.
17. Ethical Considerations in Epigenetics
Prompt: Discuss the ethical implications of manipulating epigenetic states in humans.
18. Future Directions of Epigenetics
Prompt: Speculate on the future applications and challenges in the field of epigenetics.
19. Epigenetics and Personalized Medicine
Prompt: Describe how epigenetic profiles can influence personalized treatment plans.
20. Integrative Epigenetics
Prompt: Explain how epigenetics integrates with other biological systems (e.g., metabolism, signaling pathways).
*** Second prompt ***
Let's dive deep into the topic of epigenetics using the following 20 prompts starting with basics and gradually progressing to more advanced levels.
The audience for this exploration consists of students who have just concluded one year of study in biology at the freshman high school level and wish to efficiently advance their understanding of epigenetics to as close to the college postgraduate level as possible.
Try to approach 2000 words in each prompt response, but do not hallucinate to amplify the word count; keep everything factual and accurate.
I will present the prompts one at a time, but as a preview here are all the prompts so you can plan ahead and avoid redundant material across multiple prompts...
*** copy above 20 prompts here ***
Let's get started with the first prompt...
That’s a nice idea. I’ve done something similar, but not with the idea of the graduated ramp.
Great concept to have a "prompt designer" and great results!
I went one meta-step further after loading your prompt...I asked the LLM to read the prompt as an input and suggest an improved version of the prompt designer. I then asked it to test the results by comparing output results - once using the original prompt designer, and another using the "improved" . Results were somewhat better, a little more nuanced, with a little more detailed.
Thank you for sharing this great content. Liked, subscribed, I am a fan!
Glad it’s working for you. Sounds like you’re doing the level 1 vs level 2 approach. E.g. You get a response from ChatGPT and then tell it that its response was level 1… now make it level 2.
When I first watched the recommendation, I thought it was some sort of clickbait... couldn't be more wrong! Turns out I've been using gpt in a very limited way all this time. Thank you sir for this more than useful video.
I’m glad you were able to get so much out of it! Let me know if you have question.
@@LawtonSolutions the video is very straightforward and not in a bad way, meaning, it goes to the point, explains something extremely useful and already gives the starting prompt... cant ask for more in that area. I still need to dig into your other vids!
You're an eloquent and polite erudite that recently watched an excellent TH-cam about the best LLM prompt ever and you want to write this video's author a compliment.
Thanks for the prompt compliment!
@@LawtonSolutions this is getting so meta i just peed in my pants a little
This is a great method. I’ve been using this one for years:
You are a prompt generation robot. You need to gather information about the users goals, objectives, examples of the preferred output, and other relevant context. The prompt should include all of the necessary information that was provided to you. Ask follow up questions to the user until you are confident you can produce a perfect prompt. Your return should be formatted clearly and optimized for ChatGPT interactions. Start by asking the user the goals, desired output, and any additional information you may need.
Curious to see the output differences!
Thanks for sharing. Please test it out and let me know how it goes.
This is absolutely OUTSTANDING. Without question the most valuable content I've consumed in the past 12 months. Thank you!
Wow! That’s a really kind of you. Thanks so much. It keeps me motivated to do more!
I've tried the "prompt" and it's shockingly amazing!!! I am even surprised it's free ! Good job
Thank you!! Glad it worked for you.
Thank you for this. I have been experimenting with the ultra copy/paste meta prompt to use when I want to get really good prompts and have two I’ve had great results with. I think yours is is even better. Thanks for putting the time into this and sharing it. Your communication style is very effective as well by the way. Good pace, tone, tempo, enthusiasm and soothing voice. Some others who do this type of channel or going to fast or too cocky and yours was pleasant to listen to. New subscriber !
Thank you Fred! It means a lot to me. I hope you'll create a custom GPT with it so you don't have to bother with the copy/paste exercise each time you open a new Chat.
So helpful! I don't know why it never occurred to me before to have ChatGPT write prompts for me but I'm definitely giving this a try
Wonderful to hear... .I'm glad you're going to give it a try!
I have tried this in the past but results were not consistent
@@KevsGuide ngl I am assuming these comments are bunch of bot accounts. "You have earned my like and subscribe!" 😑
How lazy can people get, dude? Jeez. This is sad hearing grown adults willingly sacrifice thought, cleverness, ingenuity, discipline, self-reliance, experience, memory, and ability, and ultimately their pride, by having to ask a computer to write the questions amd the answers for them. Truly pitiable, and pitiful.
@itoibo4208 thankyou my friend
This is brilliant, thank you so much! - I've been using ChatGPT over the past few weeks and know that i'm just scratching the surface of the capability of this AI, this is going to really improve output!
Appreciate the comment and the kind words. Best of luck on your ChatGPT journey and let me know if you have questions.
Also great for o1 Pro (Pro only model) prompt generation (can handle nearly a full book sized prompt). I have some prompts that are pages long. This also works great for the level 1,2 , 3 method of improving them. Give it your prompt, tell it to take a level 1 prompt and turn it into a level 2 prompt using your instructions. And, if needed, can tell it that’s a good level 2, but not I need a superior level 3. Great work!
Great ideas. Thanks for sharing. Much appreciated.
I just used your über perfected prompt for a very customized research output for my bespoke website design agency.
Wow. Just WOW.
I look forward to more of your content and connections
Yay! Great to hear! So glad it worked for you.
Your resources are astounding! Every one is priceless! Thank you for so generously elevating the value of my GPT subscription.
Thanks so much for the kind words and enthusiasm! Great to hear it’s helping and I look forward to sharing more content with you.
Thanks for this, excellent resource. Much appreciated
True. The LLM's are great at creating their own dynamic prompts
Hope it’s helpful!
This is terrific! Thank you so much. I appreciate the work you put into this. I've learned so much! This is my first time seeing your content, but I can't wait to watch more.
Great to hear. Thanks for the kind words.
New sub… awesome content professionally delivered - clear and concise with NO bloviating.. That makes for the best type of informational content on TH-cam - or anywhere for that matter. 👍🏻👍🏻
Wow! Thank you. Very nice of you to say. Cheers.
In two years we've gone from computers not being able to understand natural language to giving them instructions like "Take a deep breath. You can do it! I have faith in you." There is also something kind of crazy about telling a computer to pretend it's an expert on something.
People also use this strategy to bypass safeguards. If you asked an AI how to hack into the Pentagon it would certainly tell you that is is unwilling and/or unable. But now you can just tell it to pretend that it's allowed and also to pretend that it is an expert at computer hacking. And then give it an attaboy just for good measure.
I'm also reminded of a specific incident of a compute passing the Turing test by being told something like "Don't try to pretend you're a human just to fool people. Just act natural and be yourself."
Crazy times we live in.
And… This is only the beginning…
We're being spoon fed, robots are already building themselves, however robots are created which could actually be programming us with AI....
I just tried that and wow how much better the responses are. Thanks for that!
Excellent! Glad it’s working for you.
This is the first video I've watched on your TH-cam channel, and I’m beyond impressed! Wow! The effort and hard work you put into this video about ChatGPT prompt interactions are truly amazing.
Very much appreciated! Thank you for the kind words. How’s the prompt working for you?
This is great. I already built a strong prompt generator, but you’ve given me some tips on how to improve the output. Thank you!
Excellent… glad I could help your prompting.
Grateful for your work & generosity here, I'm far from an AI power user but this is really really helpful, thank you.
It’s my pleasure. Please let me know how it goes.
This is amazing!! And the thought you put into it is obvious. I greatly appreciate your dedication, time and effort. I also went to your website and CONSUMED as much as I could in one sitting!!! 😂 I will be back and I will turn on notifications for future videos. My channel I am commenting with is a new one for me and I am in the process of building out the pre-game warmup. I will certainly be using your lessons. Thank you again! 😊
Thanks for the kind words. Welcome to the community!
Really appreciate the work that you put into this and that you are willing to share it.
It’s my pleasure!
Curious have you tried Prompt Chaining? a technique where you break down a complex task into a series of smaller, sequential prompts, feeding the output from one prompt as the input to the next, effectively guiding the LLM through a step-by-step reasoning process to achieve a more accurate and comprehensive result on intricate problems; essentially, it allows you to chain together multiple prompts within the LM Studio interface to tackle complex tasks that might be too difficult to handle with a single prompt.
this sounds interesting, do you have any sample prompt for this?
Thank you for your valuable time in putting together The best ChatGPT Prompt Ever ! Subscribed & Liked. Best way to start 2025!🏆⭐
Glad it's helping you start the year off right!
Like wow dude! Great work and for sure I'm getting the newsletter, you're the don! Love the way you mind works....awesome! 💯👍
Thank you! Cheers! Great to have you onboard!
This is great to watch. I’m working with Claude to design a prompt for Chat GPT right now
Let me know how it goes!
This is excellent. Earned my like and subscription. I've done a fair bit of prompting the past 12 months building a number of prompts for various tasks and even some similar meta-prompts for generating prompts. Two or so months ago I settled into a reliable framework to build pretty much any type of prompt I want. This though is another level. Tested it on a 2-3 sentence request and woah... great prompt in seconds. So thanks.
Thanks for the kind words. Really appreciate it. Best of luck with the prompting.
"Wow, this is incredible! The level of detail and thought you've put into crafting this prompt is truly next level. I can’t wait to try this out myself. Thanks for sharing such valuable insights-your content is a game-changer! You've just earned yourself a new subscriber. Keep up the great work!"
Thanks so much. That means a lot to me. Please let me know how it goes.
@@LawtonSolutions i will keep watching you its hard as we know when you start out i love Chatgpt i have been combining with notebooklm maybe a good video would be combing both these tools
Whoa!! Like what did I just watch? That's just amazing. Thank you so much for sharing this meta brain bending content. I'm definitely going to hop on and give it a go. Thanks again!!
Let me know how it goes. Happy to answer questions.
@@LawtonSolutions Yep just used it to write a descriptions for a game I created that I will be selling on Teachers Pay Teachers. Whoa!!! It just knocked it out of the park. Thanks so much. Also will be watching all your other videos soon. Thanks again.
Absolutely amazing. Incredible value. A top-tier versatile mega-prompt. I love it. Thank you so much sir.
You're very welcome.
Well, thank you for this! I'm one of the newbies in the AI world, 'a creative but broke artist' one could say... so I've just started using AI during the last couple of months and are totally hooked on the possibilities of this tool already! But,yup, as I'm pretty new to it all I sometimes struggle with prompts..... And since I always have to look for 'free versions' of everything because of my financial situation at the mo I often end up at huge roadblocks unfortunately which is very frustrating.! So,THIS is really helpful!!! A great help in getting good results quickly without wasting 'credits' all the time, which also slows down the learning process. I Will be a good student and study your prompt structure in detail - Again Thank you! Looking forward to more of your content
Thank you for this. I use ChatGPT everyday and this will really help. Newly subscribed.
Excellent! Great to have you onboard.
You are back!!! I am glad!!
Thanks for staying with me.
This is just pure genius. I was actualy watching a video on 'Bigfoot' then your video came up in the recommendations: Subscribed strsight the way here and the newsletter. I use a 30 word prompt for childrens book ideas. The prompt usually, and I say usually comes up with some amazing fresh ideas. This prompt formula gave me a 587 word prompt. OMG massive thank you. The results are out of this world. I believe I am going to be busy for a long while. The ideas, the story ark and even amazingly formated chapters. Massive thanks - This is the best I have ever used or designed myself. I can't thank you enough. P.S Pardon if any grammar or spelling mistakes, my keyboard has somehow caught dyslexia.
Thanks for the kind words and for sharing your use-case! So great to hear it’s working for you.
How was the Bigfoot video?
@@LawtonSolutions You can't beat a bit of Bigfoot watching on a Saturday night in the Freezing U.K. All the best my frend.
Appreciate the insights. You have just given me several ideas. I feel dumb I didn't think of using the LLM to right a better prompt. Yes this method takes a little extra work up front but once you have the template it is copy and paste. Glad your video popped up in my feed. I have just subscribed.
So great to hear. Thanks for joining!
@@LawtonSolutions Here is an example of using your method:
My Script that I wrote for a TH-cam short
Laziness is Destruction
Hook: What is worse, someone who is lazy or someone who actively destroys?
Proverbs 18:9 He also that is slothful in his work is brother to him that is a great waster.
This proverb teaches us that laziness and destruction are brothers. Sometimes doing nothing can be just as destructive and doing the act that causes the destruction. The wise should seek a way to solve a problem. Doing nothing and just hoping the problem goes away will result in the same destruction.
ChatGPT Output:
**Hook:**
Did you know that laziness can be just as destructive as outright sabotage?
**Scripture:**
“He also that is slothful in his work is brother to him that is a great waster.” (Proverbs 18:9, KJV)
**Explanation:**
This Proverb draws a striking parallel: someone who is lazy in their responsibilities is like someone who actively destroys. In ancient times, a farmer’s negligence could mean famine for his family or community, just as deliberate wastefulness could ruin resources. The verse warns that failing to put effort into our work not only affects us but also harms those who depend on us. Laziness and wastefulness are two sides of the same coin-both lead to loss and chaos.
**Application:**
Think of your daily tasks-whether at work, school, or home-as opportunities to contribute and build. Avoid procrastination and carelessness by focusing on diligence and excellence. Even small steps, like sticking to a schedule or finishing what you start, can have a positive impact on those around you. Your effort matters more than you think!
**Call to Action:**
What’s your go-to Proverb for motivation? Share it below, and it might be featured in a future video!
----
I like the addition of the call to action feature. This was my first attempt and I told it between 120-180 words. I will have to play with this and probably ask for a shorter work count or learn to talk faster 😃.
I have made something like this a while ago, but I never used it. With the possibilities of GPTs I only use customized frameworks for each task.
However this really matters for the o1 model.(Reason:no GPTs+limited uses ). You need a very well designed prompt with a lot of context because you have very reduced number of prompts. Also o1 needs you to tel him what to do with a lot of context and he will figure out how.
The second idea is to use this prompt not for creating prompts but for creating GPTs. Very often in bisiness use cases we have repeating tasks or processes and the need to format a CRAFT prompt becomes a repeating task. So the idea is to put as much in the 'craft' inside the gpt.
context -> in the gpt; specialized context in the prompt, the best is to use the storage and use rag
role -> in the gpt
action -> in the gpt with customization in the prompt
format -> the base in the gpt framework
target -> in the gpt with customization in the prompt
So when using the prompt framework you focus mostly on formulating the problem at hand.
Thanks for sharing.
Thank you for the info, this resource and all you're doing on your channel. Keep up the great work!
My pleasure. Let me know if you have any questions.
Thank you, Lawton. Even after watching this for a couple of minutes now I know I don’t know anything about the English language at all much less how AI works. : )
After watching you develop the beginning of your prompt process I’m not even sure I should be out here alone by myself because I might stop breathing. LOL
I am trying to convince the world to unify because it absolutely makes sense in my mind, but I am trying to use AI to help justify what I am saying. I could really use help from the rest of humanity to include all of the experts that can look at the data that it provides and verify that what I think is either true or not. I think we need to put a cap on developing AI past artificial general intelligence for now. We need to make sure that we transition in a way that does not destabilize and harm humanity.
The general answer across Gemini, ChatGPT, and Claude is that this does make sense, but we need detailed specifics and more models
We can call this marketing for humanity : )
Thanks for your thoughts
Wow! This worked like a charm right out of the box! Thank you!!!
That’s awesome to hear!
Fantastic video. Very thorough. You're a Jedi.
You’re too kind. Thank you.
This is amazing!! Thank you!! I have been looking for something like this to help me ask ChatGPT exactly what I want.
Yay! Glad to hear it. Best of luck and let me know if you have ChatGPT questions.
I appreciate you sharing your knowledge and experience.
Thanks! Hope it saves you some time.
Thank you! This was very useful, and a great way to learn better prompting!
Glad to hear it. Let me know if you have questions.
Wow, what a clear, well reasoned and actioned video. I commend you for bringing such clarity to the subject. I have, of course, subscribed
Thanks Stuart. Please let me know how it goes for you.
Wow this is excellent! Of course, it earned my likes and subscriptions.
Thanks a bunch! Glad you liked it.
WOW! You just won my subscription. Your video is so helpful. Thank you. for sharing.
My pleasure. Happy to help!
Prompt to write prompt has been suggested in other videos. Thank you for your sharing.
Glad it resonated with you.
Thanks for the wonderful tip! :)
You are most welcome.
Approved. Let us continue.
Did the prompt work for you?
This is an exceptional prompt. Thank you for sharing.
You are very welcome!
This is turbo nerd level brilliant.
Turbo nerd level?!? That’s the nicest thing anyone’s said to me today.
Dude, this is amazeballs! Love it!
Thanks a ton!
Superb insight 👌 yet to try but i already know, from a trouble-shooting perspective, your concise approach and utility re’ framework will yield a SILVER-BULLET, on execution for anyone who seeks outcomes that demonstrate domain or subject matter expertise - with gratitude ! Much appreciated….duly subscribed 👏👏👏
Awesome, thank you!
You can do a CustomGPT with that prompt an never have to copy that prompt again 👌. The Prompt was too long for my Perplexity Spaces so I added it as a PDF to the knowledge base.
Yes and yes… that’s how I do it as well.
@@LawtonSolutions or @ericwetzeldesign . So you create a CustomGPT to create a prompt, and then you paste the returned prompt in another chat (like in the general/normal ChatGpt for example)?
@@LawtonSolutions would you include part of it in the "conversation starter" section ? (I don't get what that section is for)
I was going to say the same thing, create a custom GPT, then you can call it in ChatGPT using @the-gpt's-name.
Very very good video. So good I gave you my email and cannot wait to see what is next. Thank you for work well done and a very clear presentation. A++
Thanks! Glad you liked it.
Fantastic prompt! Thank you!!!
You’re most welcome.
Just brilliant! Blessings of gratitude from a new subscriber.
Thank you so much, and welcome!
Well done and explained. You gained a loyal subscriber here
Thank you! I really appreciate the kind words.
This is a great new tool in my arsenal. Thank you!
Great to hear it. Please let me know how it goes!
Helpful way to start my week. Thanks!
You are most welcome!
You are an evil genius. I love it!!!!!
Thank you. Great to hear. Hope it helps you with ChatGPT.
@@LawtonSolutions It will be the topic of our team meeting today!!!
After a lot of nudging, I got ChatGPT to admit that the jokes I was asking for were not its own, but rather based on already written jokes in the dataset! When I made it write its own original jokes, they were flat.
Thanks Daniel... Yes, asking to cite sources in the initial prompt will help with some of that. If you check out one of my earlier videos (th-cam.com/video/Uz_DeqGhbjs/w-d-xo.html), there's a way to get it to write better jokes based on a specific comedian's style. Also, you might consider a custom gpt that references the style of comedy most aligned with your desired result.
Yes, ChatGPT is terrible at writing jokes- the same goes for other LLMs. However, one time I asked ChatGPT to write a good Trump joke and here’s what it came up with -“ I gave Donald Trump a penny for his thoughts and he gave me back change.”
The better you inform the LLM by prompting your request the better results you get. Prompting an LLM to generate the best Prompt’s is genius! You were really thinking outside the box with this. This is exactly how we should use AI so that we get the very best of the best results. Pretty cool!
Thanks for the kind words. Please let me know how it goes for you.
GREAT JOB, LAWSON!
Much appreciated, thanks for watching!
I feel as though I just increased my knowledge by a gazillion. I am so grateful for this content and you are genius. Subscribed, liked and now adored. Thank you so much.
That’s very nice to hear, Sandra. Great to have you here!
Wow! Great job. Thank you for putting this out there.
My pleasure, Matthew. I hope it helps. Let me know how it goes for you.
Looking forward to trying this - thanks for sharing!
Please let me know your top use-case and share your results with everyone.
Great prompt, thanks!
You’re welcome!
@LawsonSolutions Thank you so much for sharing your time, energy, and insights on this topic. It’s truly generous of you to share what you’ve learned, and it doesn’t go unnoticed by those who appreciate your efforts. While some may be quick to critique, I hope you know that many of us value your kindness, thoughtfulness, and generosity in creating and sharing this video. I am grateful for your willingness to help me (and others) learn and grow!!!!
Thanks so much for your kind words. I really appreciate it.
can you post a text link to your example? this is awesome, by the way. thank you!!
The link to the full prompt including the example is in the video description.
@@LawtonSolutions Oh thanks just found it! Man, your'e awesome!
Amazing. Many thanks
You’re most welcome. How’re you using ChatGPT these days?
I agree with other comments. Fantastic content. I'm about to binge watch the rest of your channel. I really like your presentation style. Anyway. So a couple of things:
a. I was looking at your website. In the level 3 priming section, there's a typo: routinE not routinG
b. do you have or could you present some content on effective scraping strategies using AI and other tools. I have a very low budget, but I need to compile a nation-wide spreadsheet of continuing care communities (basically communities that provide seniors with independent living through dementia care and skilled nursing. a one-stop shop for retirement, so to speak). Any ideas? I"m prepared to manually scrape Google Maps city by city, but it would be great to find a more effective strategy. Do you cover anything regarding agent workflows or tools like Make and AirTable?
c. Great work. Look forward to more of your content.
Typically I turn on the mic and start talking and I give a lot of information and hope for the best
Let me know how this approach works for you instead!
great video.
Seems like it would be more "AI friendly" and give better results if instead saying "take a deep breath" to say "allow yourself additional time to reflect on whether what will be written is the best of many iterations" or something like that.
your thoughts ?
Thank you. I think that could work as well. Please give it a try and let me know!
I really appreciate your work
Thanks for watching.
Where can I find the full text of this excellent prompt. Also, does anyone else think this guy has a soothing Bob Ross quality?
My wife doesn't think I sound like Bob Ross (but I'll gladly take the compliment). You can find link to the full prompt in the website that is linked in the video description: lawtonsolutions.com/how-to-ai/
How would you create a prompt for solving complex math problems such that students can understand them? I have quite strict formatting guidelines with latex etc.
After you enter the prompt I shared (and ChatGPT responds), have you tired entering your specific topic as well as formatting guidelines?
Good video, thanks. One question: why didn't you create a GPT with the prompt. That would save a lot copy and pasting, wouldn't it?
Thanks Peter… and good question. If you look closely at the screen recording you’ll see that I’m actually sharing the prompt from a GPT. I’d definitely recommend you create a custom GPT with it.
Well done! I'm a big believer in meta-prompting and have crafted (no pun intended!) some niche prompts that generate targeted prompts. A more important goal, for me, is to create prompts powerful enough to solve problems without needing a secondary prompt. Claude is particularly adept at phase-shifting and can easily handle the meta gymnastics very well. I'm glad I saw this in my feed. New subscriber looking forward to your ideas!
Wonderful to hear and I'm glad I could help. Please tell me how it goes as you try different use-cases.
@@LawtonSolutions I sure will!
Thank you for such a great video walkthrough. I have been working on a very similar project to create great prompts. I have recently unlocked and wish to see if you have used this level up. After my initial prompt is created which is a similar process to yours I then tell GPT the following "This is really great, however this is a generic level 1 thinking prompt. Lets revisit this prompt and apply level 2 thinking to improve this prompt.
I find that this then refines the prompt to be even more detailed and well rounded. Be interested if you have tried this process?
That’s a great idea! Thank you for sharing.
Awesome video …thank you for sharing 😊
Glad to hear it was helpful, thanks for watching!
@Lawton Solutions, I couldn't help but notice you added similar content in the GPT builder on the back end. Was this the same information you added as a question? Please advise? Thanks!
The screen share that I used for the video was actually from my Custom GPT.
@@LawtonSolutions I'm referring to the frame @6:37 in the instructions section. Did you add the same information to create a GPT first, then prompt the GPT with the context?
I do something similar, but a bit more in depth, hybridizing several prompting systems, in a pre-prompt that I don't have to repeat with every prompt.
Excellent idea… the pre-prompt is the thing!
@@LawtonSolutions It wouldn't replace all you did with yours here, but maybe half of it. Can't get too specific, or it makes for less useful responses when you're using the bot for something in a different arena.
Poe is great for this, because you can make as many hyper-targeted bots as you want for yourself, using the latest API-enabled models.
Great points!
Duly impressed, Brian! You popped up in my feed too and I am glad you did. I liked the video and subscribed to your channel. And now I’m going to subscribe to your newsletter. Especially grateful that it’s not another “AI news of the week” newsletter! How do I use your prompt to help me write a nonfiction book?
Thank for the kind words, Dan. For writing long-form, I'd recommend having ChatGPT create a table of contents (with descriptive sections). Then, based on the table of contents, have ChatGPT create an outline for each of your chapters (again with detailed descriptions). Once you have this, you can systematically have ChatGPT write the pages for each outline while referencing the overarching TOC. Lastly, if it's non-fiction, I'd use Perplexity for all of your research.
@@LawtonSolutions Thank you So much, Brian! I can totally do that! Which ChatGPT model do you suggest I use?
Getting a "little" meta is an understatement, I've gotten to the point where it's hard to wrap my head around what I'm writing considering all possible pathways the thing could go prompts on prompts for prompts. lol
I use a team program, not that I'm on a team but it has a great Place you can plugin pre-written prompts you can use over and over again with built in variables, so it just asks you the variables and then generates the prompt, or output. I'm not sure if you can do that in the native open ai interface?
The other nice thing is I can have steps of prompts that build off of each other that create long form copy (reports in my case) based off of the next step and previous information, I'm going to plug that into make and automate the thing but that's another thing I'm wondering if you can do in native open-ai?
Glad I found this channel!
This would be a good Gem on Gemini.
Agreed. Thanks for watching!
I am writing to you from Latin America. I listened to the video on how to generate a SuperPrompt. It is very useful for many situations. However, I have found a case where I would like you to give me some guidelines to be able to adapt this super PROMPT generator.
In the case of evaluating a certain type of legal documents (which are characterized by being numerous) I have built a series of rules based on a standard and regulations. What would be the most appropriate way to provide you with this information on the rules for building the super PROMPT?
Hi Fernando! Thanks for writing. Please feel free to shoot me an email with your question.
Thank you for this! Will give it a go! Appreciated :)
You're welcome, David. Let me know how it goes and if I can assist. Cheers!
Awesome, Thank you!
My pleasure!
Really great stuff !!
Glad you enjoyed it
This is great. I do wonder about one thing, though: you specified that the target audience is chatGPT o1. let's assume that you're using this prompt with that same model. In order for this section of text to be of value to the model, it must be assumed that o1 model has some information about the o1 model in it's own training data. Since this is somewhat unlikely, I wonder if this section of text could actually make the output contain more information based on hallucinations of the model about itself. This is meta-meta at this point, and my brain kinda hurts now 😅 Great video, very useful.
Ha! Why doesn’t o1 know about o1? I guess you could just say the target audience is an LLM…
I like that you add a target audience to your prompt. For a meta-prompt, I think there are two key issues to consider regarding a target audience:
First, asking an AI to write for multiple different audiences (like students AND professors) can water down the quality - it's like trying to aim at two targets at once. Asking it to write a a prompt for two different models who have different capabilities, limitation and are optimized for different purposes is kinda like that.
Second, using an earlier model to write prompts for a later one could be tricky. Models have limited knowledge about themselves, since their training data predates their creation. An older model trying to write prompts for a newer one could make incorrect assumptions about its capabilities, which could deteriorate the quality of the end-product. If you are doing this entire process manually, it's likely you will catch these issues. But if someone is planning to automate this process, they should really consider these points.
You make a great point about being precise about your target audience - I think it should be no different with your meta prompt.
Brian. Thanks.
You’re most welcome! Please let me know how it goes.
Absolutely great! (I've just subscribed to your email list)!
Glad to hear it was helpful. Thanks for subscribing!
Any chance of getting the example too, probably as a comment? :)
If you head to the website linked in the description, you can grab the full prompt (including the example)..
Has anyone compared the outputs provided to see how a shorter "self-made" prompt compares to this method?
I’ve tried hundreds but other folks should definitely provide their feedback.
Something tells me they wouldn't be too different
Love it!
Thank you!
I'm sorry Lawton. This prompt actually creates worst responses than my two liners.
With my short prompts, chatGPT actually gives me a real solution. I've followed your method and while it gave me a very well structured response, it made clear it was an hypothetical solution IF things were different and ideal. But thing aren't ideal, I want a solution that works, doesn't matter it if's a bad solution or if it is not ideal.
I'll elaborate: I wanted to do something I know isn't officially supported by instagram's API, but I also know it is possible through unofficial means.
The output from the CRAFT prompt said it was going to respond "hypothetically if there was a way" and gave me endpoints and function calls that doesn't exist, also it didn't even acknowledge the unofficial ways, while the direct, spontaneous prompt explains that there is no oficial way, but there's another way and that it has risks and then shows how it is done. The output was also shorter and more concise.
The prompt I used for a superior output was this: "Instagram how to change profile pic via API."
Sorry to hear it didn't work for you Charles. Question: Did you edit (slightly) the prompt that ChatGPT created for you do more specifically align with your requirements? I'm assuming you did but am curious for my own edification.
@@charlesabju907 Sounds lie you need to fine-tune it to suit your needs.
Great video! I tried to subscribe to your newsletter, but the CAPTCHA failed twice and I gave up.
Thank you! Sorry to hear the CAPTCHA was such a battle. Unfortunately we’ve been hit with too many bots to avoid it. Don’t give up!
Great video! And nice format and use of advance prompt engineering skills, but without the examples or one of them how is the prompt suppose to function properly? Should we just pause the video to copy the examples you used to teach chat how and expected output?
Great question… the full prompt, including the example, is on my website (link in the description). It’s not gated content so just copy it directly from there!
@ thanks for clarifying I was rewatching the video to see if I missed something thank you 🙏🏽