Great Writing with ChatGPT | A Complete Hands-on Guide

1. Course Overview: Hey there, welcome to the course. My name is Luke and I will be your teacher throughout this course on creating great writing with chat GPT for anything, for whatever you want to create. You will see me throughout the course like this. We will be spending most of our time in this view going through prompts, going through the different approaches we're going to cover in this course. In this first lesson, I would just like to quickly go over what you can expect to learn in the course, what you can expect to get out of the course. So first we're going to get into some core principles, things to bear in mind as you're creating with chat GPT, then we're going to be getting into our three main approaches. These are fairly broad. We're going to be learning them by example, looking at specific prompts or steps with multiple prompts. But more importantly, you're going to be developing an intuition for how to apply them so that you can deeply understand which approach to use for what situation or how to mix and match as you need to customize your own prompts. Because if you think about it, even to very similar tasks, because they're not exactly the same, require slightly different prompts or multiple prompts to get the desired results. And so you need to learn how to become a prompt engineer. You need to know how to build your own prompts. And as you gain this intuition, we're going to be going through lots of prompts, lots of examples, lots of specific cases. As you build this intuition, you're going to start building up your own prompt library. You're also going to be learning to use chat CBT as a genuine writing assistant to not just create great writing, but also develop ideas, come up with ideas of course, but then develop those ideas that you can then turn into exactly what you had in mind when you decided to write that thing that you've been trying to create or that you don't want to spend 5 h creating. We're going to work on some advanced tips and tricks, including how to repurpose content. We're going to focus on how to automate work and simplify a lot of tasks, especially writing tasks. So we're gonna be looking at a lot of examples and prompts there. We're also going to be going over how to use chat GPT to get really good corrections and feedback and how to tune those knobs so that we can get exactly the type of corrections and feedback that we want. Knobs will be a key word actually throughout this course. So that includes, of course, simple corrections. And there are a few ways to do that, but also high level feedback. Something like a proposal that we can then take and integrate back into the original version to come up with a much better version. Again, without spending too much time on that. We're also going to be looking at how we can get specific feedback according to or against a rubric or standard. We'll be going over that toward the end of the course. And again, throughout building this intuition, creating our own unique prompts for our own unique situations, for the things that you specifically want to accomplish, building that prompt library so that you can save time rather than wasting time just randomly trying stuff. This is what we're trying to, this is what we're trying to do to get efficiency and productivity without sacrificing quality. And by the end of this course, you will be a prompt engineer. You'll have many prompts to start with, many prompts to use. But you'll also know how to build your own prompts and by the end of the course, you will have your own prompt library. So I hope you're excited to get started. I certainly am. In the next lesson, we're going to be going over a few core principles before we get into our three main approaches. So I will see you in the next lesson. 2. Core Principles for Writing in ChatGPT: In this course, we're going to focus on three main approaches. These three main approaches which are very flexible, will allow us to do basically any writing task very well. But before we get into these approaches and start working with them, I want to create a bit of a foundation first and cover some core principles very quickly that I want you to keep in mind throughout the course that I think will help you as we go through the course and as you continue to work on whatever it is you're working on to really make sure that you have that high level of quality. So that you're not just creating that generic stuff that you get with a basic prompt. We all know how to use chat GPT in that generic way of writing a resume for me, right? But go ahead and try that right now. Write a resume for me. How does that look? Right? Well, it doesn't have any of your information. And if it did, is it exactly what you wanted? Does it communicate who you are? Right? Is it you that cover letter? Is it you well know, it's kinda generic even though it has my information, right? So the first thing that I'd like you to keep in mind is this idea of being the editor in chief. That means you have a vision, you have requirements. You have this thought in your mind of what you're trying to achieve. And now you have a tool that can help you achieve it. So don't let the cart get in front of the horse there, right? If you start thinking, well, I guess chat GPT or bar, I guess they know best. Not necessarily you are the one making the decisions. You're the editor in chief. Now you have a way to maybe save some time. Now you have a way to get that result that you wanted and do something else, right? So keep that in mind. I think that's important just as a sort of perspective thing. But also remember, if you just ask that generic question, what do you expect other than a generic answer? Write a generic prompt generally leads to a generic response. And what do you expect? What else would it be? Garbage in, garbage out. It's the same way if you're trying to give requirements to chat GPT-2, any AI. If you give it bad information, if you don't specify things clearly, then it's garbage in, garbage out. You can expect anything else other than crap. And so I hear this a lot. People I know sale. Well, I don't like the quality of the writing. Okay. Well, show me the prompt you used and they show me and it's, it's crap. So what do you expect? Keep that in mind throughout the course. Keep that in mind. Also. This is gonna be one of the core things that we're using in this course. Outlines are gold. If you know how to do a good outline, even if it's a bullet point outline. If you have a good outline that is sufficiently detailed, there is so much you can do with that. And to check GPT, that is also gold, right? So just keep that in mind. We're gonna be working with that much more as we go throughout the course. Now, what about being specific? What about making sure that our requirements are actually communicated? Well, that sounds kind of obvious. I think it's obvious, but it needs to be said, right? If you don't say what you want, other people don't live in your head, right? So if you don't say what you want and other people don't do what you were expecting. You can only really blame yourself, right? So if you get back something and it's oh, that's very well-written. I like that, but that's not really what I wanted. That's not the stylist, not the tone or the voice. What do you expect? Honestly, if you're not specific enough. So maybe even if the output is great and it still doesn't quite match what you were going for. Maybe it's not because of the AI and what it was thinking. Maybe it's because you weren't specific enough, right? If you take photos and you don't know how to set everything correctly, you don't know how to set the correct shutter speed or the correct ISO. You're not going to get the results that you want. But if you know how to set those things perfectly, then you will. It's that simple, so leave nothing to the imagination unless that's what you want to get out of it. Unless imagination is what you're looking for, in which case, yes, leave much to the imagination, right? Also, there's this basic thing that we do as humans, where if you're training with me and I'm a coach or something like that. And I'm going to teach you how to do something. I will first show you how to do it. I will say this is what I want This is an example of something that's good. What I think is good. Now can you do it rather than just letting you do it first, right? That's how training works. So that's very important to keep in mind as well. If you want to create something, make something, write something, write a very good email or proposal, right? Whatever it is, if you have in mind what you want it to look like in the end, Show that to the AI that you're working with so that it can learn that and then start working. And there's a specific way to do it. And we're going to go through that. That is one of our three main methods, but there is a specific way to do it. So we're going to, we're going to get into that in the next lesson actually. So hang tight. We will get there. Just a couple more things here. Always iterate, right? If you see the first response and it looks good. My strong recommendation is don't stop there anyway. Hit the regenerate button a couple of times, keep going, right? Because if you keep going, you might get another piece that is actually quite good and you didn't think of it that way. The first thing that you see isn't necessarily the best, even if it looks pretty good. So always go through a few iterations. Unless you're very short on time for the sake of getting the best possible quality, right, and be prepared to take something from one and something from another. It's always this back-and-forth process, right? Keep that in mind as well. It's rarely just here's the thing. Result. Got it. Copy paste. I'm good to go. It is this back-and-forth thing. And so think in terms of steps rather than prompts. You see everywhere these prompt libraries are guides which are great, that's fine. But I think they miss something very important, which is often, in most cases, it's not just one thing you have to be thinking in terms of several steps that get you to the result that you're looking for, right? That gets you to a beautifully written email. That gets you to a really compelling paragraph in an essay, right? So we want to be thinking in terms of steps, flows. You could think of it that way. Or a process rather than a prompt copy paste done. Because that gets us back to that generic stuff, right? And also very important, whenever you need to do, it is probably slightly different than that thing that you needed to do yesterday, which was actually the same task, right? So always be ready to make slight variations to the prompts that you choose when you're working with any AI. But in particular chat GPT. Because if you want good results, you have to be specific. And if there's any difference between that thing you needed to do yesterday and the thing you need to do today, then the prompt needs to be different. Even if that's a one-word tweak, right? And if you can build really specific great prompts that are perfect for what you need to do. Then you can build up your own custom bespoke prompt library. We're gonna be talking about that later. And it can save you a huge amount of time. You have 30 steps or series of prompts that you use for specific tasks. And you just copy and paste those. You've built them, you've customized them according to your needs. If you have those and you build those and we're gonna be talking about how to build those throughout this course. Then think about how much time you're going to save, especially if those prompts are giving you really good results, right? So these are just some things to bear in mind as we go. Now, we're going to get into our three main approaches for getting really solid writing with chat GPT. So I'll see you in the next lesson. 3. Approach 1: Outlines and Knobs: Now it's time to get into the real meat of the course. We're going to be going through with examples, the three approaches that we're going to use in this course to create great writing. The three are outlines and knobs, building from templates and the Star Trek computer. Now this is just what I'm calling these three because I think they're memorable. But you can disregard the names if you like, outlines and knobs, building from templates and Star Trek computer. Okay, we're gonna be starting with outlines and knobs. Important to keep in mind here. What we do, what we create this process for each of these. This is what we now begin to use as a framework and approach for any other thing. So once we go through these three, with three examples, then we're going to start to explore how we can use those same approaches for a lot of other different situations. For writing stories, for writing proposals, for writing emails, for writing presentations, for creating all of those other things. But it really boils down to these three approaches. Okay, so with that in mind, let's hop over here to chat GPT. Now I'm gonna be using GPT for, for now. But you can use GPT-3, 0.5, or maybe you're watching this at another time and there's another one that's out and it maybe looks different, but the same principles apply here. Okay, so let's start with outlines and knobs and talk about what exactly that means. Well, this approach is when we have a lot of information that we want to include in something, right? I have many pieces that I want to include and I know what those pieces are. I have the details, I have the specifics, right? And I want to create something, let's say bigger or longer with those pieces. I want it to be very high quality, of course. And really I don't want to spend the time to put in all the work to format it, to make sure every sentence is perfect. I want to get to something very good that includes all of the information that I can provide. I have a vision, I know what I want, and I'm able to specify what those things are, right? So this first approach of outlines and knobs, we're going to be writing an article. I think that's a great way to apply this approach. Writing an article because it's a complicated thing. Articles are quite long. They include a lot of information, right? This is again best when we're trying to get a lot of variation, a lot of creativity. We're maybe pretty sure what we want, but we don't know every single sentence Exactly. We know the kind of thing we want to see. We don't know every single little piece, right? So we want to leverage the creativity of chat GPT, of the language model, right? Right, okay. If you have longer, more complex writing, this may be the best approach. So what is the approach? Let's get into it. Alright, so here I'm gonna, I'm gonna first we're gonna go over basically the structure and then we'll look at the specifics. So the idea here is first we specify the task we use act as, and we state what we want to happen, what we want to create in some detail. Then we create a set of knobs. Think of these as knobs that you actually turn. You don't have to use all of them, but you can choose from these knobs. Again, if you use all of them, it probably won't come out, right? I never use all of them, but think of them as knobs because these are the things that you're turning, tweaking to adjust the input so that you can get a better output. So tone, do we want to have? And here you would put in adjectives. Do you want it to be formal? Do you want it to be casual? Do you want it to be engaging? Do you want it to be insightful? What about perspective? Maybe here you want to say the perspective of and maybe an argumentative or a polemic or something like that. Or you can simply state the perspective of first-person narrative, third-person narrative. Often, if you'd just put in a prompt, write something for me. You may get back An us sometimes in a week when you wanted an I and the me. Okay. Well, that's where you would specify that style then allows you to clarify the kind of things that you would like to see. And I'll give you some examples of that. But this is where you say, use concise language. This is where you say use some humor and analogies or don't use humor and analogies. If you want to make sure that there aren't any in there. This is where you have a lot of control over the kind of output that you get. Structurally. Maybe you want to have smaller paragraphs. Maybe you want to make sure there are section headings, so you would state those, their purpose. What are you trying to do? People who read this? What do you want them to feel? I want to inspire. I went to inform, I want to educate whatever that happens to be. Now, token limit. This is the length according to the language model. Rather than word count, you could put here length and then do a word count. Or you can do token limits. And if you e.g. put 50 tokens, that's going to be very short. So the essay you get is going to be a couple of sentences. Maybe if you put in a few thousand, then it's going to be much longer. You have to get a sense for what those are. And chat GPT, four different models have different limitations there, so that's something to look into. Then temperature is 0-2. Closer to zero makes it, let's say, less ambitious, less variation, less deviation, less flexibility and creativity. Closer to two, It's going to take more risks. It's going to be a little bit more trying things. It's going to be maybe putting something in there that you may not want. But maybe sometimes those risks are good because you get something really interesting. So that's something just to play with, right? Sometimes you want to include context. This would be maybe the context of why you're writing this, what you're creating it for, right? Not necessarily the audience, but this is the third article in a series about x, something like that. And then for SEO reasons, perhaps you may want to include some specific keywords and you could change this to keywords to include e.g. okay. So these are knobs and I want you to think of them as knobs, things that you turn like knobs. And what you put here makes a big difference. Then we put in the outline, okay, so the outline is where most of the work goes. The outline is where you have to actually provide what you want to be there, right? You have to put in the work here. If you leave it up to chat GPT, it's gonna make up whatever it wants. Okay, great. But again, is that, Is that what you want to say right? Now? If you put all of this in here and you say generate an article about the history of the English language. It'll be better than just generate an article about or create an article about the history of the English language, you're going to get a better output if you have the knobs and the act as the details after that, right? But if you have specific things that you want to include, then you need the outline. But we're gonna be doing this in two big steps here. So let's just, let's just do it. Alright. First we have our, our initial prompt and our, our knobs. And then we're gonna be putting in our outline. And we're going to just briefly go through this and then we'll go to the, go to the second step. Okay, so we're not gonna go through the whole outline. These, these are just details about that maybe I know as an educator, this is actually from a video on YouTube, History of the English language. Feel free to check that. That's on YouTube, but it's my video, right? So that's why I'm using this. So these are actual notes that I created. These are notes, these are notes that I created before I started writing my script for the video. These are the notes that I used. So these are the specific things that I wanted to say. Now, how long did that take? Well, it depends on the research. It depends on how much you know, but more or less detail. I tried to have more detail in there. Alright. Now I say act as an experienced, experienced writer, knowledgeable about the history of languages. My first sentence is stating specifically what kind of person, what kind of assistant I want to be writing this for me. Right. And you would change this for each one. That's why it has to be bespoke. If you're writing something about technology, then act as an experienced tech reviewer, writing an article about maybe the latest whatever phone release, write a compelling full article based on the guidelines and outline that I provide. That part always stays the same for me unless I have something specific to change. Use interesting and dynamic syntax This one is important because it helps the variety of the different sentences. If you find you're getting these repetitive sentences, every sentence kinda looks the same or to many of the same, then you want to have something about syntax in there. This should be a full length article and an in-depth exploration of the topic. Present the history of English and a conversational style making it easy for readers to follow along and understand the information being presented. Use insightful analogies. This is key. I want to put this in here because I don't want it to read like and then, and then, and then story, right? So the way to think about this is this first paragraph. You are explaining it as a job to someone as you would if you hired them on Fiverr, you hired someone to write an article for you. I want you to okay. I want you to write this for me and I want you to do it this way. And I want you to use interesting syntax and make sure it's very in-depth. If you were just talking to that person to explain what you want it this is this is this, right? And then the knobs are a little bit more, let's say technical or a little bit more granular, right? But the last part is key here for me. Use insightful analogies and touches of humor to make the article more entertaining and engaging. Because I don't want it to be this flat thing, right? I don't want it to feel flat and boring. So I want it to pop with a little bit of, a little bit of humor, not too much. And interesting analogies that help the Ideas click. Okay, here are the guidelines and then we'd go into tone informative, engaging humerus at times, perspective, third-person narrative. Use common language, that's the style. I don't want it to use flowery language. In fact, I could put do not use flowery language. Sometimes I do that, right? Right in full paragraphs, use occasional humor and occasional analogies. Temperature is 0.8. So I want it to, I want it to not take too many risks here. That's the general, the general idea, purpose to inform and entertain, use the following outline. So let's see what we get with the outline. There we go. The English language, like a box of chocolates, is rich in diverse flavors and its history as a saga filled with unexpected twists and turns buckle up as we embark on a linguistic journey through time. Exploring the evolution of English from its humble beginnings to its current status as a global powerhouse. Alright, then We get what I think is a pretty good looking essay. The history of English is a saga of invasions and the movement of people. When the Romans set foot on British soil, they brought with them Latin prefixes like pro and Sub, which would later sneak their way into the language, like uninvited party guests. There's one of those analogies I asked for. It doesn't seem too heavy, right? Maybe one per every couple of paragraphs. But interesting, and it helps make the idea clicks. So I like that, but I'm still just using this as raw material. So what have we done? We've explained what we wanted. To whom are experts in languages, right? We've turned our knobs and these are for me, maybe that's not the style you want. Great. Figure out your own knobs. That's what I'm saying. This is what I want to see. This is what I want. This is my writing style. I've tuned to my knobs, right? That's why it looks like this. And I've spent many, many hours figuring out how to turn these knobs so that it matches how I write. But this is very, very close to how I write. So we've got this as our first draft. Great. These fierce Scandinavians contributed words like reindeer, Kindle, and Thursday to the mix. Old Norse accounts for 1% of modern English. A small but impactful contribution of around 2000 words. Accurate, clear. I think engaging syntax looks good to me. I think it's pretty solid now. Yes, I could just copy and paste this, but remember what I said. We don't just do that. We don't just copy and paste. We iterate. So what I would generally do here is do another and see if I can maybe get a different analogy. Maybe I'm not crazy about the chocolates one. Well, if I'm not crazy about that and I don't know which one I want at the beginning, then I would just get another version of this, right? And maybe there would be a different analogy. If I know which analogy I want, then I would put that analogy in somewhere above in the knobs. Perhaps. Make sure to include an analogy about biological evolution in the first paragraph. State that very clearly, I don't care. I want to leverage chat GPT is creativity, so I want to see what kind of analogies it can provide. So let's see if we try another one of these Picture the English language as a magnificent patchwork quilt. Each piece of fabric represents different linguistic influences and historical events that have shaped the language into what it is today. With threads woven from Old Norse, Latin, French, and other linguistic origins. English has a rich and diverse history that we'll explore together in this article. So grab a cup of tea, sit back, and let's dive into the marvelous world of the English language. I like this one better. To me, the quilt analogy works better than the chocolate analogy. That's pretty good. So what I'm going to do is I'm going to take this one. And I'm going to use this to then create my second version. Now you can do this or not. But I generally see very interesting results when I do this. So when we go from outline to first draft, we can see that as raw material for the second draft. And we can do it immediately. Okay. So I'm going to just copy and paste my second draft, my quilt draft right there. And I'm going to take the exact almost the exact same prompt that I used for the outline with some slight variations. And I'm going to use that with the draft rather than the outline. This will allow me to use that as the raw material rather than what I consider to be final. Now I may still take things from this draft and use it in the final one. But that's where you have to get creative and ultimately make those decisions. You're a curator of the language that you get rather than having to create it yourself. Dealing with the blank page syndrome of steering and not knowing what to write. Right? So I have put in again, act as an experienced rider, knowledgeable about the history of languages. I'm using the same thing here except for one thing. Here. I've put Draft instead of outline. And the first draft that I provide. And then below right before I paste in this, I'm going to say here, use the following first draft instead of use the following outline. Alright? Imagined the English language as a vibrant tapestry. Each thread representing various linguistic and historical events that have shaped it over time. This tapestry, woven with the strands of Old Norse, Latin, French, and more, has given English a rich and diverse history. So pour yourself a cuppa, get comfy, and lets unravel the fascinating tale of the English language. Can you see how this one is more interesting than the first one? Now, I can't pretend to understand exactly why this is, but it seems to be the case. Generally speaking, when you're going from outline to first draft, you are getting, let's say, pretty good writing. If you have tuned your knobs well, if you haven't, it's going to come out as fairly generic. So you have to really tune carefully. But it's not quite that Wu, very punchy, interesting. Something that's going to maybe make you laugh, sometimes really clever language. But if you feed that first draft back in using the same or very similar prompt, then it's almost like that bandwidth that it was using before to turn the outline into pretty good writing, is now used to turn pretty good writing into very good writing. I mean, this is quite nuanced, right? It's using the analogy, but it's also using the analogy throughout the first paragraph, tying it in the concept of the threads and using something a little bit more. Let's say, I don't know, casual like cuppa instead of cup of tea, right? Which I think is a little bit more playful and I actually like that now, we're not going to read through the whole thing. But having done this many, many times with many different writing samples and worked for many hours on tuning these knobs. I can say most of the time you get good results. It's still important though, to see what another version gives you to hit, regenerate and iterate a little bit, and then take a paragraph here and a paragraph there. You're still the editor in chief. You're still making the decisions. I don't recommend wholesale copy pasting into wherever you're putting it, right. But I do see a much, much higher level of quality with this process. What I would ask you to do is try it yourself though, and try to get a feel for it first. Do it with exactly this, exactly what I have provided. And see for yourself right? Next. Try it for yourself with your own content. See if you can create a really interesting piece of writing using this process. Because if you can get a feel for it and you can tweak those knobs and find your style up there it is. My style is engaging and my temperature is a one-point to write. If you can tweak those knobs just right and find your style, wouldn't it be great to be able to create writing that sounds like you're writing, That's actually good. Instead of generic, of course, of course. But this is just the first of our three approaches. Now that we have gone through our outlines and knobs approach and gone through a specific example. Let's go on to our next approach, which is building from templates. So I'll see you in the next lesson. 4. Approach 2: Building from Templates: Maybe during our exploration of outlines and knobs, you were thinking at some point. This seems like a lot of work. I have to provide a fairly detailed outline to get really good results. Don't have to do this every time. Is this the only way? Well, no, it's not the only way. We're covering three main approaches. And once we look at an example of each approach, we're going to then talk about, okay, how can you take these and apply them to every other possible situation which you may encounter, every other possible thing which you may need to create or right, right, right, right. Now. The next two are less complicated than the one we talked about before. Then outlines and knobs. That one again is when we want to really fine tune the results and we're writing something that's a little longer. But everyday communication, e-mails, copy for something, right? Maybe sales copy, e.g. shorter SEO content, social media posts, maybe a performance review for a colleague or ten performance reviews for ten colleagues. These are not the best use cases for the outlines and knobs approach generally, because they're shorter and often they're a little bit more repetitive. We have a style that we like to use perhaps, but it's similar each time. And so we need to be able to use that similarity. Can we? Yes, Yes, we can. Now we're going to talk about our second approach, which is building from templates. What is a template? A template is something that allows you to do several versions of something, perhaps with different content, following a similar structure, a common pattern. Now, this gets a little complicated because the way I think of templates, when we're writing with chat GPT, that also includes style, that also includes the voice of whatever you're writing, right? It includes the structure of the paragraphs. And we also want to have some variation. We don't want it to be so rigid that we can't use it for different content, right? So let's get into it and see how we do this. And we're going to be using the example of a customer support e-mail or a customer service email. So here we are back in chat GBT. And we're going to be instead of, again, going through this whole outline thing, we're going to be teaching, chat GPT, a specific style of thing that we want. But the nice thing about this is, for the most part, we do not have to specify all the details of that style. We asked chat GPT-2, learn all of that for itself. And the farther you get into this, the more you realize how powerful this is later in the course are going to be getting to some advanced principles, some advanced ways to use this specific approach. Very powerful, to be able to train two very simple training for make it look like this or make it like this. And then make ten things that follow that basic format and style and voice and everything. Wow, if I could do that, I would be able to save a lot of time, right? So how do we do this? I'm just going to dive right into the example. So here is an e-mail. This is an email that is being written to a customer. So let's say I own a pet food business. Sometimes I have to deal with customer complaints. Things happen the wrong order. They got charged too much. All kinds of different things happen, right? But I want to respond to each email with the same style. I don't want to be jokey and say, Ha-ha, well, too bad for you, right? I want to be respectful. I want to have a professional tone. Maybe for my company, I like to use I, or maybe not, maybe for my company I like to use we. And that matters, right? Some companies, it's eyes, I'm companies. It's weak when they talk to a customer. And maybe I want to be a little conversational but not too informal, but that's it. Difficult thing to balance. But if I could just capture that style and repeat it, that would be great. So learn the structure, style, tone, and voice of the following example. Once you have learned it, say I understand, do not add additional information. The example is meant to be used as a template for future emails. The wording here is very important. So generally you want to keep the wording here. Don't change it too much. And let chat GPT figure out how to do that correctly. I will then provide the details for the next email I want you to write So all we want to get here is I understand. And you can change this. You can say, oh, I don't know, Let's change this instead of I understand. Let's say peanuts. I wanted you to say peanuts. Do not add additional information. Now let's change this to a turkey sandwich. Alright, Let's say turkey sandwich. Now, we then provide a sample. So this is not necessarily where you need to put in work, but this is where you need to show what you want. Here is the email example. Thank you for getting in touch regarding your concern about order TR 7843212, you are absolutely correct that the amount billed was not correct. Okay. So this is admitting that there was an issue and fixing that issue and offering a discount code on future orders. And saying Sincerely, I lean more specialist. This is a fictional, hopefully fictional company for pet food. So what we want to see when I hit Enter here is turkey sandwich, and we'll just change this to chat CPT4. It's always a little bit better, isn't it? A little bit, a little bit. I mean, they're very similar for this sort of thing, but why not? Turkey sandwich? There we go. That's what we wanted to see. Okay. Now what we want to do is see if we can get an email that matches that in the right tone, in the right style, right? That's what, that's all we want to see. Now, if you just say write an email apologizing to this customer, the issue tends to be and this is the issue that I generally see. The issue is it's all over the place because you weren't clear, because you didn't specify, because you left it up to the imagination, which you don't want to do in this case, then you got inconsistent results. It's number one, probably generic. It's fine. It's correct, but it's generic. It doesn't match your style. I instead of a bunch of these annoying things, these tiny little things that are just not quite right. Okay, So now we put in next email and then we give a little bit of information, right? We have the confirmation. We give a little bit of context here. So all I wanted to say as next email, write an email apologizing for the incorrect cat food being shipped. Do not make excuses for the mistake. Assure the customer that a replacement will be sent free of charge. Mentioned that we are sending a bundle of cat treats as an apology order number. Then I put the order number, correct product, tasty tuna crunch, incorrect product, tasty chicken chow. Guidelines use respectful in friendly language. Now I'm just adding that as a sort of reinforcement beacon. Generally leave that out. But notice this is way simpler than our previous approach, right? This is, here's what I want. And you don't have to stress too much about writing this well, because this is just your request, right? So that's the nice thing about it for me, for this one, once I've given the template, I don't have to think carefully about, alright, I'm writing to this customer, I need to make sure it's really well-written. I'm sure we've all in this kind of situation, writing a formal or important email, spend an hour on every sentence. Alright, well, I don't have to do that. This took me 15 s to write because I'm just, I'm just talking to the machine who's gonna do it for me. Right? Okay. So let's see what we get here. Hello. Thank you for getting in touch. We're sorry to hear about the incorrect product. You receive an order and then it's the order number. We understand that you were expecting tasty tuna crunch but received tasty chicken chow instead, we apologize for this mistake and want to assure you that's one of those formal words. Assure you is definitely sort of that formal one which matches the tone of this sort of beginning here. Right. Thank you for getting in touch regarding your concern. That's that same sort of professional language that's kind of hard to, kind of hard to specify, right? Not very easy to state exactly. Now also, is it using the same person, okay. Due to human error, error on our side, we overbuild you. Okay, so it's using our and we, and I want it to copy that and use the same general perspective. So it's doing that too. We apologize for this mistake and want to assure you that we will send the correct product, tasty tuna crunch to you right away at no additional cost. As a and I love this one as a token of our apology, were also sending you a bundle of cat treats. We hope your furry friend enjoys them. Now that's the sort of playful thing that's not to jokey or silly. That's hard to specify. That is, in the original e-mail. In the original email, there's a touch, a tiny touch of playfulness, friendliness. That's not super serious formality. That is hard to specify with the knobs that CBT is able to glean naturally when I ask it to learn it as a template and it's perfect, I would be able to send this, I think without any revisions. Maybe I would change one or two things. Maybe I want to state the name of the cat treats of tiny, crunchy Turkey cat treats or whatever they're called. This is a fictional cat, cat food company that I made up. But the point of this is, this approach allows us to create a style without actually having to specify that style. And that's powerful. Now, can you use this for longer written content like the article? You Can. I get much better results with the knobs for longer things like full articles or blog posts. For this sort of thing. For an email to someone, a professional email, a personal email for something like YouTube descriptions, which I use this method for. And I will show you an example of that later in the course. I think this approach is really great. Now we're gonna be using this in a few different ways later on and you'll see how it works. It's very cool, very powerful in my opinion. Try it out for yourself. You can use my exact prompts that we practiced here, this specific example and then try one of your own, doesn't have to be an e-mail, can be something else, right? Use the same way to train it as a template maker and then ask it to make a new one of something similar to that template. Did I say that correctly? That's the idea. Alright, try it out for yourself. Let me know how it goes. We're now going to go on to our third approach, the Star Trek computer. So I'll see you in the next lesson. 5. Approach 3: The Star Trek Computer: So far in our three approaches for creating great writing with chat GPT, we've gone over outlines and knobs and building from templates for outlines and knobs generally, we're going to be creating something longer. We want to have that ability to tweak things, to adjust it, to create a model that matches our style of writing or any style of writing, that high level of precision for building from templates. We don't want to have to specify all that stuff. We want to show chat, GPT something and say learn this, and then say, do another thing like that, only with different information, right? And that makes it highly repeatable. That makes it really great for those daily tasks, like emails. But our third approach is very different from these two, but still very powerful depending on what you're trying to accomplish. And this is the Star Trek computer. If you don't know what the Star Trek computer is, that's okay. I'm going to briefly give some background and also explain why I think it's so important to know how to think in this way when you're working with chat GPT. Then we will get into our specific example. So the Star Trek computer, people on a spaceship in the future. There's a computer, right? It is, of course, far more powerful than chat GBT. It is a powerful artificial intelligence that basically runs most of the ships in the future. And there's this one episode where the crew are being abducted during their sleep. They can't remember very specifically what happened, but they have these sort of foggy memories. And eventually they get together realizing that we've all been abducted. What's going on? They get together and they go to this place called the holiday deck, which can create things from energy right in the holiday. And they have this conversation with the computer to try to recreate their memories. They're doing it collaboratively and in an iterative way, very important, in an iterative way. So one person says, I think there was a table. Yes, there was a table. A table appears, the computer creates a table. But then another person says that man, not quite that kind of table. It was a table at an angle. Yes, it was at an angle. So the table disappears and a new table is created at an angle. And over time, through iterations, they create what they all remembered, but they weren't quite clear on, but now they're all looking at it exactly. And that allows them to have an insight that accomplishes something that allows them to solve this mystery of who's been abducting us. And they're doing it in this iterative way with the computer. Now, why am I saying this? Well, this is fundamentally what the approach is. There's a specific way to do it. But the creators of chat GPT have indeed talked about specifically the Star Trek computer as inspiration for what chat GPT is meant to be. It will change over time. It will become more powerful of course. But that is what it is meant to be. This computer that you're talking to that is helping you evolve your ideas to bring your fuzzy thoughts into clear reality. The CEO of OpenAI specifically mentioned the Star Trek computer in an interview that I saw a couple of months ago. So it's what it's meant for, right? So what is this approach and how is it different from the other ones we've talked about? Let's head over to chat GPT and talk about that. So here we are. What we're going to do for this approach is we're going to make a request without supplying a lot of specifics. We're going to ask chat GPT-2 take on a role. But then ask chat GPT-2, ask us for things for information over time. That is the key to this one. And through this back-and-forth, we build up the information needed to then create whatever we're trying to create together. Not a table, but e.g. a. Cover letter, resume, or there are a lot of other ways to use it. We're not going to get into in this course that are not specifically related to writing. Perhaps feedback on your English, e.g. right? So let's try this out. Let's do this. I'm going to do this with speech So I'm going to talk to chat GPT because I want to make it seem more, I don't know, more casual, more freestyle. I don't want to have to type all of this out. So bear with me. This is going to be a conversation there. I'm using a, a, an extension in Chrome to do this. I'm not promoting any specific one. There are a couple that work. I'm not saying one is better than the other necessarily. I'm just using one of them. Okay? Okay, So let's paste in our prompt here. And then it's gonna be a back-and-forth conversation. Please act as an expert cover letter writer. We want that expertise, we want that focus and direction. So state that in the first sentence, very important before you write my cover letter. This is also very important. Again, the wording, the structure of this, it does matter. You can't just randomly say it and get the same results. You have to be very specific. I'm not saying this is the only way to say it. I'm saying you have to be specific in the right way. Before you write my cover letter, ask me specific questions one at a time. Only. Ask one question at a time and wait for my answer. So I don't want to get five questions. It'd be overloaded and have to provide the answers to five questions. I want a back-and-forth conversation. That's the point of this. That's what makes it back-and-forth. That's what allows it to be iterative. Once we get a result, once we get this, we can then give some feedback. This back-and-forth process is what we want. Don't ask more than ten questions. You're going to make that 54, whatever, once you have all the information you need. Right. And impactful cover letter. If you want to be more specific there, impactful professional put some adjectives in that you want. But here we're going to leave the exact style and format up to chatty PT, hopefully because it has all the information that it needs. So let's think what is the name of the company you are applying to? And then we'll just make this voice. So that's not it. Oops. I am applying to Google as a senior UX designer in New York City. What are the top three accomplishments or projects in your career that highlight your UX design skills and experience, okay. Okay. Just making this up by the way, this is not real, it's fictional. I'm doing my best here. So I have over ten years of experience and UX design and product design in e-commerce, Fintech, and startups. I'm currently a product designer at Bard bank, working in their digital transformation team, where I work with developers and key stakeholders, I helped barred bank and its major digital transformation, including working between the front-end and back-end developers. And the app is one of the top rated banking apps in the App Store. And I also worked in a startup called potato bus as a product designer where I helped design the first enterprise dashboard for that company, which involved interviewing clients to understand their needs and then turn those insights into simple interfaces that would work for all clients. And I also received an award for a UX design boot camp among 20 other candidates at flat iron school. What specific skills and software proficiency do you possess? And that would make you particularly valuable for this position. Okay? Alright. Love that. Bear with me. Okay. I'm just gonna talk through this because I want to show you that it can be conversational. I mean, my wording is not great, right? This is what I said. Doesn't even have there's no periods, there's no commas. You know, it's not perfect. I'm just talking. Just recording it. You could record somewhere else if you want to just talk. You could record in an app like Otter, which is great. And then you could copy and paste it in. That'll be fine too. Okay, so let's try this one. I work in Figma Sketch and the entire Adobe Suite. And I'm familiar with Python as I've worked closely with developers to deploy a lot of solutions. In my previous experience, I'm very good at working across disciplines. I often act as a link between product managers, developers, and clients, which requires understanding all angles. And then getting to the heart of the issue. I also excel at collaborating in fast-paced teams and work best in a dynamic environment. What do you find most appealing about working at Google and the senior UX role in particular. Alright, let's see what we can come up with for that. Google is a culture of doers taking responsibility, sharing insights and knowledge. And that way of working really fits me perfectly. Every project I've ever worked in has offered opportunities for me to pick up new skills and share what I know with my colleagues. People who aren't really familiar with design or the UX process, they're usually very curious to learn and I love sharing what I know in terms of skills. Since I've worked across industries through my ten years of experience, I'm able to see things intuitively that many other designers miss. It's an intuition that I've built up and I think that makes me a great fit for Google. Is there a specific department at Google you would like to address in your cover letter to alright, let's now move this over to text. We've had the indulgence. Let's now get back to the business of just clarifying a few things. The design department, I guess. I don't know. Okay. Here we go. We didn't have to answer ten questions, just a few I tried to answer more than was asked for knowing that those details would need to be provided. I'm excited to submit my application for the senior UX designer position in New York City. As an experienced professional with over ten years of UX design and product design across e-commerce, Fintech, and startup industries. I'm confident that my diverse skill set and proven track record of success will make me an invaluable addition to the Google team who not bad, pretty punchy, right? Butt. We could try to make it more punchy. We could say, make it more concise and direct. We can say, add some humor. We could say, make the syntax more dynamic. So this is what I would like to try to get across. After you get this, you don't have to take the first one. In fact, as I've said, regenerate, try again, iterate. So first maybe I want to see what a second version would look like. Okay, so maybe it's going to lay this out a little bit differently. It may give me is that actually the address of I have no idea. Dear Google Design Department or sorry, dear design department at Google. I'm excited to submit my application for the senior UX designer position at Google's New York City office with over ten years of experience in UX design product design and a proven track record of working across industries such as e-commerce, Fintech, and startup. I'm confident that my skills and passionate make me a perfect fit for this role. Yeah. So maybe that's, I don't know if it's a better variation yet. I mean, we want to look at it. We want to sort of play with it a little bit. We might take one paragraph from 1.1 from the other, but at least it's including a lot of the information we need. Now, the iterative part is where I use verbs to give specific feedback. Maybe I don't want barred bank mentioned. Maybe for whatever reason, I think it looks good here, but whatever it is, use a verb and then state what you want to happen or what you want to be different in the next iteration. Remove barred bank, and that's it, right? Make the syntax more dynamic. So you want to use verbs like remove, add, make, use, adjust, change, improve. This kind of verb followed by what you want to happen to then get different versions. And if you want specifically just one paragraph than state that you want that paragraph changed and everything else left the same. So maybe I say remove barred bank and make the syntax more dynamic. Alright, and let's see what we can get with that. In my current role as product designer, I have been instrumental in driving digital transformation, collaborating closely with both front-end and back-end developers, as well as key stakeholders. This collaboration has led to the creation of a top-rated app, showcasing the strength of my design capabilities. That's pretty good. But maybe I want to be very specific because I don't see enough short sentences. Use more short sentences. Use, use an analogy And maybe I want to say make the first paragraph more punchy. And I'm not sure if chat GPT knows what I mean when I say punchy, but I know what I mean when I say punchy. So try different words that represent what you think you mean. An experiment. Again, it's very iterative. So make it more punchy. And maybe I want to say use a, and this is now getting into the territory of maybe a little too much feedback, but we're just playing around here. Use a rhetorical question as a hook at the start. Okay, So this is the specific feedback. Now, the way I like to think of this as is, the way I like to think of this is I like to think of this as I'm giving feedback to a person I hired to do this for me. I'm too busy to write my own cover letter. I've sent you a quick voice message letting you know generally what I want to focus on, you write it. Now. I'm going to give direct feedback to hopefully make it better. I want a couple more drafts from you. Well, the good news is you're doing that in a much faster iterative way than working with a person on it, right? So you can go through these cycles much more quickly and get to something which you like. Now for a cover letter, it's much more subjective what good means? Because for a cover letter, you're trying to represent yourself. So all of these cover letters have pretty good writing, right? And you don't want to get too crazy and creative with a cover letter to a potential employer. You want to be professional, you want to be generally respectful. You want to catch their attention so that they might look at your CV or resume. But the idea is that it has to represent you. So now you have to be the curator deciding if it fits your criteria for what good is, if that makes sense. So that's always true. It's true for the article, right? But we can clearly say for our first approach, just asking chat CBT to write an article about the history of the English language turns out to be very generic and everyone can agree on that. It's not very good. But when we do it the other way, when we use the knobs and we use the outline, It's obviously much better. So yeah, it is subjective, but that is clearly way, way better. For this. It's more of you deciding what better is. So I think that's an important distinction for the iterative approach. The final result is ultimately what satisfies that little thing in your brain. That was your spark to decide to do it. That little vision you have sitting in there and when you see it, you say, Ah, it's like carving the statue of David. You're kind of just chipping away things that shouldn't be there so that you can get at this thing that should be there. That feels right to you. You are the editor in chief. So I don t think this is going to be necessarily better, but let's see what happens if we ask for a rhetorical question at the start. What does it take to create truly exceptional user experiences? As a senior UX designer with a decade of experience across e-commerce, Fintech, and startups. I believe I have the answer, whoa, that's very confident. I'm excited to bring my passion, skills, and expertise to Google's New York City office. Maybe changed that to Google in New York City or to Google periods stop right there. I think that might be what I would do. And then in the feedback, I would say remove the mention of New York City, right? And then because I don't want that there, I want to remove it like a conductor leading a symphony. Okay, That's the analogy I asked for. I harmonized diverse teams to create stunning user experiences. Now, maybe you know what the analogy is. So there you might say add an analogy of a and then choose an analogy that makes sense to you that you already know the conductor. I'm not crazy about it because that feels like the leader of everybody. So maybe I have another one. The point is here you want to iterate until you get to where you feel like it matches your vision. That's the idea. So start with my example if you want, and then try it yourself. Answer this with your own information. Imagine you're applying to a job at Google for x position, change the position perhaps, or a job somewhere else, pick a fictional one, answer the questions that chat GPT asks you. And once it has the answers, once it has all the information, read through the first version, what stuff sticks out to you and you owe. That doesn't seem right, that doesn't feel right. I need it to be more whom, what? Then? Use statements that start with verbs to craft it, to iterate it until it feels right, until it feels like, hey, I could have written this. This sounds like me. I like that. Then you don't have that blank page syndrome of where do I even begin? And I have to write 500 of these for all 500 jobs that I'm applying to. Let's be honest, we're not doing that. But we could make each one unique if we're able to hone it and sculpted in this way, e.g. change Google to another company and make it a little less formal or something like that. You could do a variation of that. Again, does it fit you? Does it feel right? I'm not telling you to do that. I'm just saying that's something that you could do and we're going to be looking at more examples. Next. Now that we have these three main writing approaches, we've got our Star Trek computer, we've got our outlines and knobs. We've got our templates that we can build from. Now that we have these approaches, can we start thinking about, what about social media posts? What about a presentation outline or maybe a lesson plan or an outline for a video. What about interview questions? What about organizing a meeting agenda? What about SEO optimized writing for a short post, a social media post? What about a business report? What about writing a story for my kids and interactive story? What about performance reviews at work which take forever, right? What about all of that? We're going to explore these specific cases and situations and more. Using these three broad approaches, we're going to try to build up an intuition for which approach applies best to which situation. And there's not always one answer. But you'll develop an intuition for yourself that will allow you to approach anything and be thinking about the best way to do it, to find the best way to do it, to get the best results, to get the best quality of writing. So let's get into our specific prompts using the approaches we've talked about four different writing applications 6. Building a Resume or CV: Now that we've gone through our three main approaches, let's build some intuition. What do I mean by building intuition? To get a sense for which approach may be best in a given situation. And you have to decide that for yourself. Now it's also important to keep in mind that these approaches can be mixed and matched. E.g. if you are writing a resume, you may want to write that resume with the outlines and knobs approach by clearly stating the tone that you want, the format, any specific style requirements that you have. But then ask that to be a back-and-forth conversation in the Star Trek computer style. And that is totally doable. You have to, again, build your own bespoke approach for each thing that you need to do. The nice thing is once you have those, you can often reuse them with very minor changes each time in the specific situation. Alright, so let's now go through sort of rapid fire, a bunch of these, we're not going to spend a lot of time on each. We're just going to take a quick look just to sort of think through these. And I want you to think along with me. I want you to think about how you might approach each one. And I'm gonna give you an example, an example prompt for each as well. Okay, So we're going to start with, we're going to start with a resume. So how would you, how would you write a resume or CV? Well, if you don't really care exactly how it looks in terms of the format and you want to just go with whatever chat GPT knows to be the format for a resume or CV. Great. But maybe we want to try it the other way. I would be tempted to use the outlines and knobs approach because that allows a little bit more fine tuning of the voice that I want to have, right? Of how I want it to feel when you're actually reading through my work history. So let, let's try it that way. Let's In an outline of some key information. So act as an experienced resume writer and career expert, write a clear and well-organized resume based on the guidelines and outline that I provided. Right? We've done this before. Now. Maybe I do this way because I want to specify that I really don't want to have bullet points and I want to use very active language, maybe that's very important. So I'm going to try this approach. You can try the Star Trek computer style to, I think it would work perhaps just as well, use precise yet ordinary vocabulary. So I want to avoid using a lot of very formal, flowery Shakespearian language. Perhaps do not use flowery language. And I want to persuade and inform. That is the overall purpose that I have with my resume. Persuade potential employers to hire me, right? So with this one, we should get a pretty clearly organized and a resume and a CV of course, are different. If you wanted to use the template style, you could find a template that you like. There are many CV resume templates and then you would go that way. I want to follow this specific template. Maybe it's better to use the building from templates approach. Okay, so the summary. And again, maybe another reason I choose this approach is because I want to have that summary at the beginning, which hopefully will convince the person who's reading it to continue reading and checkout my education and work experience. Highly skilled software developer with a strong background in computer science, web development, and data-driven applications, proven ability to collaborate effectively. So we have here the language that starts with e.g. proven ability. It starts more active. It doesn't use a subject at the beginning of the sentence like I. That can be a good way to sort of get a person's attention at the beginning. In this case, it works for a resume, works for a CV. It's not often the way that we write, but for this type of thing, it's very common. But if I have an issue with this, then I may want to make some revisions. So I would then do the Star Trek computer style, where I would say maybe make the summary more concise and move the education section after the work history section, work experience section. Okay, so I want a variation of that. So now we're in Star Trek computer right after our initial one. So here you can see how we're sort of mixing and matching skilled software developer with expertise in web development, data-driven applications, and effective teamwork, passionate about mentoring and community involvement. So that is more concise, absolutely. Work history first, and then we can assume that education is after. So I think this is a good case, this specific one, a good case where I think any approach could work. And it would really be up to you to try each one. See which one gives you the best results. And don't be afraid to blend as needed using the iterative approach by requesting things after you get the first version or after you get the second version, make the summary more concise and move the education section after the work experience section. There you go, Right? Very clear. So let's, let's look at another example. 7. Creating a Social Media Post: What if I want to craft a social media post? Maybe I have a specific style on social media. If I have a very specific style of posts and we're gonna look at an example very soon for YouTube descriptions for how we might do that. If we have a very specific style, probably building from templates would be best. I use a previous posts as a template and then I change the meaning, right? I changed the content, okay? But maybe not, maybe I like to be more flexible. Alright, so in that case, it might be better to use the Star Trek computer approach. Again, I'm not saying that it has to be the way to do it, but I think it's a good way. And for this one, I want to probably limit the number of questions I get. I don't want to have to answer a bunch of questions and give tons of information that would be pretty exhausting, wouldn't it? So please act as a social media expert and helped me craft a social media post to share insights about using chat GPT, using the Star Trek computer approach 0, very meta. Before writing the post, ask me specific questions one at a time. Only ask one question at a time and wait for my answer. Don't ask more than five questions once you have all the information, draft the post, okay. And I guess technically there should be a comma there, right? Okay, alright, so let's, let's answer, let's answer the questions. And I'm going to speak what specific features or applications of chat GBT would you like to highlight using the Star Trek computer approach? It's a heck of a question. Click on the wrong thing. I would like to focus on the Star Trek computer approach being highly iterative and taking full advantage of the nature of large language models. Which platforms will you be sharing this on? Linkedin and Twitter? I could have typed that for sure. Probably faster. What tone would you like for this post-professional casual or a mix of both? I would like it to be very playful and use a very clever style. I want to use a question to encourage comments. Are they general users of AI or people with specific interests in chat GPT. This post is for working professionals who want to streamline their written communication without sacrificing quality. Thank you for the information I will not draft a playful and clever, so don't announce a chat GPT, that's a little silly. Linkedin. Captain. We've discovered the secret to efficient, high-quality written communication. It's the Star Trek computer approach to chat GPT. Alright, so now I might say it's really jumping on the, the Star Trek theme with the live long and prosper spock hand gesture. So maybe I would say I'll make it shorter, make it more concise. I'm curious if I say make it less cheesy and more concise. I'm not even reading the other post, but and more concise. I didn't even read the other one discovered the secret to efficient high-quality written communication with the Star Trek computer approach to check GPT, okay, for LinkedIn, unleash the power of chat GPT with the Star Trek computer approach, engage in iterative conversations with your AI to refine your written communication quickly and effectively. Working professionals. How could this approach impact your daily tasks? Share your thoughts in the comments. Not bad, actually, not bad. What I would remove here is it has asked me for the audience, but if that is the audience I'm sort of aiming at, then I would actually just delete that, right. The working professionals. If in my social posts, I don't want to say hello, working professionals. I think that's a little cheesy. I could just say, Don't mention the audience or don't include the audience in this one. Or I could say simply or I can just delete that and say, how could this approach impact your daily tasks? Assuming that my audience are the ones reading the post, right? This may be then the benefit of the template approach. Where if I have a standard style for social posts, then I can ask GPT, chat GBT to analyze that style. And once it understands that, then simply say, give me a post about the Star Trek computer approach using chat you bt. And it'll fit what it already knows as my style because I have it in a template, because I'm using a template, right? So keep that in mind and consider the one that fits your use case best. I think this is pretty solid and I would actually post something like this. Minus of course the working professionals thing, Twitter embrace the Star Trek computer approach to chat GPT for swift and effective written communication. I would there remove the exclamation marks and that would be something that it could learn in a, in a template that I don't like to use exclamation marks very often. So it would have to learn that. Or I could, I could do iterative feedback where I would say, make it less excited using no X. Goal. Mation marks. Don't address the audience. You discover the power of techy bt using the Star Trek computer approach, engaged blah-blah-blah. Strategy. You do the test for working professionals. How might this strategy influenced daily tasks for working professionals? Yeah, I suppose that could work as a question. I think that's a little bit better. But I still think when you're asking your audience, you would say, how might this strategy influence your daily tasks? That might be a little bit better. Tap into the Star Trek computer approach to chat GPT for effective written communication. Full stop, no exclamation. Great. So I'm getting to something better and it's I think, a pretty fun process, the back-and-forth. So try this out for your next social posts. I'm not saying use this specific approach. Try out the template one as well and see which one gives you better results. Again, using the outlines and knobs approach for social posts, maybe not the best idea. I think, as we've talked about, that's generally better for longer form content. 8. Generating Content for a Presentation: What if you need content for presentation, for a video, for a lecture, for something that you have to talk through. And you need the content for that. And you want to build it out in a specific format. I use this sort of thing all the time because I make a lot of YouTube videos, I do a lot of live streams. So often I know what I want to talk about. I have the specifics, but it's nice to have clearly written, well-organized notes or an outline or content that I can read through as I progress through, let's say e.g. the live podcast. So for that, I want to follow a similar structure because if I don't, then I have to learn a new style each time, right? So for this, I like to use personally, personally, I like to use the template approach. Building from templates is a great way to get the same kind of thing each time with different content. Now this will have different degrees. If you're very clear on what content you want and you just need an organized, then you need to provide more after you get I understand or after you get confirmation that chat GPT has learned the content. If you want to get something around a specific topic and you don't know specifically what it is, then you might leave that up to chat GPT as well. So there is some flexibility here. It's not totally fixed. And I'll show you what I mean. Okay, so let's say e.g. I. Have a specific format for teaching idioms to English learners. Alright, so learn the structure, style, and format of the following presentation. And once you have learned at say, I understand, do not add additional information. The example is meant to be used as a template for future presentations. I will then provide the details for the next presentation I want you to write. Here is the example. And then I have idioms for making decisions. And then I have three idioms on the fence, having second thoughts. And in over my head, three idioms that English learners may not know that they may need to learn about this general topic. But then I have an example under each one, I'm on the fence about whether to accept the promotion. Yes, it means more money, but also more responsibilities. Now, this is where it gets very interesting. Because based on what I say Next, let's just make sure we get I understand. Okay. We've got I understand. Okay. So why is this interesting? Because I may now provide the next three idioms I want to talk about. And I know exactly what they are. But I don't want to make the examples myself. But because I've provided the template, I don't have to say, give me an example under each one. That's the beauty of the template. It now knows that it needs to make an example under each one. But I don't want to think of the example. I just want an example that's going to work to explain the point to the learners, right? So that's pretty cool. But maybe I could go a step further if I'm a lazy teacher and I would never do something like this. Of course, I usually like to decide what the content is. And then if there's something that I can outsource, that's less critical but still needs to be there. I might outsource that to the template to chat GPT. So in this case, I can do it either with just the example sentence underneath in over my head, supplying each of the three idioms. Or I can give a general theme and asked for all three idioms, not knowing what they're going to be. And then do the Star Trek computer approach. If I don't like the ones that I get, right, or I could give the specific idioms and the example sentences. And it will just structure it for me. Now in this case, because it's so simple, It's a simple example that would be kind of pointless because it would almost be easier to just do it myself then, right? But if you have a bunch of notes that you've taken or voice dictation rough notes, and you want to format those into a specific structure, then it becomes a great formatting tool to organize things. Pretty soon, we're going to be looking at an example where we have a bunch of notes and we want to put it into a specific structure so that people can easily read it and review it. So there's a lot of nuance here, right? It's very flexible. Okay, now we put in the next topic, Idioms for complaining. But I want, because I'm a lazy teacher in this specific case, I want to ask for the idioms as well. I think a responsible teacher would come up with this specific idioms and then ask it to be organized in a specific way. But in this case, super lazy, no time, very busy, can't do it. So here's what, here's what we do. And Jack GPT complies. But I think I would have to add this, create three idioms with examples, and then it will fit those within the template that we have provided. Barking up the wrong tree. I think that's a pretty good one. Making a mountain out of a mole hill. I think that's a good one for complaining. Squeaky wheel gets the grease, that's a great one. I know it's annoying, but sometimes you have to be a squeaky wheel to get the grease. Barking up the wrong tree might not be the best. So here's where I might give a little feedback. Maybe change the first idiom or swap the first idiom out for another one. But I think you get the idea. Now apply this to more complex presentations. I wanted to use a very simple example because I wanted to highlight the different layers. Layer of number one. I may want to generate everything because I have a general topic but no specific content. Number two, I have specific content and I want some of it to be generated, but I have the main points. In this case, that would be the idioms but not the examples. And then another level where I have basically everything in an outline or in bullet points at least roughly. And I just need it to be organized. All of those are great for using the template approach. 9. Coming Up with Great Questions for an Interview: The next one we're going to look at could have different applications for very different types of situations. So we're gonna go through it, but as we do, I want you to be thinking on your own. Okay? What would I actually need to use this for in my life, for what I need to do, right? We can't cover everything, but there are so many different places where you need questions. You might need them for an interview. You're interviewing someone, or perhaps you're preparing for an interview. And you want to anticipate some of the questions. Well, you would create a prompt that's different for each of those two cases, but both are in a professional setting. Okay. But what about a book club? Well, that would be very different because then the questions might be more open, more why and how questions or what do you think questions, rather than telling me about type questions or prompts. And so it varies widely again, depending on what you need questions for or what you need a list of prompts for. It may not be questions, it may be prompts. So if you're a teacher in a school and you need 30 prompts that are each different for a classroom situation. Well, that would be one thing and you would need to really sculpt your own bespoke prompt for that. Again, if you're preparing for an interview, that would be very different. I'm going to prepare some interview questions for conducting an interview. But I would also use this for a book club. Sometimes I run book clubs and in that setting, I want to choose some open questions that are going to get the conversation going for the book clubs situation, I might do more of a hybrid between the Star Trek computer and the outlines and knobs approach. Because I might want to provide a basic outline summary of reading for that week. I'm not going to assume that chat GPT can immediately pull that specific reading, right? So I might create an outline and provide that outline as the raw material, right? And then use the iterative method to get to the questions that I need to run the discussion. This gets into a lot of those nuances and you have to think through which approaches or combinations of approaches will work best. So this is how I'm going to approach this one for the job interview situation. I'm going to I'm going to do the standard thing where I asked for I asked for chat GPT to become an expert in career in hiring. I want to specify that I need ten questions, right? I haven't said anything about the job. Now I could put that in the prompt, but maybe I want to save that for the back-and-forth. Either way, I think is fine. So that would make this longer and the back-and-forth potentially shorter. But I do want to make a slight hybrid here because I want to kind of bring in the template style where I show it a type of questions so that I get more like that. So this is more of a hybrid between the Star Trek computer and building from templates. Ask me questions. I've limited it to three questions because I don't want to have this very long back-and-forth every time, maybe two or three questions, that's okay, by the way. And then once you have all the information creates the questions. Here is a sample question. What is something important you've learned thus far in your career? So maybe I want to focus on a particular type of question more about you. Questions about your values questions, how you work questions rather than specific skill questions, right? So let's see what we get with this. Well, we're gonna get a question, right? What is the specific job position for which the candidate is being interviewed? Well, that's a very roundabout way to ask the question. This is a marketing position focusing on social media marketing activities, including strategy and creating content. What level of experience is required for this position, entry-level, mid-level, or senior level. This is a mid-level position requesting at least three years of experience in social media marketing. What is the industry or sector in which the company operates? This is a pet food brand with customers in Europe and North America. Now we should get questions now because I said no more than three. Based on the information provided here are ten interview questions. So let's check out the questions that we get. Can you share an example of a successful social media campaign that you have managed in the past. What made it successful and how did you measure its performance? I liked that second part of that. I want to know how you knew it was successful. That's an interesting one. How do you approach creating engaging, sharable content for a, or how, well, how do you approach for pet food brand? So I would make a change there to wood Assuming that the candidate hasn't worked in pet food before. So I'm asking a hypothetical, but okay. Not bad considering our customers are located in Europe and North America, how would you ensure that your social media strategies cater to the cultural differences in preferences of these markets. I think that's a good question. Now, maybe because we're in the Star Trek computer approach, I want to iterate this. Maybe I just do a straight regenerate, but perhaps I want to focus not on a specific strategy things, but focus the questions Around values, working style, and personality. Okay, I want to, you could maybe do this twice. Then the next time, focus the questions around career achievements and social media strategy questions, right? But here I want to learn about this person. I don't necessarily want to just know what they know. I don't want to necessarily just access their strategic mind. I want to know if this is a good cultural fit for the company. How do you personal values align with our pet food brands, mission and vision. Okay, maybe that's, I mean, that's a question that an interviewer would ask. I wouldn't ask it personally, but I probably wouldn't use all ten questions either. It's not exactly my style. But that is something that somebody would ask because I don't think it's fair to us to make people say in an interview, why do you want to? I didn't even like that question. Why do you want to work here? Most people are going to answer, are going to think, I want money and then answer some, some answer that they think the interviewer wants to hear because it's not a fair question. What do you believe are the most important characteristics of a successful social media marketer? Yeah. And how do you embody these characteristics in your work? How do you handle stress in tight deadlines and if fast-paced a marketing environment, That's a great question. How do you handle constructive criticism and feedback on your work? That's great. Collaboration is crucial in a marketing team. Can you describe a situation in which you worked with a diverse group of team members to achieve a shared goal. Again, great question. I personally would not choose the first, But then again, I would probably go through a few versions of this specific feedback and then pick no more than six to eight questions from, let's say 20 questions that I that I get, right? So part of this process, when we're in the Star Trek computer approaches, we expect higher-quality from the outlines and knobs approach because we are highly precise in what we want. Tweaking those knobs. In the building from templates approach, we expect a match with the format we have provided with the style and the voice that we have provided, right? And we want to get something that's pretty good, but this sort of repetitive idea, we may need to write 30 emails responding to customers. And this is a good, a good way to do that with a few changes here and there. This approach, because it's more iterative, we're giving less at the start. And as a result, there's more refinement required on the backend, which is once you get the response to, start giving feedback, to hone it and sharpen it. Bring your mind back to that example that I gave from that episode of Star Trek. Okay, you get the table. It's not quite like that. It's at an angle. There was a some scissors or a sharp object coming out of one side. So that's the idea, right? We don't want to just take wholesale copy and paste. Those are good enough with the first response we want to iterate through. And then cure rate. Cure rate have a curators mindset picking out the things that are useful and providing feedback to get things that are going to be helpful for what we want to accomplish. Okay, So I like eight of ten of these, but I'm only going to use six. And but I also need specific career expertise related questions for this interview. So I'm going to keep going and I'm going to ask for more career-oriented or skills oriented questions. And then from all of those, I'm going to create a list of 15 to 20 questions that I may use in the interview. Great, but that's from a larger list of maybe 40 or 60 questions that I have generated, right? So that's the idea, That's how you want to use this approach. Think of it differently from the other two, I think. Alright, so how could you use this for something that you need to do in your life? I gave two possible examples, the book club one, I'm not going to show you that one, but I would I would approach it a little bit differently. I would still provide a sample question for a book club, but I would also, as I mentioned, provide a summary of the reading as well so that we're basing the questions on something that's relevant to everyone in this book club, meet up, right? Maybe you're preparing for an interview. Can you come up with questions that you might be asked in a very similar way. Maybe you need to teach a lesson. Can you come up with prompts that are going to challenge your students to write an interesting essay. Again, there are so many different applications. Think through the ones that you would need. Think through the situations that apply to you, and then customize, create a prompt like this one you can start with mine if you like. That is going to be applicable to your life, right? And if you have any questions, let me know. Let's keep going though. We have a few more of these to cover 10. Integrating SEO Keywords Into a Blog Post: When we write a blog post, usually we want to make sure that it's SEO optimized. Now that is a whole big process and this is not an SEO course. But one of those steps might be to make sure that we've integrated or included a few important keywords into that blog post. This is, besides everything else that you might need to do. Now, maybe you've done research elsewhere to decide what keywords you want to include in that specific blog post. Maybe you're not sure yet. So this is again where you need to think through which approach is best or which combination of approaches is best. I almost said our best, but that should be is best. Which combination? That's the thing, right? Right, right. So here it depends. Maybe I know the keywords there. I might just say act as maybe an expert blogger in this field. And then use my first approach, the outlines and the knobs including those keywords and draft or create a first draft of the article from an outline or from a first draft that way. But I might not have the specific keywords. Might want to make this more of a back-and-forth process. Well then maybe the back-and-forth iterative Star Trek computer approach would work best there for your first draft. Then you would bring that first draft into our first approach, into the outlines and knobs approach. But maybe I have a list that maybe comes from a friend or an expert in SEO marketing of best practices for SEO. Things you have to do to make sure your article is fully optimized. And I first want to show chat GPT, that list of best practices, have it learned that, and then go into the Star Trek computer approach. So you see how these work together. The best way to, and we're going to look at a specific example of this very soon. The best way to make sure that you're following best practices is not to hope that chat GPT knows the best practices, but it's to say learn these best, best practices and then go from there. And that's a very different way of doing it. But you get much more precise results that way that are directly put up against the standard or the guidelines or the framework that you first provided and taught to chat GPT. Now if that sounds like a bit much, don't worry. We're going to look at a specific example of that in a moment. But maybe I want to, maybe I want to do this in a slightly more playful back-and-forth, back-and-forth way. So I might say act as an SEO expert. I will provide you with a short blog post. After reading it, I want you to ask me questions to help me optimize the keywords and discoverability. Use your knowledge of SEO to help me generate an optimized version of my post only asked me one question at a time, then wait for my reply. Do not ask me more than two questions. I want to keep it limited. Maybe this is an exercise. It's meant to be exploratory, playful. This is not meant to be my final draft of this article. This is not my serious draft. I've just done a rough one or have generated a rough one. And I want to make sure I include some specific keywords. Here's my post, then we have this post. Let's not worry about the content of the post. It's about, it's about YouTube shorts. It's like a commentary on the trend of of YouTube shorts. Okay, so what are we getting from this? Well, in terms of an SEO expert, it's not super clear. So if that really matters to you, consider doing a step before this, where you bring in a framework. You bring in guidelines using the template approach, teaching it that, and then doing this approach or our outlines and knobs approach, right? Or you want to play around, you're going to hope that expert means expert. That chat GPT actually knows some best practices. Again, you get a lot more variation there. It's not a certain thing. There's no standard that it's being compared against. What are the main keywords or phrases that you would like to target for this blog post. Now, here I could say, well, I want to make sure I'm using, Let's see, content creator or something like that. But maybe I don't know yet. Well, then I'm not a very responsible, I suppose, social media marketer or whatever I am in this case. But perhaps I don't know. Why don't we always hit that one. I'm not sure yet. I was hoping that you could help me develop some keywords to use. Based on the content of your blog post. Here are some potential keywords and phrases to target. Youtube shorts, video sharing, short-form videos, mobile optimized videos, content creators, daily news. Yeah, okay, So that gives me a better idea. And again, Not necessarily best practice because I would need to do some additional research to decide if these are the best keywords there are, of course, many places and ways to do that externally. I want to focus on content creators, short-form videos, and YouTube shorts. Those are the keywords I want to include. Great, to optimize your post. Youtube shorts, content creators and short-form videos consider incorporating them more frequently and strategically throughout your text. Make sure they are included in your title headings and Meta descriptions as well. While it would be a tough job to make a title that includes all of those. Alright, also ensure that your content remains natural and engaging while doing so, okay? Yeah. Okay. Thank you for that. Can you please help me write my first draft including these keywords and best practices? Of course, here's a revised version of your blog post with the chosen keywords and SEO best practices title. Youtube shorts, the rise of short-form videos and content creators. Not as bad as I thought, not as hard as I thought. I don't know, I might want to iterate that one a little bit. Youtube shorts has taken the video sharing world by storm, becoming one of the hottest video trends in 2021, right? Yes. Well that's a little outdated. I suppose. One of the key reasons for this success is that YouTube shorts is incredibly user-friendly. The creation tools are simple with features like speed control, music, and filters. In conclusion, okay, so what we've got here is I would say my opinion pretty generic, right? I mean, it's okay as a starting point. But then I would go on and either continue iterating, specifying style, tone, voice, format, all of those things that we've been talking about as actions. Remember we use verbs to follow up, make it more, make it less, Add, Remove, right? Don't mention, mention. Include those sorts of verbs. We want to state what we want. Or we would take this as a first draft and then plug it into our outlines and knobs approach, ensuring that in the knobs we include keywords or keywords to include. And then those keywords that we want to make sure are in the post. Again, this is just one approach. We've talked about, a few different ones that could be used in this situation. Also be thinking about other ways that you could be using this style where you have an early version of something, you get to an early version of something that has something important in that back-and-forth that can then be used elsewhere with the other approaches, either of the two approaches, honestly, and I'm really looking forward to showing you how we can use a framework or guidelines which is coming up soon. 11. Crafting a Structured Meeting Agenda: If you were going to create a well-organized meeting agenda, which approach would you use or which combination of approaches would you use? Well, it depends. Maybe you have a particular format that the whole company uses or your team uses every single time. In which case you may want to use the template approach or maybe mix the template approach with the Star Trek computer approach. Providing the specific details of the meeting. Great. But maybe you don't really care if it follows a particular format. It just needs to be simple and well-organized, easy to follow, alright? In that case, maybe we could use the outlines and knobs approach as I hope you're starting to feel. We could use any of them and they each have their pros and cons. This is all part of this process of building your intuition around these approaches. And by the way, I'm just calling them these approaches because that's what makes sense to me. I don't care. You can make up your own names or no names at all. The important thing is that you start building that intuition that allows you to easily apply the best approach for your particular situation so that you can build out your library of prompts that allows you to work more effectively and efficiently. That's what I really care about. Not really if you call it the Star Trek computer approach, I don't care. Right? It's just what makes sense to me. Alright, so let's, let's try using the outlines and knobs approach to create our agenda. Act as a professional Assistant with an expertise in creating effective meeting agendas. This first sentence, notice I always structure it the same way. And the reason I do that is because the large language model wants to be everywhere, wants to be everything. And so you have to use your first, well you don't have to, but it really helps to use your first statement as a thing which pulls it down into a specific situation, using it as a localizer. Hey, just for a second, splash of cold water in the face. Don't be everything for just a second. Pay attention, focus. Be this just for a few minutes, okay, while we do this thing, then you can be everything again. You can, you can generalize again, but I need you to focus and just be one thing for just a minute, right? That's what the first statement does. Then you go more and more specific. So you're kind of building this funnel of focus that then gives you specifically what you're looking for. And then you release your little genie of localness back to the Cloud of being very general and multipurpose, right? I think that's a good way of thinking about it. That sounded crazy coming out of my mouth, but that is really how I think about it. Okay. Crafted, concise meeting agenda based on details I will provide ensure the agenda is well structured and simple. My main criteria, right? But then I need to create my, my, my details and my knobs. So Here are important details that I want to provide, right? But then I have knobs here. I have an objective not knob. I have a tone knob, a style knob, and a format knob, but it's not so specific that I need to use a template, include the duration of each section rather than the time. So I don't want to say it's 2-45, this section. I want to say it's 45 min right? Now, please create the meeting agenda using these guidelines and details. Okay? So what we should get then is something that is clear enough but doesn't have to follow a particular format. Objective, review sales performance, discuss future sales strategies and explore collaboration opportunities. Okay, so here I have my 20 min, 30 min rather than my specific times. That's what I want is it's simple. It's using bullet points, that's what I want. But it is it is committing a an error, not an error on a fence, a grave offense. What I want is no Roman numerals have I specified that? Do not use Roman numerals. So that style is an issue. Well, let's see if we try that again. If we don't get roman numerals because that's a big no-no for me. I don't want Roman numerals. Now, if you get something that you don't want, okay, So we didn't get roman numerals the second time. One way is to just try regenerating and see if you get something that is closer to what you were expecting. This is definitely closer. I said I don't want roman numerals. I got Roman numerals. So unfortunate, but there are things you can do there. Number one, you could add a knob that is temperature. And the closer to zero you go on temperature, the less ambitious chat CBT is going to be. So you might set a temperature and then move that to 0.5 or something like that, then you're going to get more strict adherence to your criteria, to the knobs that you created, right? Or if it's just one thing that bothers you. I don't like the Roman numerals, then you just say no Roman numerals. And that should solve it, right? If you have one specific thing. So then you kind of go, go into the iterative style of feedback. That's hopefully just one or two iterations before, before getting something that you're happy with, right? So I think this is a pretty good agenda. I like the numbers, I like the simple bullet points. I liked that I have the minutes for each section. And that looks pretty good. So I would go with that. Now, if I found that it didn't have enough detail, maybe I want to include names. Well, you'd have to then supply that above. So in my in my details here, I have The objective, I have the, I have the participants broadly, but I don't have the names of people, so maybe I want to include the names of people. These are all things that you would provide above any details that you want to be included. So that works pretty well, but I would encourage you to try the other way, try using it with a specific template for the next meeting you have. See if you can set up an agenda that fits your criteria, that works and see how it goes. And if you have any questions, let me know 12. Interactive RPG Storytelling!: Let's get creative for a second. What if you wanted to create a story featuring your child as the main character? That could be a bedtime story that you tell them better yet, what if that story could be interactive? It could be responsive to things. They say, two actions they take. That would be pretty cool. Now, this is kind of the direction that I think video games are going to have AI generated characters or to have characters that integrate AI so that they can respond to your actions and what you say in the story. But we can do this inside chat GBT as well. This is a text version, but still very cool and very powerful as a way to create a compelling story. Alright, so here's what we will do. We're going to make a request. I want you to act as a children's storyteller, then helped me write an interactive children's fantasy story that includes my son, demander, as the main character. The story should feature dragons, magic, and involve interacting with Magical Creatures, having conversations with NPC characters, problem-solving challenges that will help a child think creatively and minor battles that aren't too violent. So this is where I state my general requirements. And again, this is where you could also bring in some knobs if you have some specifics and mine around style, narrative style, right? E.g. you might say, style, JRR Tolkien's novel or something like that, right? Butt. Maybe I'm not too particular about that. I don't really care how it's written. I just want to create this cool, this cool story. I did state that it is a children's story, so I'm not expecting it to use really difficult words. I want you to tell the story in short sections. This is the key part to make it interactive. Each time pi manner needs to take an action or converse with another character. Stop and wait for my input. In this little paragraph, what we're doing, and this is kind of crazy. We are building a video game. It's a text video game, but we're kind of building a role-playing video game. Now, what if you have rules in mind there? You would want to include those either as criteria, maybe after this or before, or as knobs. You might add a knob called rules. Rules. Battles are governed by dice throws or something like that. So you have to bring in some dice. If you roll high or more successful with an action, if you roll low, you're not as successful. You might fail that action or tried to persuade an NPC character, you'd roll high or low. Those are the basic Dungeons and Dragons rules. That's sort of how a lot of the video games, role-playing video games work on the backend of the mechanics are based on a rolling system. But anyway, we don't want to get too complicated with this because this is just the interactive mechanic we're trying to create. I will tell you what action was taken. Use that input to adapt and continue the action of the story. Once upon a time in a magical land full of wonder and enchantment, there was a young boy named pi Mender. He lived in a small village on the edge of a dense forest. One day, commander was playing with his wooden sword when he stumbled upon a secret cave. This is pretty good so far as payment or peered inside the cave, he saw a large majestic dragon curled up asleep. The dragon was covered in shimmering scales that reflected the little light that entered the cave. Commander took a step back, unsure of what to do, should demand or tried to sneak past the sleeping dragon, attempt to talk to it, or leave the cave and return to the village. Now we have a choice to make, okay? So there you would ask your child, what do you wanna do? What action do you want to take? I want to try to sneak past. Okay. Okay. Okay. Great pie. Amanda tries to sneak passed the dragon without being detected. Okay. What happens, what happens next? Payment or took a deep breath and tip toed carefully pass the sleeping dragon, making sure not to disturb it. As he move deeper into the cave, he noticed the walls were adorned with intricate carvings and glowing gemstones. He finally reached the end of the cave and discovered a small dusty bookshelf. One book in particular caught by demanders. I, it was an old leather bound book with a golden clasp. Commander decided to pick up the book and read the title, the lost art of Dragon magic. Suddenly the dragon began to stir. It opened its eyes and spotted pie Men are holding the book, the dragon and lifted its head and said, I see you've found the book of Dragon magic. It has been in my care for centuries. What do you intend to do with it? Young one, should demander respond to the dragon. So you have a, you have a question now and you need to figure out how to, how to communicate with a dragon now or hopefully not be burned to a crisp, right? So this goes back-and-forth and can go on indefinitely. Now, you may need to make adjustments to this as you go, as you get further into the story, you would need to paste at least some of the previous story to give chat GPT context. That context is going to help it continue the story if it's not able to do that in one particular chat, this is a current limitation which is soon to be solved, I think. But anyway, the key idea here is you set up the parameters and then you have this video game. And you can use it for a lot of different things. It's just like setting up a role-play for a job interview for all of the situations where you would need to have a back-and-forth, you can use this mechanic. Now, the reason I'm not talking about how to do a job interview this way. The reason I'm not talking about how to maybe have an English practice partner this way is because that's not the focus of the course. This is a story that we are writing, right? This course is focused on writing. So we can sneak in this, this approach, this style, using the Star Trek computer to do something very interesting. But note that this approach can also be used for a lot of other interactive activities. And I will leave that to your imagination. 13. Turning Notes Into a Business Report: We haven't yet tried to use a longer, more complicated template to create something longer. And that's what we're going to try. Now. We're gonna be working on a business report, right? And we need to know why we might choose a template rather than e.g. using outlines and knobs, or perhaps the Star Trek computer approach. Well, again, it depends on what you're trying to achieve. If I went to leave it up to chat GPT to create the format, I can trust it with that. Then by all means, maybe use one of the other two. But I have in mind a specific format that I want to use. Maybe this is after I've found some best practices for how to structure a business report, may be it is from requirements that I have. It has to be like this. Well then we really have no choice. We have to use a template, right? So this is when you don't want to leave the structure, the format, style up to chance. That's why we would do this rather than using one of the other two. So this is going to be very familiar and we're going to start moving faster here because I think hopefully you're starting to get a feel for this, right? An intuition for it. So we're going to start moving pretty fast on these learned the structure and format of the following business report template. Again, first sentence specifies the whole context. Once you have learned at say I understand, do not add additional information. The example is meant to be used as a template for a future business report. I will then provide the details for the report I want you to write. So here are the guidelines I have or the template rather that I have. And notice that each of these points says what kind of thing I want here. So if I have the content, but it's just a bunch of rough notes, That's the ideal situation here. I spent a lot of time just writing down a lot of research in bullet points. A lot of bullet points. I've spent months working on this report, but it's very poorly organized, right? And it's great content, it's great information, but it's all over the place. I'm citing research for this. I'm working on something that's completely unrelated, but those two bullet points are side-by-side. So that's why for this, this template that I'm creating, this example I'm showing, I wanted to specify not just the headings, but I want to specify what kinds of things I want to be in the bullet points, right? So I want to have and analysis of the results under the discussion section, I want to have a comparison to previous research or industry standards. And then if I've done it correctly, my notes will be added to a little bit. It will be sort of fleshed out a bit and then put in the specific place that I want it to be put in. I want this thing to be here basically, right? So if we do it correctly, then it should look right. Everything should be in its right place. So what we want to see here now is I understand. And then let's change this from I understand to boat. Boat. We can't spell boat. Boat shoes. Okay. So we wanna see boat shoes. And we see boat shoes. Okay, so that's encouraging. Now, use the information I provide to write a complete business report based on the template that you have learned. If there is not enough content for a specific section in exclude that section from the report. Now this is where you would think through whether you want it to make up something. Hopefully not. If it's a business report that's very important to get that right. Or whether you would want to just exclude that section for now and then maybe add it later, right? So think this is not a copy and paste situation. None of these are. I want you to think through carefully the situation and the approach that is best for that situation, or the combination of approaches best for that situation. And then very carefully, the specific wording of the prompts that you're going to use within that approach or combination of approaches. That sounds hard, It sounds like hard work, but putting in that work makes a big difference. Use the information I provide to write a complete business report if there is not enough content for a specific section, exclude that section. So don't even put that section there. Then I just have a bunch of notes that I've taken over time, just a bunch of things. And this is actually pretty short. Maybe it would be extremely long and that's okay too, right? So we generate a first version. Now because this is dummy content, it's not going to be as high-quality as if you have actually put to the time and effort into generating great content, right? Into generating great notes from all of the research that you've done, everything that you've Collected, but it is according to the structure that I have asked for, it is putting things in the right place. I have introduction methodology, results. Let's check results, presentation of findings. Let's see. Let's go down to K. We have introduction results, presentation of findings. Sales revenue increased by 15%. Over the past year. Successful product launches in new markets contributed to revenue growth. Now, the introduction, it's written this way, but under results, it is written in bullet points. And maybe for that, I don't want it to be written in bullet points. Maybe I do. Maybe I don't. For results it makes sense to me to have that as bullet points. Right? Because that's usually information that you want to grab quickly, whether it's incomplete sentences or not. For a particular use case, maybe you want to have that written as paragraphs. You would then need to specify that in your second version in the results section, make sure to use complete paragraphs. Again, not sure that's the best idea in this case. Discussion. Acme Corporation success can be attributed to its focus on customer satisfaction, product quality, and delivery speed under discussion. And we're just really checking that it's following our requirements here under discussion analysis of the results comparison to previous research or industry standards. So that seems to be what we're getting. So pretty, I think, pretty deep ability to generalize, too abstract, to synthesized, and then organize things according to a specific framework, template, or guidelines that you provide. But make sure if you're not getting what you want, that you're tuning things in the initial prompt or in the prompt that's requesting the, that's requesting the full version here, right? That this specifies the style that you want to use. Anything that's specific, that you want to be there or not there, right? And yes, you can do this after using the Star Trek computer approach, or you can do it before, or a combination of both. So here I might specify using knobs, like in our outlines and knobs approach. Or I might just make a separate statement here. For the Summary section. Use bullet points. Do not use bullet points, e.g. okay. So there's a lot of fine tuning to be done here. But that fine-tuning done upfront saves you a lot of time in having to sit down in front of a blank page and create this from scratch, right? And then from here we iterate. We add sources, e.g. references. We need a list of references. We haven't included those. So they're not here, right? So it's kinda requesting that because it's putting putting this in brackets saying, Hey, this is missing information. So give this a try. A business report would be one thing, but it certainly could be used for something like resume. It could certainly be used for a lot of different technical things that have a particular format, particular style that we need to use. 14. Writing Performance Reviews Conversationally: Performance reviews, one of the most boring, tedious things that anyone has ever had to do in the history of humanity. Performance reviews, whether it's for one person, for yourself or for a bunch of people being able to do this in, let's say a less boring way would be nice, wouldn't it? Now you could use if you have a template in mind, the template approach. But I think it's better to do as a back-and-forth because then I can talk. I can just sort of explain what I think about each person's performance. And then it will be fit into, let's say, an organized performance review. And there I could bring in a template if I want to. So if I have a particular format or style that I need to use than I can do a hybrid between the Star Trek computer and building from templates, act as a personal assistant and helped me write yearly performance reviews for three people on my team. I want you to ask me questions to get all the information you need and then build the performance reviews in a clear format with precise and professional language. Asked me about each team member one at a time. Only ask one question at a time. I hope you're also getting this. As we go through we see, Yeah, everything that we're doing is kind of one of these three approaches or some combination. But each case is still very unique and you have to be very precise in the wording. Because if you say asked me about one team member at a time, well then you might get a bunch of questions about that one person. Alright, well then I have to basically answer a bunch of questions. Isn't that what I have to do and I do a performance review anyway. So that's not saving me a lot of time. What I want to do is have this nice back-and-forth an interview. So not only do I want to talk about each person one at a time, I want to answer only one question at a time. Because then I can watch a TV show, my headphones on and I can drink a coffee. I don't have to sit down and look at a bunch of questions. So think deeply about what you're trying to get that's going to make your experience doing it better. Because if it's better and it's less painful, well, that's the goal of either automating a task or doing it with an AI, right? So here we go. Great. Let's start with Abby. So starting with Abby, great. Very clear what our Abby's primary responsibilities and key achievements over the past year? I don't know. Let me think about Abby. Abby was responsible for reaching out to pet food retailers across the United States to sign them up for recurring orders of our most popular items. She was able to close over 500 retailers, which is more than any single sales member had closed the year before. That's impressive. What challenges did that we face and how did she overcome them? Not only did Abby half to innovate in her sales approach due to the economic situation. She also had to overcome communication challenges. As many stores wanted to communicate by email rather than getting on phone calls, which required a lot of adaptability. And Abby handled it with grace and professionalism. This Abby is starting to sound like quite a lady. Abby, what a pro. Thank you for the information. Can you provide some feedback on Abby's collaboration and communication skills? Abby is a great mentor to most of the teams since she has a lot of experience, she is proactive in considering other people's needs and what's going on in their lives. And she is an exemplar in the office. Thank you. Now, let's move on to the performance review. Abby Abby's yearly performance review, key responsibilities, key achievements and challenges faced and overcome. So we have a pretty good review here. I think it's got pretty good detail. But maybe I don't like that. It's written with numbered bullet points. So I would say make each section a paragraph, rather than using bullet points. Add more detail to the collaboration section Collaboration and communication section. Okay? So I want it to be more qualitative or I want it to be a little bit more fleshed out, right? I like this better. Abby has been responsible for reaching out to pet food retailers across the United States with the goal of signing them up for recurring orders. Basically what I said. So It's communicating what I said, fleshing it out a little bit and making it more qualitative. Again, maybe the bullet points is best. Maybe I get this the first time. I want it to be bullet points. So then I say make each section bullet points instead of paragraphs. But this is what I want because I want someone to be able to read it and not just see Abby as a person who I want them to read it and qualitatively understand Abby's contribution right? In the team in terms of collaboration and communication, Abby has proven to be an invaluable asset to the team. As a mentor and experienced team member. She is proactive and considering the needs of her colleagues and offering support. Okay. Now I want let's see. Okay. So now that's a version of her performance review, but I still want to maintain the section heading. So I want to say use section headings. In the feedback. I want section headings. Okay, performance review, primary responsibility. And now we have the paragraphs, which is what I want, and we have the challenges facing overcome. So I wanted to bring back those main section headings, but I didn't want bullet points. Now I've got what I want. Perhaps I want something to be fleshed out before we move on to Raj. We're not going to do that now because I just wanted to show you how this works. But we have this iterative way of getting there and we're still on track to go onto the next one. So we have a long memory. Yes, We did a bunch of little things and feedback specifically for Abby's. And now we move on to Raj. And there's a memory that that's the thing we're trying to achieve. We have to talk about each team member. But now when we go on to Raj, we're going to stick with this style. So once I'm happy with the style, I want paragraphs within sections. That's what I want. Now, we're going to see that same style reflected in Rogers performance review. And in Who's the other person wins the other person. I forget that they're made up people. Nina, Nina is performance review, right? So make sure that you're giving feedback at the right times. And again, I recommend to keep it simple when it's the iterative style because you can do as many iterations as you need. If you write a very complicated thing, it can, it can get lost. You can go off the rails with this. So I recommend using simple verbs, keeping that feedback pretty simple and focusing on iterating to where you want, rather than trying to include everything in one message, you can try both, but I usually get better results when I iterate with shorter messages starting with verbs rather than a very long description. And if you're having an issue with the format, then you should provide that in the initial prompt, follow this structure. There you would bring in the template approach to say, here is the template I want you to learn. Now let's do the performance reviews. So yes, it's getting a little complicated. This is somewhat nuanced because we have sections within performance reviews and then different people and we have a lot going on here. But if we have the right approaches and we know how to use them and we can be adaptable. We should be able to accomplish anything 15. Creating YouTube Descriptions and Titles in Bulk: This is going to be the last example we look at in this section, where we've been going over specific use cases and examples to build up an intuition around the approaches that we've learned in this course. And hopefully you're starting to feel more confident about using the approaches. Hopefully you've started to think about how you can use these in your own use cases, in your own situations. And hopefully you've started building out your own bespoke library of prompts. Now, in this example, we're going to be looking at Building YouTube descriptions. Now, this is something that I often need to do. I have to create descriptions for videos and often I'm doing this in bulk. So I might create one description. And then I need to schedule 20 videos at one time, right? So I want to create 19 more that look just like it. Now this gets tricky because I want part of the description to change and part of it to stay the same. Because in the description, the links that I want to provide are the same every time to a free course, to a website, a store page, and e.g. to a podcast, right? So that stays the same, but the description at the top should change and it needs to have a particular style. I want it to feel a certain way. So I'm going to use a template, but I'm also going to specify something very important in terms of style that may not be captured in the template. If that sounds like a lot. Let's just go through it. And this can be used for a lot of other things as well. But this is where we really start stretching the capabilities. Because we're asking a lot of things to happen here. This is where I have seen a difference between GP T4 and GPT-3 0.5. So I think future versions will continue to get better at this. But four is definitely better than 3.5 because it's more consistent with these really challenging tasks. It can be a little bit all over the place. So learn the structure, format, style, and tone of the following YouTube video description and title. I don't care so much about the title. I can take that or leave it, but we might as well get an idea, right, including the links and emojis. Once you have learned it, say I understand, do not add any additional information. The example is meant to be used as a template for future descriptions. So we know how this works now. Now notice here, this is what I want to change each time and the title, and this is what I want to stay the same. The get a free course. The English community, the store page, write this stuff, stays the same if it changes the hashtags here below that. Whatever, that's not a big deal. But I want the links to stay the same each time, okay? That makes it much more repeatable. That makes it much more scalable. That saves a lot of time and energy, especially if the descriptions are good, right? So we should see, I understand, there we go. Now what we want to do is provide the next topic. And I'm going to specify a couple of things here. I want to specify the topic. So learn English pronunciation through shadowing and not very well-written here. A little bit of detail. I wrote this very quickly because I'm just trying to get the ideas across. I don't need to write it well. Find role-model English speakers and improve your pronunciation by copying their pronunciation. That's not very good writing, but I'm just getting the ideas across now. Guidelines use imperative sentences. Don't use exclamations. Don't change links or hashtags. We're going to try to not change the hashtags will see what we get. If they're changed. I think we're right up against the limits of what's possible in terms of, you know, all of these guidelines and following the template but still getting different results for one part of the description. I think we're really trying to push it here. So we're right up against the edge of what's possible. I'm going to try to have the hashtags stay the same, these here, but we'll see what we get. The important thing I want you to note though, is this imperative sentences. I have decided that in order to give a sense of vibrancy and oomph to my descriptions, I want to start with verbs. Learn this, do that. Well, this is called imperative sentences, giving a command to someone, go over there, that's a command. That type of sentence is called an imperative sentence. A question is called an interrogative sentence. And there are different sentence types. This is where it might be useful to learn a bit of English grammar. The names of things, types of sentences, grammar structures You know, the different names of punctuation marks. If you don't know those that can be useful to learn them and understand them because then you're able to better fine tune the results you get because you're fine tuning the prompts that you make, right? So I don't want to see an accidental we will teach you. I don't want to see it. I don't like it. I don't like that. We think there, unless I want to write with a, we don't want it. So I want to use imperative sentences. If you just ask for a description based on a topic, again, it's all over the place. The style, the voice, right? The point of view, it's all over the place. Don't use exclamations. I don't want a lot of false excitement in there. I just don't like it personally. Okay. So here we go. We've got a title that looks okay, master English pronunciation and shadowing technique. Yeah. With shadowing technique. Yeah. Okay, discover the power of the shadowing technique to perfect your English pronunciation. Choose role. Model English speakers to 0 and replicate what's wrong with me. Choose role-model English speakers and replicate their pronunciation. To enhance your speaking skills. Improve not only your pronunciation, but also you're listening and self-awareness through this effective approach to language learning. I like it, That's pretty good, nice and snappy. Now have the links stayed the same. Get a free English course that's the same. Join the free English community. Same all courses, same, listened to the audio, same other links. It's not registering this as a link because it's a weird one, it's a dot e, but that's fine. The YouTube description box should do that automatically. And then hey, we have our original hashtags. Fantastic. So now we have a real template. Now we don't necessarily have to specify that we want imperative sentences, that we need to maintain the hashtags. So we used a template and then now the results that we get from using that template to ask for the next topic are kind of like momentum. A second template that is a bit more refined and it's almost like we're getting chats UP tea into the habit of using it, which then creates that feeling of momentum. So all I have to do next time is say next topic and state it. Now it does happen once in a while where it just gets off track and goes off the rails. It happens for this sort of complex thing where some things stay the same and some things change. Again, I think we're right up against the limitations of the model and it will continue to improve over time. So you may have to start it, start over if you do six or seven. And again, not just for descriptions for whatever task you're doing, that's a sort of repetitive task that you want to save time on. Bring it back, right? If it starts going off the rails, start a new chat. You use the prompts, train it quickly on the template and then keep going. And that should do the trick. That is it for the use cases, the examples that I wanted to explore with you to build up this intuition for using the approaches in this course. The three approaches, the three main approaches that we've been talking about. Now I'm hoping you're starting to see how they fit together when one is the best or maybe we, we find a way to blend them. And I hope you're starting to get a sense for what's really possible here. Because again, we're only focusing on a narrow set of things. You can now start using these for all kinds of other things. The possibilities begin to open up. Hopefully not in a Pandora's box type of way, but the possibilities begin to open up and you start to see how much can happen, right, with these three approaches. So start exploring those. Let me know if you have any questions. In the next section, we're going to go on and talk about advanced tips and tricks. And you're not going to want to miss these because there's some really cool things that you can do. Now that we have this intuition built up, now that we have a really good understanding of the three approaches, there are some really cool things that we can do to build on that knowledge. So I'll see you in the next lesson. 16. Repurposing Transcripts for Anything: Now that we've built up this intuition for creating customized steps and prompts for whatever we may need to do. Let's build on that with some advanced tips and tricks. In this section, we're going to be exploring some things that can improve efficiency, make things easier, make them faster. We're going to be looking at re-purposing videos so that we can create other things and use information from videos for a lot of different things that we may need to do. We're gonna be talking about responding to messages at scale. We're gonna be looking at automating repetitive tasks. And we're gonna be talking about a little bit, talking about building up our bespoke prompt library. So that's what we're going to talk about in this section and we're going to start with re-purposing from videos. So what we're going to do now is look at how we can use a video as raw material to create other things. Those other things might be discussion questions. It might be a lesson plan. Those other things might be a script for something else. The other thing might be an article. So there are a lot of different ways we can use the video as raw material. Maybe we're just quizzing ourselves. Maybe we're trying to check our comprehension of something so many different applications. I can't even begin to list them all. But as we explore this, I think you'll start to build up this sense of how you might use this approach. So here we have a video. And the basic idea is that we take the content of the video and we figure out how we can capture that content and use it for various things that we may need to do. Now, this might be familiar, a history of the English language. That is then we talked about that already. Yes, exactly. This is a video that I created. This video is the raw material that I used to create the bullet point outline that I then used to create the article that we did at the beginning of the course. So you can see how in this way, information becomes, I guess we could call it platform agnostic or something. It's the information that we want. And we can use chat, GPT, and perhaps other tools too. Move that content across different forms. And that's now quite easy to do if you know how to do it correctly. And that's what we've been spending our time on thus far. Let's say I wanted to write a blog post or article version of my video. How would I do that? And again, there are a lot of other ways to use a video. Ted e.g. is a great source of things like lesson plans. If you want to generate discussion questions, you have a group of students watch a TED Talk. You get the questions based on what we're going to do right now using the process that we're going to go through right now. And then you generate a bunch of different ones for different groups so that they're unique. There are so many possibilities here and it's actually quite powerful. So here's what we do. Basically. Again, it's different for different platforms, but this is on YouTube. And so all you do, and this changes over time, I suppose, but they have a transcript. So we go to the transcript. I want to be clear. It doesn't matter that you do it this particular way, you just need to get the transcript. Now, actually, why is this interesting? Because videos tend to be pretty well vetted. When people create videos, they tend to tend to, I can't speak for everyone. I certainly try my best. They tend to be quite careful. They tend to put in the work the diligence to make sure things are accurate. And so if you have an accurate transcript, that means the information is going to be accurate. The spellings of words, all of that is going to be fairly accurate. So that gives you a really powerful starting material. To the extent of maybe it's better if you want to write something to just talk through it, because that's more comfortable for you. Then turn that into a transcript, double-check the spelling, and then work from that as your raw material to, let's say, create your lesson plan or create your blog post or create whatever it is you need to create. Enough. I'm gonna get rid of the timestamps. I don't want those. And I'm just going to copy the whole thing. And then we go over to chat GPT. We paste the transcript. Now, basically all you need is a simple requests for whatever you need. Do you need a detailed bullet point summary? Do you need a list of general topics on the broad themes? So you have to ask yourself what you need from that, right? In my case, I want to ask for a bullet point summary of the following transcript. Now what I might do is say something like Do not do not exclude any details. Maybe I don't want to exclude any details. I want it to be very clear, very specific. I always ask for bullet point summary because I don't want to have a generally I don't want to have a structured outline. Okay? So I could generate an overall summary of this, maybe with a specific outline structure. If I want to state that, again, you have to use, use specificity to get what you want. Now, if this is starting to look very familiar, that makes sense because this is the process that I did to generate the outline, the bullet point summary or bullet point outline that we used at the beginning of the course. So yeah, that's what it is. Now I would get rid of that channel subscription request and courses advertisement right at the beginning. Don't need that. I might read through it quickly and I might ask myself if everything is relevant to whatever I need to create or not, right? And then I have to decide what I want to do with that. So you use this, let's say, to generate a list of broad themes. You might expand on this to focus on one area may be the way that language is changed specifically and have that fleshed out or add a little detail, you can make changes here. Let's keep it simple and say I want to simply generate a list of 15 open discussion questions based on the summary. Focus on asking why and how opinion questions rather than knowledge related questions. So you do have to be specific because I don't necessarily care about the exact format. I'm just making a request here, right? Maybe I do care. Then you might want to think about creating a template. You might want to think about using some knobs, right? But here it's such a simple request that all I'm asking for is why and how questions. Why do you think the English language has evolved and changed so drastically over time compared to other languages. If you had watched the video, you would have a pretty good idea about that. How have historical invasions and migrations influenced the development of English? In there, you might make a modification and say, of languages in general, in your opinion, why has English borrowed so many words and phrases from other languages throughout its history? Very interesting questions, I think. So. Maybe we just want those discussion questions, but there are some other things that we might decide to do with this. Perhaps, I want to use a specific video as a way to challenge students or quiz them as a teacher. So I don't want to have a discussion, but I want to test their knowledge. Now, the request is going to be a little bit different. Again, we're focusing on the classroom setting, but there are a lot of, a lot of applications here. Based on the summary, create a list of true false quiz questions to test memory and comprehension. So we're trying now to avoid opinion questions we want to focus on only yes, no true, false. Is that right or not? And we want to make sure that whoever has watched the video understands what they saw and what they heard. Now, you might specify here, make all the questions true, make 70% of the questions false, or something like that. In this case, I don't care. I just want to get some good questions, true or false. The English language has only two distinct eras, Old English and modern English. That is false based on my knowledge of the video, what about Middle English? English is a Germanic language and is directly derived from German. Ah, that's a tricky one. That's a really well-written question. It's sneaky because it is a Germanic language, but it is not derived from German. It is a cousin of German. Interesting. Alright, so we have some really good questions here based on this. And I hope from this you're starting to see the potential of doing things this way. Here you have vetted content, content that has been checked and carefully planned. So start thinking about how you can use this approach for things that you need to do. If this is a work related thing and you have a meeting recording, as long as you have everyone's permission, of course, then you might take that meeting recording the transcript from that meeting recording, and you might then generate some specific things from that meeting recording, right? So we're just looking here at an education-related video. That's my background, that's my application. I might want to write an article, of course based on something that I created. I might want to test students. I might want to generate discussion questions, but I might have a transcript of a meeting that happened yesterday was a pretty long meeting. I don't expect everyone to go back and listen to the recording. I'm going to use the transcript of the meeting to generate some high-level insights, right? Provide some high-level insights based on this transcript. I might use the transcript as raw material to simply organize all of the content of the meeting into a more readable form, right? There are a lot of different ways you could go with this. You might even use those meeting notes to generate an outline that you can use to build slides for a presentation. Start thinking about the possibilities. Start thinking about things in terms of outlines. I believe this is the sort of center starting point for so many things. And when I say outlines, I mean that broadly bullet point outlines, in my view are better than well-structured outlines, just my opinion. But start thinking about how you can use this approach in your own work and life. If you have any questions, let me know. In the next lesson, we're going to go on to talk about how we can use another really interesting approach to respond to messages and then how we can automate that 17. Responding with Full Context: Remember when we were talking about building from templates, we looked at a customer service email for a pet food company, where we used a previous e-mail as a template to create a new email with a different issue that we were dealing with. The template approach. Well, what if we're not so concerned that it has a specific tone or voice, or style, or structure. What if we're more concerned about how relevant the response is? Well, in that case, you might want to focus on providing context rather than focus on providing the template. And you can combine the two. But for our example that we're going to look at, we're first going to talk about how we can use context to generate replies to messages or emails that include some knowledge that we have access to that we would like to make sure we're communicating with whoever is sending us that message or email. Then we're going to look at a way that we can automate that because you would want to, wouldn't you? Maybe? Now to be clear, this is not a course about building bots. This is not a coding course, this is a writing course, but I think it's still useful to know a simple way to automate this kind of task. Okay, so the first thing we'll do is teach chat, GPT, a little bit of information, share a bit of context. Context that can be used to answer a question to respond. So I've created some fictional, some fictional details here. And this is quite long, so we're not going to go through the whole thing, but I wanted to sketch out the idea. So act as a customer service representative for potato landscaping landscaping company that provides high-quality full-service landscaping services. Full-service means they do everything related to landscaping. Your job, Jessie beauty is to respond to customer questions in the message box on the potato landscaping website. Even if that's not the case, maybe I want to frame it that way. I'm going to use it to respond to e-mails or whatever. But I want to frame it this way perhaps because I want to put Chad CBT in that headspace of responding to messages in a concise, maybe respectful manner. Use a respectful yet conversational tone consistent with customer service backprop best practices. Keep messages concise and I'm not saying you can't combine this with the template style to achieve a specific style of response. But if you were using it in a sort of messaging capacity rather than answering emails, that might be a little more challenging because you might get very short messages or very long messages, right? So you might want to have that flexibility. And this goes on. Information about potato landscaping. I've provided some fictional details here. I've given some background. I've talked about the services that are available. If you understand my request, say I understand and then wait for the first message. Okay, So this is what we're going to set up as our context. Now, just for testing purposes, let's ask a simple question. Let's put in a very simple request that someone may ask. And again, you might be thinking, okay, well, that sounds a little tedious to copy and paste all of that. Yes. I tend to agree if it's for a specific message, but if you're not responding to messages that often, you're not responding to emails that often. Maybe it's not a big deal to quickly grab that from your prompt library or go to the most recent time you used it and simply paste in the next message. Because if it's in the same chat, you can just paste in the next thing and it's going to apply the same general rules. But again, we might want to automate that as well and we'll talk about that next. Okay, So a simple question here. I want to get some quotes for tree removal as soon as possible. Okay? This is the simple message that we get. Hey, there, we would be happy to. And notice that the tone is, I could change it in the initial paragraph, right? But I'm pretty happy with that. Hey, there we'd be happy to help you with tree removal. To give you an accurate quote, will need some more information like the size and location of the tree. Please provide those details and we'll get back to you with a quote as soon as we can. I'm pretty happy with that again, because this is a generic version of that. Then what we get here is going to, I think, ring a little bit generic. Yes, we can help you with tree removal. But that's also based on the information that is provided. So if you include in the above the hours of the business and the contact email or phone number, and the specific tone and style that you want to use to make it more conversational or less, or don't use formal language. You may think about then adding a few knobs there, right? We mix and match, right? We use what we need to build bespoke prompts for specific purposes. So bringing a knob or two until you can tune it so that you liked the responses that you get at the granularity that you need to be able to provide the level of detail that you want to in your responses. And also bear in mind what you're going to be using that for. Is it for messages to respond to quick messages, or is it for emails? In which case, maybe we would try to combine that with a template. But what about automating this? Well, let's take a look at how we might do that. 18. Automating Writing Tasks in Zapier: When you're trying to accomplish something specific in chat GPT, it's nice to have that back-and-forth to be able to iterate. Because you're creating something, you're writing a script, you're organizing meeting notes, you're building a presentation, a slide deck. So it's great to have that back-and-forth. But sometimes we want to leverage a very specific thing that chat GPT can do without having to have that back-and-forth every time. We were just talking about answering, responding to messages or what about emails? Well, this is not optimal to have the back-and-forth, to have to paste things in every time. So can we automate? Now, if you are very experienced and you know how to code, you probably have the answer to that already. You can probably build a bot in 5 min. But what if that's not the case? Well, I want to take a quick look at how we might, might build a workflow with something called Zapier. Now, if you don't know about Zapier, you've probably heard of it, at least. Zapier is quite simply an automation tool. And they have some free features. They have a paid version of free version. And so there are limitations here, but I just want to go through this because it is, It's really cool. I use it for all kinds of things. I use it to link together different websites as I need to. I use it with chat GPT. Now how do you do that? Well, you can connect to chat GPT another way via the chat GPT API. Now there are actually two right now, there's an open API API and there's a chat GPT API. There are a little different, they work a little differently. So I don't know where that's going to go. I don't know if they will keep both. Where it will always be the OpenAI API and the chat GPT API. The OpenAI API. That's a mouthful, is, let's say it's a little more broad in a sense, they're more knobs, more things you can tweak. Whereas the chat GPT API is more like creating a specific role or function or thing with chat GPT via Zapier or something like it, and some other app or platform that you connect to through Zapier. So it's actually pretty easy to use. You just connect accounts, your YouTube account or your Google account or your OpenAI account, whatever. The only kind of complicated thing about this is that you have to get an API key. So you have to sign up for a specific account on OpenAI website and then go to your account and create an API key that then you use to connect to Zapier. Basically, all this is doing is linking together apps and actions inside of apps. Not everyone is available, but that's essentially what's happening and it's doing it with actions, triggers and actions. So there's a trigger, that's a thing that starts something that then causes another thing to happen. That then causes another thing to happen elsewhere. And you can accomplish a lot with that. E.g. I. Use the chat GPT API. With Discord. I have a discord community. And every day at a specific time, There is one trigger at a specific time that sends out basically a little ping that says now at a specific time in the morning, that then triggers the chat GBT API. And there's a prompt inside of that in the Zapier settings that instruct the chat CPT API to create a totally custom word of the day kind of announcement thing that goes out every morning. It has a word of the day, it gives some example sentences, and then it creates a quiz that can test your knowledge on the word of the day. And then that connects to discord. So the trigger and then chat GPT and then discord. And it's just a three-step, a three-step process, right? So we're going to do this with Google Sheets and the chat GPT API. And this is already basically set up and it's very simple. So we're just going to look at a super simple version. Now, I get better results with Google Sheets. But you might use another one. You might use Notion for our purposes, google Sheets actually works pretty well. So the first step is a trigger in Google Sheets. What is the trigger? The trigger is that there's a new entry in a specific column in Google Sheets. And that looks like this. So this new entry is, this new entry in Google Sheets is a message. This is just to sort of demonstrate it, right? That's the trigger. And you set that in Zapier Great. So let's just quickly go through this process. And again, it's very straightforward. The accounts are already connected, so we won't go through that. The action that I want is new or updated spreadsheet row. And you can choose from different, different ones, but I want a new row in a specific column. Okay? And then I'm going to, I'm going to choose my account and I'm going to choose my trigger. This is the specific thing that's going to happen, that's going to begin these steps. So which spreadsheet is it in? Of course demonstration? It's worksheet one. And the trigger column is called text. That's the name of the column. So the text column over here, at this column, this one here, text, right? This is the one, this is the one that we want. Okay? So that's fairly straightforward. Then we might do a test, but we're going to just continue. Then we can act. Chat GPT were already linked. So again, you just need the OpenAI key. The only event available currently is conversation. It's the only thing you can do. You connect the account fairly straightforward. Then you basically are just creating a chat GBT prompt. That's all it is. It's not much more complicated. Respond to the text message following the guidelines I provide, use only lowercase letters. Do not use emojis, use a thoughtful and considerate tone. Do not use exclamation marks. Ask a follow-up question only if it makes sense to do so. Here is the text I want you to respond to. Then this part is dynamic and changes depending on whatever the new entry in that column is. And here it is. You haven't, you have an option, right? So you can, you can here essentially just summon one of the things from the spreadsheet because it's connected. Very easy to use, actually, very straightforward. We're going to use, should we use GPT-3, 0.5, or should we use GPT for? Let's use GPT for, why not? Alright. And we're going to name our assistant Kathy. We're not going to test right now. We're just gonna go for it. And then we want to update the column beside it. We want to update, we want to add a new row right beside that so we can see the response. Now I want to be clear, I'm doing this in a spreadsheet, but you would link together perhaps other apps, okay, message received here, right? Then that goes into a spreadsheet column and you maybe review it first and then send it yourself. Or it goes directly back into wherever the message came from. There are a lot of different ways that you can do that, right? So basically, I'm just choosing where I want that to go. I want that to go back into Google Sheets, continue, publish and turn on. And this is gonna be version. Okay, so now that it's on, what we wanna do is test it. Alright, so our Zappa's on. We're gonna go over to the spreadsheet and what we want to see is a message here generates a response over here, and that is successful automation. Now, we want the response to be all lowercase. We want it to be respectful and we don't want to use exclamation marks. Okay, So what should I say here? I am having trouble thinking of an interesting text message that I could send here. Any advice. Okay. Alright, there it is. Hi there. It's understandable that coming up with interesting messages can be difficult. Sometimes I think about what you're passionate about or something that's been on your mind lately and let that inspire your message. Okay, well, very dramatic, but just a demonstration of how the infrastructure works and how fairly straightforward it is. But now you can start thinking about, could I take the template at approach and put that into Zapier or something like it to generate those responses very easily and I can just quickly review them at one time. Or could I have those messages? Are those emails sent automatically? I mean, maybe maybe not. Depends on how comfortable you are with just sending things that have been generated by chat GPT without checking them first, personally, I like to check them first. What about messages that you're receiving? Well, what about other things too? I mentioned the use case that I have on Discord. This is where you might start thinking about the approaches that we've explored in this course and the specific use cases that you have for those approaches which times do not need to have the back-and-forth, where it would be nice to be able to do that without having to do it while you're asleep. Then which pieces need to be connected together in order to make that happen? And what is the exact single prompt that I need to make that happen with consistency every single time, whatever it is, right? And this gets also into productivity and other areas outside the scope of our course. But useful to start thinking about this if you haven't already. What can you automate? Where does it make sense to do that? Maybe it doesn't work for you. Maybe you don't like to really do things that way. Totally fine. But at least now you can start thinking about that for yourself, where it might make sense, where it doesn't really make sense. You have to really think through all of this to create your own processes, to optimize your own workflows in a way that works for you. So that brings us to what we are going to be talking about in the next lesson. In the next lesson we're going to very briefly touch on the prompt library 19. Crafting Your Prompt Library: Before we wrap up this section, I want to very quickly focus on the prompt library. It's something that I've mentioned throughout the course, but we haven't actually talked about what that means exactly. Although I think you know what I mean, building your own prompt library, essentially, creating your own steps and prompts for tasks that you need to do. Now, far be it from me to tell you how to organize that. Do it in a way that makes sense to you. Just as a little bit of perhaps inspiration, I'll show you how I do it. The way that makes sense to me. But again, whatever, right? Whatever is best for you. This is a sample one or one that is based on mine because well, I don't know. I just I didn't want to share all of the things that I need to do in a given day. So I just made one that is close to my original prompt library, the one that I use on a daily basis. So how I do it is I use Notion. Notion is a pretty cool tool for organizing things, creating notes, creating calendars, all kinds of stuff. And I do it as drop-downs because I find it easier to scroll through very quickly. Again, you can make it a table. You could organize it into individual pages. There are so many different ways that I think you could structure this and organize it. Once you reach a point where there are so many and I'm starting to do this, then you might have prompts within tasks. So I start with general things I need to do. I need meeting notes, I need meeting agendas. I need video descriptions, I need announcements, I need a cover letter. I need replies to emails, right? But I might then begin. And this is again something I've started to do recently, creating bigger categories with these inside of larger categories. So Notion allows you to create drop-downs of anything so you can do drop-downs of drop-downs. And I personally like that style. I like being able to just quickly dropped down and see what something is rather than opening a separate page. But by all means, do it wherever you're comfortable. We all have our own way of keeping notes and our own way of organizing things. So e.g. I want to create meeting notes. It's right there, dropped down. Boom. And I can just copy and paste that template that I use for meeting notes. And when I'm done, I'm back at the main, sort of main view. And it's so simple, it's stupid, simple, but it works for me. If you want more granularity, maybe you want to create broader categories. And you want to have again pages inside of those. But start thinking about it. Make sure it fits how you think, right? I think in terms of tasks, maybe you don't think in terms of tasks. Maybe you think alphabetically, well, okay, then organized them alphabetically, organized them in a grouping that makes sense to you and no one else, right? I think one theme that we've been really covering and touching on throughout this course is, yes, there are general approaches that are useful. Yes, we can explore these, these examples for specifically how to respond to e-mails or write blog posts or organized meeting notes, or generate an outline for a slide deck or whatever. We can do all of this and these have things in common, but the one that is going to work for you is unique. You're going to have to customize your own and build your own process, your own steps and your own prompts that work for what you need to do through testing, through iteration. The same way in the way that you organize it if you just sort of, let's say follow what I do, but it doesn't quite work and it's hard for you to find things and it's frustrating. Again. This is a new area and it's constantly changing. So constantly evolve and change your approach. And think deeply about everything that you decide to do. Because if you build really good prompts that allow you to accomplish a lot and produce great results almost every time, then you're going to save a lot of time. It'd be much more efficient and effective in the way that you do things, right? But that comes down to you thinking through all of this, organizing your prompts, creating your prompts, creating your library, and the way that fits and suits you. I think that goes without saying, but I think it is worth saying because I'm trying to get across that there is no standard of final knowledge on this stuff. It's relatively new. It's, it's very new. In fact, it's going to continue to change and evolve dramatically over the next however many years. So I think it's better if we, instead of thinking of the perfect prompt that I found on this course, I'm going to copy and paste it instead. Think in terms of this is working for me now. I may change it in a month, right? This makes sense to me now. I may adapt and change in the future. We're always going to be able to keep up with those changes and surf the wave of any innovation that comes in this space so that we can continue to be competitive, producing great stuff, getting great work done, right? So that's a bit of a soapbox rant. But I think hopefully worth saying. Now, in the next section of the course are going to talk a little bit about ideation before we go on to talk about feedback, proofreading, and correcting your writing 20. A Soapbox Rant on Mindset: In this section of the course, we're going to very briefly talk about ideation. How do we come up with and develop ideas that we can then turn into really solid writing for whatever. We're gonna be, looking at some specific prompts, a few approaches that are useful for overcoming blank page syndrome. And then something a little bit more specific about how we can use chat GPT in a back-and-forth way to really develop ideas in depth. And we'll look at how we can do that in an adversarial way, which is very interesting this before we go on to our next section, where we'll be focusing on feedback, editing and proofreading with chat GPT, it's hard to get good at anything unless you do it often. So we've been doing a lot of hands-on stuff throughout this course. And hopefully that's starting to build your intuition for how you can use chat GPT for whatever you need it to do, particularly focusing on writing for this course. But in order to really build out that intuition, really test the limits and maybe you're already doing this. It's very important to habituate to thinking. Can I do that with chat GPT, or whatever AI tool you want to use. Can I do that with this? So instead of listening to me, hearing me say, you can do this with chat GBT, Here's the template approach, e.g. okay, great. But if you really want to continue to see what's possible and find the best approaches and tools that are going to allow you to be more effective, be more efficient in whatever you're doing. I would recommend you, and this is what I do. So that's why I'm recommending it. Ask yourself for anything you need to do. Can I do that with touchy PT? And just starting with that, assuming that maybe you can then allows you to really see what the limitations are, right? That's how I start everything at this point because I always want to figure out a new approach, a new method, a new type of prompt, how I can change something in tweak something. So I always start with the assumption that, Hey, maybe I can do this. Can I do this with chat GPT? Maybe I can, let me try. And then once I try, I might do some research, I might read, I might go back and forth and then figure something out. So this is also a way to habituate, to thinking in a certain way so that you're always pushing the limits, so that you're always getting the most out of it. And again, whatever tool you happen to be using so that you can get the best results so that your life is more convenient and what you create is ideally better, right? The other thing in terms of just the overall mindset, is to be able to distinguish different types of creativity. I think we can say there is easy creativity and there's really deep creativity. I think chats CPT is very good at easy creativity and not so good at deep creativity. So what do I mean by deep creativity? Well, not to toot my own horn, but I gave the example of the discord community and having a trigger go out every morning. That then asks the chat GPT API to create a word of the day with examples and a quiz at the end, right? So coming up with that idea is a sort of deep way to integrate Chat GPT into a process and learning process, a community, a workflow, a way of creating content or articles, a way of teaching stuff, right? These sorts of things where we come up with an idea that solves a deep problem that we have. This is, this is what we're good at because we're able to see the whole field, the whole picture. But while I could come up with the word of the day and I could create the examples for the word of the day and I could, I suppose build the quiz. That's easy creativity. The easy creativity is the one that I feel better to outsource. Yes, I could do that. But we'll anybody care if I do it or if Chance GPT does it, If it's basically the same quality, that easy creativity of pick an interesting word, come up with some examples. Build a quiz where you choose one a, you know, a, B, C, or D. No one's going to know if I did it or chat CBT did it. It's easy creativity. And that's just an example. But being able to distinguish between what am I good at? What is chances are pretty good at, will allow you to start to make optimized decisions about what, where to allocate your time and energy and knowing what to outsource? Yeah. Probably maybe don't outsource Deep level problem-solving at your job, right? Because there's just, there's not enough information to get the answer, but you can reflect on that. But what processes in your job are the type of creativity that you can outsource to this wonderful little tool that can help you right now that you know how to build really good prompts to get high-quality output, high-quality writing, even better, fantastic. Write your life should be easier now. So I think it's important to think about where those lines are and continue evolving them over time. It also helps me to see myself more as a lot of ways. E.g. writing an article, writes The History of the English language. Am I there? I created the outline, so my creating it. Well, I'm creating one form of it. I'm creating the ideas. But when it comes to the actual written part, the article itself, I'm more of a curator. It's my taste that is making me decide to go with the first three paragraphs, but get rid of the fifth, the fourth, and the fifth, and use the fourth and the fifth paragraph from two other versions and then get another. Do it again, go through the first, the first approach that we talked about, outlines and knobs again and iterate that I am there a curator. Alright, so think about the clothes that you wear. You don't make the clothes. You buy the clothes. So what is it that makes your style your image? It's not that you make everything, it's not that you do everything. So I don't think we should feel we shouldn't look down on ourselves for not doing everything. We're now, we're shifting into a new way of working and creating where we do create stuff, but also we automate stuff and we outsource stuff. And we create with tools as well. And when we create with tools, we become curators. So just as you're a curator with your style and the clothes that you choose, like why am I wearing a blue corduroy shirt? I'm not sure why am I in front of a 1970s looking pastel spaceship living room, right? It's curation. That's what it is. And as a curator, your choices define the final product, not you doing everything. And just shifting into that mindset opens up this new vista of, Oh, well, that means I'm actually now capable of way more than I was before. It's not that I'm being replaced that by thinking differently now or different than I'm able to now work in a different creative landscape which is broader and I think personally more interesting. Another way I like to think of chat GPT is having 10,000 interns who are pretty good at many things. Wouldn't it be great to have 10,000 interns? Well, yes, that's great. I can solve many problems that can generate ideas. I can do so many things now that I couldn't do before. Fantastic. I can get summaries of videos and all of that. But in specific areas then you might want to move from pretty good, too. Good, great, excellent. And so there, you need to get more involved. It's like picking out one intern who's your email writing in turn and training them a little bit, giving them more skills, giving them more feedback until they get beyond pretty good, right. But then in all the other areas, you still have those 9,999 pretty good interns. And then over time you start to develop the skills of the ones that are most relevant to what you need to do. I don't know if that helps. For me that helps me build a framework of okay. So I've got a general-purpose army and I can train individuals in that army to do specific things as needed, like writing articles, like organizing a meeting agenda, right? Like answering emails. And I could do that for all of them, but if I know how to do it for a few key ones, that's a big win. So those are just a few insights that I've uncovered working with chat GPT essentially every day since it was launched. Things that helped me navigate this very useful tool that continues to improve over time. So in the next lesson, we're going to be talking about how to overcome blank page syndrome. 21. High-velocity Brainstorming with ChatGPT: Is there ever a time or place for a very simple, very generic prompt like write a resume or write an email? I think the answer is yes. And the reason we haven't focused on that approach in this course so far is that it's very easy. You just say what you want and you get something back. But it comes with a downside. And the downside is that if you're expecting high-quality writing or something specific, you're probably not going to get it. And then it's this endless back-and-forth where it's not the Star Trek computer. It's kind of this just endless hacking away, changing things, copying and pasting, hoping you get somewhere feeling lost and just ending up with something that's kind of blah, right? That's what usually happens. But there is a time and place for it. And it still has, I think, a function, uh, use those sorts of very simple prompts. So let's just say write a resume, okay, now why in the world would we do that? Write a resume, we get something that's fairly generic. It doesn't have a lot of specific details, right? What we're getting here is a fairly rough template of a resume with some specific language around descriptions. Eager to contribute my expertise in blank to create value for the company and further my career development. It's quite generic. I mean, that is the ultimate generic sentence at the beginning of at the beginning of a resume. A resume, It's not something you'd use. You would want to rewrite that completely. But what are we getting? Well, we're getting a structure. We're seeing at least how we might start thinking about the real resume. We're getting some ideas about if we have no starting point. Ah, oh, yeah, I should think about my certifications. Oh, I forgot about the one that I got back in 2015. Should I put my skills after my education or my education after my skills? I think my skills are more relevant, so I'm gonna put my skills first, okay? So we start to feel this. Now that we have something there, we start to feel this desire to create something that's better, right? And that's not to say you would begin iterating from here, but just starting from a very simple, Let's say generic prompt and getting something, anything can spark ideas, can get some wheels turning, right? And that's not necessarily a bad thing and it's often a good thing, right? What if we want titles? Give me, give me ten YouTube titles, titles for a video about hamsters. Now these aren't going to be specific to a topic that I have that specific about hamsters, maybe I have an angle that I want nothing like that. I'm just asking for ten titles. But what I might get is some thought or inspiration that then I might use or refine or add to one of the three approaches that we've talked about. I might create a template where I generate titles based on a specific style. And I got some ideas about that style from the ten titles that I got. Hamster Olympics, the most entertaining hamster agility courses fit this. Nothing, nothing to do with my hamster video because I know what my hamster video would be about if I did make one hamster care one-on-one The Ultimate Guide for new owners. But maybe that responses and negative responsive fashion. I don't like this thing where we have this little title, where there's a colon and then there's a longer title after it. I don't like that. So now I've I've gain something from what I don't like. I've gained something in relief. I've gained something and it's sort of response, a negative response to something I see here. What's with all, what's with all these. So maybe I get some ideas about my video title, Knowing that I want to use just this style without this sort of colon thing because it produces a bad taste in my mouth. Again, that sort of thing is an idea. That sort of thing is, it's not even a starting point. It's something, right? And so this is, I think the value of a very generic request of any sort is it gets the wheels turning, it sparks something, it gives a response, right? So you're leveraging chat GPT is willingness to just do something. You're never going to ask. Well, I suppose that can happen if you ask something that's against guidelines, but you're never going to ask something that is okay to ask. That then makes chat GPT say, I've gotten nothing. I don't have any ideas. I'm sorry. Will you might a friend might do that. You might ask a friend, Hey, I'm trying to think of a clever title. You got any ideas? What's it about hamster video? Man? I don't know. I had a tough day. I'm gonna go take a nap. Chat GPT, is there ten ideas? They're all crap, but there they are, right? And so you're leveraging that as the raw material to sort of get those wheels turning. And I think even if it's even if it's terrible, even if it's bad, that it's kind of pushing you in some way that is getting you started. Because often the biggest roadblock to getting started on something is the blank page syndrome of staring at a blank page, of not knowing what I actually wanted to talk about, of not having an outline. So I have an idea for a video. Oh, what's the idea for the video? I want to do a video about hamsters. About what? This training hamsters. Okay. I'm talking to myself. Okay. Okay. Training hamsters but I don't know anything about training hamsters. Okay. So Chad, CBT. Give me give me an outline for a video about training hamsters. Alright. No requirements. No, nothing Roman numerals allowed. Hamster training, one-on-one teacher hamster, fun and practical tricks. There's that Roman numeral introduction, brief overview of hamster training and its benefits. Setting expectation what hamsters can and cannot do preparing for hamster training? No, I don't want to have preparing for Hampshire training. Preparing isn't important. It's all about building trust. It's all about the techniques. So I would skip something like this, but I would definitely keep something like this. But this is really the meat of the video. Okay, so I'm going to, I'm going to not develop ideas based on this spark that is allowed me to overcome my blank page syndrome. That hopefully will turn into, let's say, a solid outline that I create. That then I can develop maybe using one of the three approaches. Or maybe if it's a video, I'm just talking through the outline, whatever. But at least I had a starting point. I had that raw material that gave me a positive emotion and negative emotion, frustration, right? To get started, right? Frustration is actually, can be a great emotion if you feel frustrated with yourself, if you feel frustrated with some dumb outline that chat CBT gave you for a hamster video. Use that frustration. React against the chat GPT to create something cool, to create something interesting, to start the creative process in earnest. So food for thought. Next, we're going to talk about how to actually develop ideas. So you have an idea and you want to develop and refine that idea with chat GPT fast what we're going to focus on next 22. Summoning a Thoughtful Interlocutor to Develop Ideas: Maybe you've got an idea for a project, a social media campaign, a business idea, a team-building exercise. You've got this little nugget of an idea and you want to develop that, refine it before you write a proposal, before you create, maybe a business plan before you take the next step to make something real with it, you want to brainstorm a little bit, you want to hone it, you want to develop it. Before we go on to talk about feedback and proofreading with chat GPT, getting back really to our primary focus of great writing with chat GPT, I'd like to give you one way that you can refine an idea, that you can develop an idea with chatty BTS, your partner. Now, like we've done with a lot of other things that we've worked on in this course. We want to set up some parameters. You could call these knobs of some sort. Now, this approach, I suppose, is closest to the Star Trek computer approach. We don't have a specific template that we need to use because we're not really creating anything yet. And we're not drafting an article or something longer like that. So we're not really using an outline, although you could add knobs to this if you wanted to. But really it's closer to the Star Trek computer approach. So whatever that idea is, you want to put that after the criteria, after the prompt that tells chat GPT what you want it to become, right? So we might use something like this. Act as a potential business partner, co-founder to help me develop an idea that I've been thinking about. Okay? Ask me questions and add your own ideas. Just as a potential business partner Would. You can add comments, agree with me, or play devil's advocate. Now this is something along the lines of a knob because we are creating the type of person we want. If we don't want agreement and we don't want extra comments, but we just want devil's advocate. Then we would just leave that here, play devil's advocate and come up with arguments that go against the idea that may weaken the idea. That's a useful role. Very useful, okay, make it a natural back-and-forth conversation. Only ask one question at a time. And as we looked at in the Star Trek computer approach, often we want to do one at a time rather than, okay, here's my idea. And then we get this giant list of questions about it. And then it's not really a conversation. So I'm trying to create a potential business partner here. Here's my idea. I'm thinking about starting a food business that sells high-quality breakfast food from trucks. Alright? So probably we're going to have to supply some details. We're going to have to think through it a little bit. And then we may have to answer some, some tough questions. So let's just do a couple of questions here just to sort of get a sense for this. Okay? That sounds like an interesting concept. What kind of breakfast foods are you planning to offer from the trucks? Generally, I would like to offer traditional home food, British style breakfast, American breakfasts, and other breakfast related cuisines from around the world. Eventually, offering a variety of breakfast options from different cuisines is a great idea to cater to diverse taste. How do you plan to ensure the quality of ingredients? I haven't really thought about ensuring the quality of ingredients, although it would be very important to have high-quality ingredients, It's definitely crucial to maintain high-quality ingredients for a successful food business, you might consider partnering with local farmers and suppliers to source fresh, organic, high-quality ingredients. How do you plan to market your food truck business and attract customers? I have no idea. I guess I might advertise on Google. Advertising on Google is a good starting point, but you might want to consider other marketing strategies. For instance, for instance, you can leverage social media platforms like Instagram, Facebook, and Twitter to showcase your menu items and daily locations, collaborating with local influencers. I like that suggestion. And fluid, fluid bloggers, food bloggers can also increase awareness and attract customers. Have you thought about how you will set the pricing? Know I haven't thought about that Any ideas. Pricing is an important factor in determining the success of your food truck business. Here are a few ideas to help you set the right pricing for your menu items, cost analysis, competitive pricing, okay? So we're just talking it out and getting some rough ideas. But one part of this is it's just forcing me to say it out loud or write it out loud so that I can get the idea out of my head, right? And if I don't know what my pricing strategy is going to be and I don't know how I'm going to source materials and I'm not sure about marketing. I can use this at least as a starting point to begin answering those questions. Now, if you want to refine this, you can add more upfront, right? If this is the Star Trek computer than we might want to add some criteria. Or we might, we might want to add some detail about who exactly is this potential business partner. Using the example we looked at where we provided business details, right? To answer messages on a website if you remember that example that we looked at. So you could give a background to chat GPT in this case, to build essentially a character that you're going to have a conversation with. So instead of maybe this being a potential business partner, I want it to be a very critical investor who maybe is not very interested in investing in the food space and can only come up with reasons why they don't want to invest in my idea. And then I will really have to answer some tough questions again, just to force me to think through everything. But if this is a movie idea, a short film idea, an article idea of video idea, a team-building exercise idea, an activity for my kids, idea, whatever it is, I can create a partner, a back-and-forth partner to help me work through it. And I find that to be very useful. I do this pretty often when I want to sort of develop something a little bit further. For this one, I think there's a long way to go in terms of how good that conversation can be, how deep that conversation can go. I think in the future that's going to become much better. But at least for me, it's very interesting and very useful as a starting point to flush things out a bit. Okay, So let me know how that goes. We are going to move on to our last section of the course. We're going to be focusing on feedback using chat GPT for feedback, corrections, proofreading. So there's a lot more to cover, and I will see you in the next lesson. 23. Correcting Grammar and Spelling: With a solid foundation for how to create good writing with chat GPT for any situation. And an intuition about how to mix and match the approaches based on what you need to do. And then build out your own bespoke prompt Library for those things that you need to do with a sense for how we can come up with and then develop ideas for the next thing that we may want to create, for the next thing that we may need to do. The last thing to talk about is feedback and corrections. And so we're going to spend some time on that. Now. Now we're gonna go through this from, let's say, simple to slightly more complex. We're going to start with a simple way to get direct corrections, but see what has been corrected. And then we're going to go on to talk about slightly more in-depth feedback with a revision and an explanation of those revisions. Then we're going to look at how we can compare something that we've created are written or perhaps said, e.g. if you're working on your English, if it's not your first language to a standard. And there are a lot of interesting use cases there. We're going to look at specifically how we can do that to get a particular type of feedback on what we write, what we create, and perhaps what we say. So first, simple, simple feedback. Now, one issue that may not always be an issue. One issue when you ask for a revision on something, let's say it's an email you've written. You wrote an e-mail, you wanted to write it. Great. I'm not shaming anyone for writing their own e-mails that go ahead. One issue is you ask for a revision and you have to then read through both versions, your version and the revision to see what's different. Right? Now it depends on if that's just a revision in how it's worded or if it's actually corrections. But let's say you want it to be a revision with corrections. What you want to be able to see very easily is what changed. What did you actually fix? Okay, so let's look at a simple way to do that. Now, this is similar to something like Grammarly, and I have nothing against Grammarly. Grammarly is great. So if you just want a basic corrections on spelling and simple grammar stuff, great, right? You can use Grammarly and you have to paste stuff back and forth, but okay, now you can also generate texts using AI inside Grammarly. But eventually you hit a paywall there, at least currently, I believe I don't know where they're going with that. Right. This is just me, but I like to have a lot of control over all of the knobs, over the outlines. I like the iteration, the Star Trek computer, right? I personally, personally, and this is why I'm making a chat GPT course rather than a Jasper course or Grammarly course or something. I personally like the vast range of things I can do that 10,000 interns, which includes writing, but also productivity. Productivity in all kinds of other things that I'm able to do. I like that flexibility. I like the iteration, I liked the back-and-forth. So that's just me. So can we recreate something similar to grammarly using chat GPT? Let's take a look. Now, the simplest prompt I think we can use is something like this. I want you to correct the spelling and grammar of the text. That's pretty straightforward. Now, you could just say correct the spelling and grammar of this text colon, and then paste whatever it is you want corrections on that paragraph for that essay or whatever you have. But again, if you do that, probably you're going to get just a revised version. And then you have to look through and compare. Now we go further, show the corrections in bold so that I can see what has been corrected. Here's the text, so that's a simple change, a very simple twist that gives you at least some sense of, okay, well now I know what's different between these. And I might decide to actually go back to my original version for some, some things. So let's see what we get with this. Now. What I paste it here, by the way, is, is a pretty bad text. It's got spelling issues. It's got a lot of problems. We won't read through it, but it's, it's not good. We have a U instead of a YOU. It's not good. So the first one is, I hope you're doing well. A classic pet peeve and a big time. No, no. Y-o-u-r instead of Y-O-U apostrophe R-E. I hope you're doing good. I wanted to catch up since it's been awhile, since we last talked. I've been really busy and there's been bins. So what did we have before? It's BIN and I've BE n Okay. That's an issue, right? How's life been treating you lately? I've been thinking about the past and how much fun we used to have together. Okay, again, don't focus on what this paragraph is about. This is just the approach. The point here is that what we want is to get simple grammar and spelling corrections. And that's it, right? We can maybe get rid of this. Let me know l m k, because the original, the original one has L, just lm k. So it's keeping that but offering the option to keep l m k, I don't think it's common enough to keep maybe we get rid of that. Let me know when you're free and we can plan something now. Is this then very good writing? Well, no, I wanted to catch up since it's been awhile. I've been really busy. This is not necessarily good writing, so we would have to take it a step further. This is just our first step in the simple stuff and all we want to do is see what's different. Now you can tune this a little bit to say something like, I want you to correct the spelling and grammar of the text. Also. Maybe make revisions to improve the style and quality of the writing. And then you could add a couple of knobs. Then you can say, right, make sure to show the corrections in bold. Fair enough. We might do that, we might try that. But I want to explore an approach where we can actually get a list of things that are different in addition to a revision. So with this simple approach, I encourage you to play around. Start very simple with this one, right? And then see what you can do with it. And there are quite a few things you can do with it. The simple approach of asking for revisions and feedback using the knobs that we talked about earlier in the course to hone things like style, syntax, and structure. Making sure to ask for those changes in bold so that we can compare the versions easily. So now let's go onto another way to get slightly more in-depth feedback on something we've written 24. Proofreading and Full Revisions: We've looked at a simple method for getting corrections on grammar and spelling by asking for whatever is different to be in bold so that it's easy to see. Now we're going to take it to the next level. Now, this is a bit more in depth. We might make more changes. We don't just want corrections on spelling and grammar. How can we do this? Well, to do it, we need a proofreader. A proofreader is someone who goes through your writing and tells you, well, you need to change this if you want it to be more natural. There's a syntax issue here. You've repeated this word too many times. In addition to, here's a grammar issue, Here's a spelling issue, right? So they're checking more than just grammar and spelling. It's broader, right? And we want a full revision. So act as a proofreader and make improvements, not just Corrections, make improvements to the proposal I provide. Make a full revision. So I want the revision, right? I want to have the new version, but I still want to see what's different. We could do it with the Show Me in bold approach, but we're going to do it slightly differently this time, which has different applications, make a full revision with the improvements. Maintain the overall voice and style of the original. Just the voice and style, not the wording, not the phrasing, just the voice and style. Maintain the overall voice and style of the original, but feel free to reword sentences to improve overall syntax and style. Now, feel free is kind of an odd phrase to use here, right? This is the one that I found to work quite well, basically to give permission to chat GPT to make changes. Instead of saying you must make changes because if something is fine, leave it alone. And I encourage you to experiment with different ways to phrase that. But I found feel free to work quite well, sort of like saying, Hey, you make the call, right? But if it's great as is, leave it alone. Okay, and we want to improve the overall syntax and style, meaning we don't just want to have the corrections in the revised version after providing the revised version list all the revisions, improvements, corrections that were made. Now, this has different applications, e.g. if you're a teacher and you want to provide those students on something they've provided to you, then you would build this in a way that works for you. And again, I'm not saying that this is the one you have to use, play around with it, experiment until you get the results that you're looking for, right? But also think about how you want to see the changes. Do you want that as a list? And you could send that list to someone to say here's what's different in an updated version or maybe for keeping records, or is it easier to see it with the bold text? Whatever works, use what works for you. So here's the proposal. We're not going to go through the proposal, but I wanna, I wanna give my proposal for considering. It's a great idea. You definitely need to consider it already. We can tell it is no good. So here we go. I would like to present my proposal for your consideration. It is an innovative concept that I believe is worth your attention. So there we have not just corrections. If we were doing just corrections, what we would see there is in the original, I want to give my proposal for your consideration. It's a great idea. You definitely need to consider it. So that would be the grammatical correction. But would that be good writing? I think anyone would say, no, why not? Well, it doesn't really sound very professional. It doesn't sound very clear. And also there's the repetition of the word consider, which is not good, right? Especially for something like this. So this is where the giving permission to chat GPT to make those style revisions comes in very handy. I would like to present my proposal for your consideration. It is an innovative concept that I believe is worth your attention. The proposal focuses on the development of a new product that has the potential to gain widespread popularity. Although I have not yet conducted thorough research, my intuition suggests that this gadget will be a significant success now, is that better than the original? So my proposal is about a new product that will be popular. I didn't research it yet, but I feel it will be a big hit. So is it better? Yes, it is. It's way better. It's much better, right? My intuition suggests that this gadget will be a significant success, right? But we still might feel that that doesn't quite striking the right tone. Now here is where you might want to add some knobs. So this is not just corrections. This one is quite flexible because you can now bring in some of the knobs we looked at earlier in the course, or perhaps introduce a template to have a specific style. Maybe I want to use conversational yet professional language, so I don't want to use flowery language. I want to use conversational, yet professional. So there, it might be a little bit more snappy, it might be more youthful, right? This is a little formal. And although this is pretty natural as a sentence, I don't think I would say it this way, right? So this is where we get down to now fine tuning it so that we can still achieve a specific style, a specific overall tone, right? Although again, it is much, much better. Alright, so let's see what we get in terms of the list of revisions, improvements, and corrections changed in formal language? I want to I would like to okay. Revised sentences for better clarity and flow? Yes. Added more professional language to convey the idea. Okay, So that's clear. More professional language, definitely there. Again, if we wanted it to be maybe not professional, still casual, but naturally would have to specify that in the original prompt. Let's see what else addressed the lack of research and added plans to conduct research, improve the sentence structure for discussing the price range, clarified, the clarified that manufacturing and marketing strategies will be developed later, okay, so you can see that this is the next level of improvements to something you may want to write. We go from corrections, simple corrections, grammar and spelling too. Now, a full revision with much more in-depth improvements rather than just Corrections. And I want to get it across that this is not use this or use that. I want to tell you to try to find the right way for what you need to do. Whether that means a combination between the first and the second method we've looked at here. Or whether that means using this e.g. and then introducing some of the other concepts, some of the other ideas, some of the other approaches that we've worked on throughout the course. Again, bringing it back to the idea that everything you use, every, let's say, step or process that you use to create a proposal e.g. should be unique to that process, to what you specifically need. So just as an example, what might this look like if we wanted to create another revision of this proposal using some knobs. Let's explore that next. 25. High-level Feedback and Revisions: We looked at how we can improve writing by summoning a proofreader who cannot just correct spelling and grammar, but also improves syntax, improves style, make it better. But we talked about bringing in e.g. knobs or something else that we might use to make that revision even better, e.g. to match a particular tone or style. Let's explore that possibility. Let's see if we can use our knobs approach to improve the first revision that we did on this very poorly written proposal. So we have this very poorly written proposal. We asked for the proofreader and we got a much better proposal. But still, I'm not wild about. My intuition suggests that this gadget will be a significant success. I'm not wild about that sentence. I see that there are a lot of corrections that have been made. This is still useful, right? But I want to see if I can make it really good, right? I want to take it to the next level without even going into another chat. I'm going to get some feedback on this. So give me general feedback on how to improve my proposal considering that I want to accomplish the following. So notice, I'm not yet asking for a revision. I want to get feedback first. Maybe I want to stop it. Feedback. Maybe I just want feedback. Okay, fair enough. Great. Again, maybe you're trying to improve your writing and you want to get feedback based on a certain criteria. We're going to look at that next. Okay, so what do I want to accomplish on now? I introduced something that should be very familiar. We've talked about this, the knobs, tone informative, engaging, professional style, use common and concise language. Use the style of a modern proposal temperature is 0.8. Purpose to persuade. Alright, so I'm just asking for feedback here. Now. I might actually get a revision. Sometimes the chat CPT goes overboard and gives me the revision. But regardless, we want the feedback. Okay, So let's, let's get the feedback here to improve your proposal while maintaining an informative, engaging in professional tone, consider the following suggestions. Begin with a strong introduction. Clearly state the purpose of your proposal and briefly summarize the main points. So we're getting proposal feedback here in addition to language feedback. Again, whatever knobs we choose, That's the power of knobs. Call it whatever you want. But that's the power is. I can specify the kind of things that I want to have. And I can either get feedback on those things or I can write something based on those things, or I can improve something based on those things. Use concise language, empty, use short sentences and simple words, avoiding jargon or overly technical terms, focus on the benefits. Clearly, explain how your product or idea will benefit your, your target audience. The target audience include evidence and research, okay? So this is good feedback. Let's say I just want to integrate it right away. I want to see, now, can chat you PT, then take that feedback and create a revision? Let's just ask why don't we integrate the feedback above into my proposal as simple as that, because we've already specified the things we're looking for, right? So here's what we get. I'm pleased to present this proposal for your consideration, which outlines and innovative concept that has the potential to transform our target market. The proposed product is a versatile gadget designed to perform various tasks, meeting the diverse needs of its users. While the specific features have yet to be determined, we're confident that the combination of brainstorming and market research will result in a product that is both functional and appealing to customers. Going down through the sections. And you can see this is much longer because it's integrating that specific feedback. While the details of manufacturing and marketing have not yet been explored at this stage, we plan to develop comprehensive strategies for these aspects as the project progresses. Now, clearly the style, the quality of writing of this is way better than the original. Think back to the original, think back to the second version. We had some awkward phrasing in the second version that is no longer here, because again, we've fine-tuned it. Now I want to just note one perhaps obvious thing is, this is really vague. It's not being clear about which specific product it's talking about. There's no there's no detail here. Well, that's because of a concept we talked about very early in the course. Garbage in, garbage out. Well, you get out what you put in. The original didn't have a lot of specific details. And so this doesn't have a lot of specific details because garbage in, garbage out here, not garbage in terms of the quality of writing. Garbage in terms of the content. It's, yeah, it's fluff, right? It's placeholder content, but it is better. It is more clear, even though it is very vague. What I want to highlight is that we've gone from correcting simple grammar and spelling mistakes, which has its merits on its own, to a full rewrite that has included feedback, improvements in the specific things that we've defined. And now a complete revision based on feedback that we've asked for in specific detail, fine tuned with knobs that we created after getting the first revision, right. So going through this step in an iterative way is first of all, very customizable. And second, a process of thinking about what you need, right? What do I need out of chat GPT in this situation? Now, what steps should I use to get what I need? And so I'm not trying to outline a particular path for you to follow. I'm trying to show you the range of what's possible for feedback, for revisions, for corrections on things like spelling and grammar, right? And with a sense of what's possible now, I want you to start building your own flows. Start building your own process based on specific situations, based on what you need to create, right? So we've covered this fairly comprehensible range of what's possible in terms of Corrections, feedback, and improvements, revisions. But there is another piece here. And that is when we want to bring in an outside standard. And this is where it gets, I think, very interesting. So let's go on to that next 26. Checking Writing Against a Standard: When it comes to writing well or what is considered good writing. There are many schools of thought, many methodologies and approaches. One, perhaps one of the most famous, is a book written 100 years ago, I believe more than 100 years ago, called The Elements of Style. And this established a kind of framework or approach to writing in a concise way, in a very clear way to avoid misunderstandings. And this really got sort of imbued in what we now call, I think, modern writing. So whether you subscribe to this or not, I want to use it to show you one way to get feedback on your writing. This is to show chat GPT, a standard detailed tell it to learn that using the template approach, basically the template approach. And then ask it to give you feedback. Even a score based on that standard, based on the breakdown of that. So in order to do this, well, in order to do this in a way that's useful, it has to be pretty specific. First you introduce the criteria, ideally fairly detailed. And if there are rules about how things are scored, that should be introduced as well. And then you provide the sample that you want to be, let's say, judged on or given feedback on. So here's the one I'm going to use. I want you to learn the scoring criteria for the elements of style. If you understand the criteria, say, I understand, don't say anything else. Very simple. Template style, prompt, right? Scoring criteria. So we have clarity and then some specific details under clarity, grammar and some specific criteria under grammar style, under style, conciseness, coherence, organization. And each of these notice has a score one-to-five. To use this rubric, read the piece of writing carefully and score each criterion on the one-to-five scale, then calculate the average score for each category and the overall average score. So this is just the first piece. That's all we want to see. I understand. Now. We're going to we're going to ask for feedback and suggestions and we're going to show it something very familiar, something that you've seen before. It's something that we've kind of used as an example throughout the course, the classic history of English essay. This is an original version that was used as a script for the video. Okay, so this is actually not a version that came from chat GPT. This is, this is from the original script. And what I want is feedback and suggestions for improvement based on the specific criteria and scores, focusing on areas with the lowest scores. So I don't really want to break down or details if it's a five, but if there's anything that needs to be improved and hopefully I get a pretty good score, right? I want to know more about what's not as good, right? So that I can really get better. Be sure to include a full score breakdown. Now, when you're crafting this, you really want to think about what you want information or feedback about specifically. So you might not include that focus on areas with the lowest scores if you don't want to bias it in that way. So think about things in terms of biasing. So you bias it one way if that's kinda what I want more of, or leave it unbiased if you want it to be neutral and simply give you feedback on everything, right? Or maybe the highest scores or whatever. Okay? Score breakdown, clarity for grammar, for style three, conciseness, three, coherence for organization. Three. Organization three makes sense because it was just a fragment, right? But now let's see, let's see what we can get in terms of feedback on, on each category. Overall, the language is clear and easily understood and cut concrete language is used to improve further, avoid asking rhetorical questions and provide a more straightforward explanation instead. Again, this is just a fragment and so, you know, it is what it is. Your grammar is mostly correct, but there is a minor error in this sentence. Just like animals languages go fade and extinct. Okay, I think I actually remember that when I was originally reading the script, that my mind just jumped over that and I corrected it in my head. Interesting, consider revising it too. Just like animals languages fade and become extinct. Interesting. Style three. Oh dear, the tone and voice are inconsistent as it shifts between conversational and educational, maintain a more consistent tone throughout the text. Additionally, consider using more engaging and appealing word Choices, conciseness, some parts of your texts can be more concise, e.g. and I loved that it has specific examples here. You can revise the sentence. Can I stop here and remind everyone that languages change and evolve over time to languages change and evolve over time. Okay. Now again, it doesn't know what my goal was. There was a spoken script, so allow me to defend myself because it was originally written as a script to be read. And so I had a particular style that I was going for. And this is judging me according to writing best practices. And so they're not exactly the same thing, which is why I used this example. I wanted to see how we could measure up to writing best practices with a, with a script for a video. The flow and progression of ideas are generally logical. To further enhance coherence, consider using more transitions and connecting words. So we're getting very specific feedback, right? We're getting very useful feedback for this approach. And again, include the score thing or not. Set this up in a way that's useful for you. Whether that means you using The Elements of Style or develop your own criteria by which you want to be evaluated to make sure you're sticking to a particular voice or tone, to make sure your writing is consistent, to make sure you're staying on track. And again, you can use this just for feedback or you can use it then to improve your writing by then going to that next step, like we did in the previous example, to say now integrate that feedback into my original version. So depending on what you're trying to accomplish here, now, in the original, in the original prompt here we say provide feedback and suggestions for improvement. So every time I'm getting feedback and suggestions for improvement, to improve, right? Each time it's saying how I can do that. If I want it to be a little bit more detailed, I might say something like make sure to cite at least two specific examples for each point that I should improve on or something to that effect and play around with that wording so that you get those results, that you get what you're going for. But the key to this is thinking about number one, what kind of criteria can you set up as a rule? And then how can you make sure that you're getting the feedback that's actually useful so that you can either improve your writing or learn more about something or get a sense for where you are currently or whatever it is you're trying to achieve. This is a process. I think it's really a way of thinking and it's actually very powerful. Especially if you start thinking about what other types of scoring criteria there might be that you could use to check your writing or to check your English, e.g. if you were to paste in a transcript from yourself speaking, using a rubric or scoring system from an exam, right? That sort of thing. And so that gets into a very interesting territory where it becomes almost like a teacher there with you giving you consistent feedback. So play around with this and let me know how it goes. In the next and final lesson of the course, we're just going to wrap things up and I'm going to leave you with a few things to keep in mind as you continue your journey, improving your writing and using tools like chat GPT. So I'll see you in the next lesson. 27. Course Wrap-up: Hey, welcome to the end of the course. Great job. Congratulations. I hope you've enjoyed the journey thus far. I hope you've learned a lot, and I hope that you're applying what you have learned throughout this course. In this last lesson, I just like to do a bit of a recap to talk through what we've covered in this course. And leave you with a few thoughts, a few things to bear in mind as you continue your journey through this vast and fascinating landscape of creating stuff with ai after going through some core principles we focused on are three main approaches, outlines and knobs, building from templates and the Star Trek computer. Again, not that each of those names is important, but that they each represent a unique style approach to creating something. In the case of outlines and knobs, perhaps something more in-depth, something a little bit longer where we want to really fine tune the final result. In the case of building from templates that we would be able to teach chat GPT, something that it can use as a template, then create something else. And then from the Star Trek computer to have a highly iterative process for creating something like a cover letter. And then continuing to hone are fine tune that. But even more importantly to not see these necessarily as three clearly separate approaches, but to understand how they work and how they might be combined as needed according to the situation that you find yourself in, right? According to what you need to accomplish to build up that intuition of, yeah, I could take the knobs idea and apply it here, but I still want to use a template to create your own processes, to create your own custom Prompt library that you can continue to iterate and improve over time. That's going to be that thing which allows you to actually become very efficient, to actually be productive without sacrificing quality, which is what we're really trying to solve here. Can we get the efficiency? Can we get the productivity without sacrificing quality? And once you build up that intuition and you start creating your own library, your own approaches. That's what you get again, if you do it correctly. So we focused on all of those examples to build up that intuition, to then start building that library. And then we got into some advanced tips and tricks. We've talked about using transcripts from videos and how we can re-purpose from those transcripts to other types of content, to other formats. How powerful that can be. Really, that is just the tip of the iceberg. I think that's important to keep in mind. There are a lot of possibilities there and we just touched on it. And really it's the same with the automation part. We looked at using Zapier to automate sending messages, creating messages, and how we can link apps together using chat GBT or the chat GPT OpenAI. But again, tip of the iceberg. There are so many possibilities there that we really didn't go into a lot of depth on. But so many possibilities if you start working with it. And then we talked about, we looked at how we can build out that bespoke prompt library before getting into ideation. How can we harness that almost childish confidence that chat GPT has to just throw out an idea to never draw a blank to help us get over that blank page syndrome. And then once we have a few ideas that we like, then we can start developing it by creating a conversation partner, maybe a co-founder, a devil's advocate, someone who can ask questions one at a time to help flesh out the idea or find flaws in the idea to refine it so that we can create something really good without having to create it first and then decide it's not worth it, right? Then we talked about corrections and feedback, starting with basic corrections, spelling and grammar. And then going all the way to high level feedback on something like a proposal. Then taking that high level feedback after tweaking it by bringing in some knobs. And again, that's just one example. One thing that you might do as part of this process of improving your writing, taking that and then coming up with a revision, integrating that feedback into a much better version, right? Then we looked at how we can compare e.g. writing or something else against a standard, against a rubric. And once again, that is the tip of the iceberg. If you start working with that, finding other ways to integrate rubrics and standards, there is so much you can do. There's so many possibilities there that I encourage you to explore. So that's what we covered. And I just want you to keep in mind the importance of building a radar so that you don't get fixed in a certain way of doing things right? I hope that I've communicated this throughout the course that I'm trying to give you a approaches, but I don't want these to be things that you just copy and paste and use, but did you work on and develop so that they're actually useful to you, right? But continued doing that beyond the course, join communities, discord groups so that you can continue to learn about best practices, new ideas. This is a very rapidly evolving area. And so things are always changing very quickly. So keep an eye out, right? Try out new stuff. Don't be afraid to explore. Don't be afraid to change something. If you find a better way of doing it, to always be optimizing, continue to hone what's working for you. If something, you start getting generic results when you're trying to create something, break it, see if you can find a better way. Continue to hone what's working and use that radar to find new ideas, right? From others, from communities that you can bring into your processes, that you can add to your prompts. And I think that's part of maintaining a sort of bespoke mindset of nobody else's solution. Nobody else's approach is going to be exactly what you need. It has to fit what you're trying to accomplish. And so you're the curator. You're the one who has to decide what's working and what's not. And so it's your responsibility to customize, come up with your own way of doing things and continue to learn. So good luck as you continue to explore and iterate and hone. And most importantly, create great stuff. If you haven't already done so, feel free to leave a review. Let me know how you felt about the course. That would be most appreciated. And you can check out my other courses if you like as well. I have a lot of them. You can see those on my profile. And again, good luck. I hope to see you in the next course.