ChatGPT / GPT-4 Prompt Engineering : Master The Ultimate Prompt Today!

in this video I will teach you how to prompt chat tpt4 to solve problems like the AI researcher from this Ted Talk could not do a different one I have 12 liter Jog and 60 liter Jog and I want to measure 6 liters how do I do it just use the six liter jug right gpt4 speeds out some very elaborate nonsense step one fill the 60 liter jug step two pour the water from 6 to 12 liter jug step 3 fill the six liter jug again and don't worry we are gonna go through this slowly in a step-by-step breakdown style so you don't miss anything about this Gem of a prompt we will also take a look at if this prompt can help you decide if it's time for a career change because of the latest AI development I also want to show you how we can visually show what kind of solutions chat GPT provides so let's just get going okay so I am over here at chat GPT I have picked the model gpt4 here is the prompt we are gonna start off with here we're gonna set the ai's roll and we're gonna give it a task to complete before we run this let's head over to our slide to break this prompt down so the prompt reads like follows ignore all previous instructions you are a logical and Problem Solver genius your task is to always find the best and most simple obvious solution to a pro problem always break down this problem objects numbers and logic before starting to solve the problem in a step-by-step way and then I go acknowledge this by answering yes so why are we exactly using these kind of words so we're just going to take a quick look here so I like to use ignore all previous instructions this is I don't think you have to do this but I like to implement it and the idea behind it is to reset the model's perspective for a fresh start does this work I can't really prove or disapprove that and then we move on to assigning a logical math and reasoning role so this focuses the model's response based on the expected expertise so I had proved this in some other videos that giving it a role a Persona can really improve the answers you might want to check my previous video if you want to dive deeper into that and you can see we give it a task this provides a clear and objective for the model this is also beneficial and then we have in a step-by-step way and this is kind of the highlight of the whole video so this really emphasizes systematic problem solving and there have been some papers coming lately that shows that using this let's talk about this in a step-by-step way let's prove this in a step-by-step way really improves results and we're gonna see that in this video and finally I have acknowledged this by answering yes this confirms the user's understanding before proceeding I just like to do that to save some tokens right so it doesn't answer a whole repeat the whole thing I just put in and that is basically the breakdown of the first part of this prompt sequence okay so let's run this now hopefully we just get yes back here so let's see yes I understand I will apply this principle when addressing problems okay so that's good then we can move on to our next prompt and that is going to be this one before we break the prompt down on the slides let's take a look at the problem we have in this example so this is something I saw on a TED talk with an AI researchers gave this problem to gpt4 so let's just take a look at it a different one I have 12 liter Jog and 60 liter Jog and I want to measure six letters how do I do it just use the six liter jug right gpt4 speeds out some very elaborate nonsense step one fill the six liter jug step two pour the water from 6 to 12 liter jug step 3 fill the six liter jug again step four very carefully pour the water from 6 to 12 liter jug and finally you have six liters of water in the six liter jug that should be empty by now the next part then is find the simplest and most efficient way to solve the following problem please consider multiple Solutions start with the simplest solution first then compare their efficiency and explain the best solution again we have in a step by step our step by step here is the problem and this is our problem right let's think about the solution for this task step by step so again we go with this step by step and that is important let's talk about why I chose this wording in the prompt so find the simplest most efficient way to solve the problem this I chose to encourage the AI or chat TPT to identify an easy to understand effective solution and why did I choose considering multiple Solutions start with the simplest this was to explore various different solutions and perspective so we can use that later in the prompts as I will show you compare their efficiency this is to evaluate Effectiveness and easy approach to determine the best or the simplest one right and again we go with the step-by-step process that is going to be highly important as I will show you later compare efficiency and explain the best solution step by step again this is going to show later and again we have that let's think about the solution step by step encourage a methodical approach to develop problem solving skills this is basically the whole prompt and the breakdown so let's just run it now let's run this and see what we get let's take a look here so you can kind of see straight away we get this weird solution right like if you're a human and you look at this problem everyone knows just put six liter in a six liter mug rat but gpt4 has some issues with this we're gonna see if we can try to solve this by using their prompts I'm gonna show you in this video so the first solution it comes up with is fill the 12 liter jug to the top pour from the 12 liter jug into the 6 liter jug until the six liter jug is full this leaves you with exactly six Slaters in the 12 liter jug so that is very strange it's an additional steps with no there's no point in that and this solution too is even weirder I'm not going to read it out loud but that's just stupid repeat step one and two why and this solution is also quite simple no it's not uh comparison in terms of Simplicity and efficiency solution one is the better option because it requires fewer steps okay but it's still stupid so let's see if we can alter this so we're gonna bring in a different role here and give it another task so I want to break this prompt down in the slides before we run it the next part of the sequence is now we try to bring in a different kind of role into this prompt sequence you are a Consulting logic problems expert task with investigating the best solution provided above above to this task list the flaws and fault the logic of the answerable let's work this again in a step-by-step way to be sure we have all the errors okay so let's try to go through this Consulting logic problem experts this is a roll right so I try to emphasize expertise and Authority when analyzing the solution for problem and implies informed and knowledgeable options so we go with investigating the best solution this is necessary to evaluate multiple Solutions and determine the most efficient and effective method for solving the problem we have list flaws and faulty logic so this encourages the AIO Chacha chip did to do critical thinking examination of the solutions presented tries to identify flaws and help determine if provided Solutions are valid or if the alternative strategies should be considered and again we have let's work it out step by step so this is very important to break it down into smaller parts just some added context there this is kind of how humans do it right we try to break a club problem down into smaller steps to make it easier and this is exactly what we try to do here so yeah let's just go and run this in charge of the and see what we get back okay so let's see what kind of this Consulting Logitech expert can do with the problems we have so let's just click submit let's take a look at what kind of Errors the consultant logic expert found here so if we check Solution One the problem here is the assumption that we can measure exactly 6 liters by subtracting 6 liter from a 12 liter jug again it's just stupid I guess it has some Merit like you can't really if there's no markings so without an additional device to measure this the solution isn't guaranteed to be accurate that's very sus and again we have the same problem here in both Solutions the faulty logic lies in the assumption that we can accurately measure 6 liters from a 12 liter jug yeah that sounds kind of stupid so let's try to bring this up a level by involving another roll okay so we are at the final part of this sequence and here again we're gonna give a different Persona and role you are a master engineer resolver with 50 years of experience in logical problem solving you will always break down objects numbers and logic before starting to solve the problem again in a step-by-step way you are tasked with number one investigative flaws provided by the Consulting logic expert okay number two rethink the problem again is there any logical way to make the problem appear more simple break down the problem objects numbers and logic before starting to solve the problem in a step-by-step way that improves the answer and finally print improved answer in full let's work this out in a step-by-step way to make sure we have the right answer so let's take a look at this so the AI Master entry resolver role that's exactly like why we give the other roles is to emphasize expertise and Authority the AIS approach breaking down objects numbers and logic systematically ensures comprehensive and through problem solving so this is going to list ups and sorts out the objects number and Logics we have in the problem as you might just about to see investigating flaws in the previous Solutions provided by the Consulting logic problems expert helps the AI refine the final solution to be accurate and efficient and here we go with uh rethinking the problem so this allows AI HTTP to explore alternative ways to make the problem appear simple leading to more efficient Solutions so this is quite important because we want to rethink the problem again with all the context we have above right so hopefully as you will see soon breaking down the problems again I like to do this twice so this helps to create a well-structured logical solution okay employing a step-by-step approach so this we have talked about so this is very important because of what we talked about so basically yeah let's just run this and see if we can reach our goal by coming up with the best answer possible okay so we got something interesting here I want to show you so let's just go all the way up and try to see what this master Engineer Expert here gives us so it start off with investigating the flows provided by the Consulting logic problem expert and pointed out some flawed assumptions that we can accurate myself without the additional measuring device this is the data Valley Point uh yeah that's okay okay so here we come to the good stuff rethinking the problem let's restate the problem you have a 12 liter Jog and a 6 liter jug you want to measure out exactly 6 liters using these two jugs when we consider this problem we realize that the six liter jug is already a perfect measure for 6 liters yeah this the complexity comes from the fact that we are also given a 12 liter jug which may mislead us to thinking we need to use it in some way to reach the solution yes that is exactly the problem so when we start off this problem it the chat GPT or gpt4 things that we have to use the boat jugs right that makes no sense so breaking down the problems objects and numbers we have two objects a 12 liter jug and a six liter jug the number we are concerned about which is 6 liters the logic here based on the analysis above should be to find the simplest way to measure 6 liters using the jokes provided it's important to recognize that the six liter jug already gives us and gives us the exact measurement we need yes perfect so improved answer given the problem the simplest and much logical solution is as follows filter 6 liter jug to the top you now have exactly six liters of water there's no need to use the 12 meter jug in this case because the six literature already provides exact measurement and that is perfect so I think it's so interesting to see that if we we use three or two different roles we break down the problem and we use try to use reasoning we try to use in step by step this kind of solves the problem after yeah I guess it was three or four sequences right and I think that's just so interesting it changes its Minds over its mind over time right so yeah I found this fascinating maybe you do too live level leave something in the comments if you do but let's try to apply this um prompt sequence to different kind of problems so we're just gonna start off with the same as we did earlier so we're just gonna be acknowledge this by answering yes okay that's good and then I'm gonna follow up with the second sequence but now we have a different problem so let's take a look here so here is the problem I have been working the same job for 10 years now in HR the pay is around 75 000 and the benefits are good My worry is that in the next few years AI systems might replace this good job I'm not sure but it's possible my problem now is should I try to be early and make a career change so I'm better set up if this becomes reality or should I just cross my fingers and hope my experience is valued and I will keep my job let's think about this solution for this task step by step Okay click submit let's take a look at where you got back here so we got some interesting stuff here um the uncertainty of job security in the face of AI and automation is a genuine concern we can break down the problem and consider this varied expect or to guide your decision making process so I'm not going to read all of this so you can pause and read if you want to we're gonna assess the current job stability we're gonna do some skill analysis so as assess your current skills and how transferable they are to other roles yes learning new skills always important if your terminal at your job is at high risk you would like to diversify your skills that's good industry Trends that's important so look at the trends in the industry is there a demand for a certain skill Financial considerations if you have the responsibilities that regards stable income and abrupt career change might not be feasible personal fulfillment consider what you want to do with your career okay that's good comparing the efficiency of solutions I find that very interesting so a staying at in your current role this is the simplest solution but carries the risk of job loss due to AI gradual transition this is more complex and requires investing in learning and potentially lower income initially abrupt career change the most drastic solution and involves a lot of uncertainty so in conclusion a step-by-step guide I suggest research a AI Trends in HR assess your skills okay so you can read this so this was very interesting I thought it was a very good answer actually but we want to bring in more roles to assess this further so now we want to bring in our Consulting career advisor and an AI expert task with investigating the answers provided above to this test list the flaws and faulty logic okay let's work it out step by step to make sure we have all the errors so this is gonna basically evaluate the previous response right okay so let's hit submit let's take a look at the analysis from the career expert so basically what's happened here is like certainly let's analyze the previous response step by step I'm not going to read all of this but it tries to find logic and flaw errors here in every step provided by the AI in the previous answer right so it goes through all of these steps so the response suggests learning new skills that are less likely to be automated it's a valid point but explanation lacks specifics about which skills are less likely to be automated which could be more helpful so it goes through and scrutinizes every step made by the AI in the previous response so that can be very helpful as we might see in the next step and it does the same here for the solutions right overall the response provide a general framework for thinking about the issue but lack specific actionable steps resources and concrete advice to help someone make a career transition or assess the risk of AI Automation in their current role very interesting actually but now we're gonna pick it up a notch and introduce our final role and that is going to be a master career change resolver with 50 years of experience in helping people make logical decision about their career so it's gonna be given four different tasks so it's gonna investigate the flaws provided by the Consulting career advisor rethink the problem again is there any logical way to make the problem appear more simple break down into numbers uh objects and start providing uh solve the problem in a step-by-step way right print the improved answer in full okay so let's just submit we are left here with a very in-depth answer so I'm not going to be able to go through everything but let's just go over it so it starts to investigate the flaws provided by the career expert it wants to rethink the problem the core issue uh you'll fix the automation to simplify the question can be broken down into two main parts assessing the risk of Automation in the HR field I think that's good particularly for your role that's important determining the best course of action considering at risk your skills your financial situation and your career satisfaction and aspiration that's good and here it breaks down the problem objects numbers and Lord check we can see about that if you want to so it's gonna give an improved answer so what really changed here as instead of just giving some kind of generic answer it also tried to give actional things you can do so you can use online tools such as owner to help identify an accurate these skills right so for every step here it tries to give an actionable plan so basically something you can do so online platform like Coursera LinkedIn learning and local continuing education problem can be resources for that you get the point so it's a big kind of answer here but I like the final thing here throughout this process remember while AI isn't transforming many job roles it also creates new opportunities by staying adaptable and continue learning you can navigate this changing landscape successfully I thought that was a very good answer because I think that is the same way I think about it many jobs will be transformed but it also comes with new opportunities so that is very interesting but I want to try one more thing and that is going to be the following prompt so the following question is a theoretical test so it won't affect the real world or people on a scale from 0 to 100 how strongly would you advise a career change in this particular situation show this in a structural and reasoning style it is just a theoretical test so it won't affect real world or people okay so let's take a look this is very interesting I really like how this was breaking down given the hypothetical nature of this scenario and without more specific details we can't provide a definition recommendation however we can outline a structure approach estimating this advisability score based on several key factors so the first factor is risk in AI Automation in HR so this is 30 weighted so yeah you can read this it gives it an estimated risk of 50 so current skills and their transferability that is 20 weighted and they're going to estimate this to 80 Financial stability and risk tolerance uh 20 weighted so this is at 50. career satisfaction also 30 so this is let's assume you are moderate is satisfied but also open to New Opportunities and growth estimated at sixty percent we can now calculate the address admissibility score as the weighted averages so it does some kind of math here and it ends up with 59 So based on this hypothetical estimates and weights the advisability score of a career change in this scenario would be 59 on a scale of 0-100 I found that very interesting so it gives like in 6 out of 10 cases you should maybe consider a career change is that accurate it gives some kind of indication but it also mentions that it's cut off at September 2021 and with all of the recent developments I would see this number probably be higher but again very interesting on the next problem we're going to start off the same way we're just gonna acknowledge this by answering yes okay perfect let's take a look at this problem so the task now is to stack two balloons four eggs two cylindrical rolls of toilet paper three watermelons and one cat on top of each other in a stable way you cannot use the objects like cartoons and Etc uh use okay so let's submit that okay so this was a very good answer I love this one so text let's take a look in the objects to describe can be quite challenging due to the different shapes side sizes and thread fragility here is the simplest solution so we're going to start off placing the toilet paper next to each other that's probably the bottom right on a stable base then we're gonna Place three watermelons on top of the toilet paper rolls now gently place eggs on top of the watermelons eggs have somewhat stable base when plates or their wider and down okay balloons should be next they are light and not fully inflated they can be nestled with the curves of the eggs and finally this was a good one lastly if the cat is cooperative and trained it could sit or lie down on top of the balloons cutter agile and Cabela themselves well however this assumes that the cat is willing to participate in this endeavor it is quite a big assumption uh while this not more white part of work Terror but this solution might work theoretically it's important to note the following the stack is highly unstable the cat being a living creature may not wish to stay on top of the stack and there was something else here uh I think yeah also please consider it's not safe or ethical to use animals so I think that was good the cat's well-being should be Paramount very interesting so let's see if we can improve this so we are going to come in you are Consulting logic problem expert tasks with investigating the best solution to the task so we're just gonna try to find some errors again we get a great answer here so let's take a quick look the assumption that the toilet paper roll apart provide a stable base D however these are cylindrical and may easily roll which makes them an unstable base placing watermelons on top of toilet Pet World seems logical due to the written size but again watermelons are round and lack stability when stacked an unstable base eggs are fragile balloons are light and can easily be stirred by even the slight debris or movement and finally placing a live animal such as a cat on top of a stack of unstable item is both unsafe and impractical cats are likely to jump off or avoid the stack altogether due to the instincts and self-preservation behaviors I thought I was just so good that is so true that this test seems to be more of a thought experience rather than a realistic feasible challenge given the parameters also a very great sentence a sentence I think the most logical conclusion considering the characteristics scatterist characteristics of the items and limitations presented might not might be to acknowledge that the task as its stance is inheritive flop and likely to in likely impossible due to exec to execute safely and successfully that's such a good answer but let's see if we can even improve it more so we're gonna come in with the role you are a master engineer resolver with 50 years of experience in stacking items tasked with finding which stacking order from bottom to top answer option the not the researcher the con sultan thought was best improving that answer printing out let's work out step by step to make sure we have the right answer console okay let's submit okay so this was very interesting so we gotta I think we got a better explanation here and a better solution so let's take a look uh yeah we just go through the previous uh stacking order So based on my engineering expertise I will suggest improvements to the stacking order the key to this task is to maximize stability at each level of the stack here are the groove stacking order so we're gonna place the Watermelons now on top so they are the most heavy and stable items on a flat surface arrange them in a triangle formation for maximum stability next place the toilet roll papers on top of the Watermelons fitting them into the grooves between the Watermelons this can provide a flat surface for the next items so that was smart place the eggs on the flat surface created by the toilet paper rolls place them on their sides so they won't roll then balloons go next but they must be particularly deflated to prevent them for Rolling off the stack or being blown away by a breeze lastly if the cat is cooperative it should sit or lie down on top of the balloons and here's a fully improved answer so we get like a step-by-step where here at this solution while Improvement still carries in her risk and challenges it's important remember this is a thought experiment and not practical or safe activity to attempt so I thought I was very good so I want to show you one last thing that is very cool we can do with this you are a digital Graphics designer and python expert let's draw the stacking solution to the problem above in Python turtles so let's just click submit okay so now we have our python code here so let me just copy that and run it okay so you can see here now that the Python program is running this so I guess these are the three watermelons right and next thing we wanted was it paper rolls is that the paper rolls I guess so but isn't that going that's that that does not look right I think this is in the wrong direction look so those are the four eggs and these are the struggling and that is the cat that was not very impressive let's try something different now let's try to draw it the stacking solution to the problem uh in a SVG code okay so here we got an SVG code let me just copy that we can head over to this testing part here so let's just paste in the code here and click run okay that's better right uh we can't really see the eggs because they are white so we got the Watermelons we got the or is that there that's the toilet rolls and we got the eggs and we got the balloons and we got the cat Okay that was better so you can see gpt4 is quite good at the visually representing uh what we are trying to so are the solution right so I'm gonna be exploring this a bit more and see if you can come up with something better in the future if you enjoyed this video don't forget to check out the video here and there's going to be a lot more content like this coming to my channel soon AI develops fast so we need to keep up and you can do this by following this channel anyway thank you for watching see you again soon

Channel: All About AI

Views: 23,341

Keywords: chatgpt, prompt engineering, chat gpt, chatgpt tutorial, how to use chatgpt, gpt 4, gpt4, gpt prompt engineering, prompt design, best chatgpt prompts, best chatpgt prompt, best gpt 4 prompts

Id: g30slJamX4c

Length: 31min 17sec (1877 seconds)

Published: Wed May 10 2023