Episode Transcript
[00:00:00] Foreign.
[00:00:05] Welcome to ChatGPT Curious, a podcast for people who are, well, curious about ChatGPT. I'm, um, your host, Dr. Shantae Cofield, also known as the Maestro, and I created this show to explore what ChatGPT actually is really, though, are the files in the computer, how to use it, and what it might mean for how we think, work, create, and move through life. Whether you're skeptical, intrigued, or already experimenting, you're in the right place. All that I ask is that you stay curious. All right, let's get into it.
[00:00:38] Hello, hello, hello, my curious people, and welcome to episode, uh, six of ChatGpt Curious. I am your grateful host, the Maestro, and today I'm giving you an introduction to using ChatGPT for content creation. So a few updates about ChatGPT before we even head into the main topic for today. First off, we are back with all the models. Literally all of them, right? So much for model unification. If you don't know what I'm talking about.
[00:01:05] Uh, ChatGPT, when all the ChatGPT, when it rolled out, when Open AI rolled out, Chat GPT 5, the whole goal was model unification. You don't have to choose what model you're using, it'll just pick for you. People complained, and then they rolled it back and they said, okay, we'll bring back, uh, chatgpt4o because it's friendlier. That's what really people really wanted.
[00:01:25] And now I went in there today and all of them are back, literally, like, all of the legacy models. So at the top of the screen, you can choose the model you use. You can use five.
[00:01:36] Uh, you can use. There's different types for five. And then you go. And there's another dropdown for legacy models. And, like, all of the ones from before are back in there. So who knows what will actually be there by the time that this episode airs, because I record a week in advance. So today is August. The day that I'm recording it, it's August 20th. This will come out August, whatever the fuck that date is. August 28th.
[00:02:00] Uh, which means everything could have changed by then because that's a whole eight days from now. And literally things change every five minutes, apparently. So there's that. The second ChatGPT update is that, yes, there were recently two ChatGPT outages. Um, the first one, I actually didn't notice. It didn't affect me. We know when there's outages, not everyone gets affected by things. But the first one I was using CHAT GPT and Lex was like, oh, I read that there was an outage and I was like, oh, um, I had no idea it was working fine. But today, again, the day I'm recording it, August 20th, it's Wednesday, um, there was an outage and I, I have no idea why. I went to try and find out why before recording this episode and no one was talking about that. So I don't know. But I do think that maybe perhaps, uh, it could be because they added like a five dollar tier in India. Something like that's very low priced, um, as like their, as like an intermediary tier because like $20 too much, $0, you get throttled. So they added this smaller, uh, you know, smaller, lower priced tier.
[00:03:02] Um, so maybe it's because of that, but either way, where my brain goes is that I could foresee them doing this in the US right? I could see open it, foresee open AI adding this lower tier. We see it across the board like subscription models where they're like, here's the ad free one and then like here's the sometimes add one and here's like the sometimes maybe add one. And you just have all these different pricing, you know, um, options on this. So I wouldn't be surprised if at some point they roll that out. Uh, I would actually love. If you're listening to this, hit me back. You can shoot me a text. I do have a phone, uh, line for. Don't call me. I'm gonna say phone, don't call me. I have a separate line for my online, uh, business side of things. You can text me 310-737-2345. You could text me, you can, you can email me. I don't really want an email, but you could go to the website and contact me that way. You could send me a DM at the Movement Maestro.
[00:03:57] Um, or ChatGPT. Curious. Literally you have a million ways to get in touch with me. Uh, but let me know if you are a paid user or a free user.
[00:04:06] Maybe after this episode you might become a paid user. But I'd like to know because I do think that most people are free users and I'm wondering how they're going to close the gap on that.
[00:04:15] Um, and how they're gonna make money because they're not making money right now. Maybe I will do an episode about this. I um, read a lot of stuff about them and I'm like, they are not making money.
[00:04:25] Um, so I'm wondering if at some point they're gonna be like, hey, we're taken away though. Like the free Tiers are so bad and so throttled that you're like, okay, well, I'll pay $5 to not have that. So I don't know. Um, but yeah, I'd love to know if you're using the free tier or the paid tier. I think a lot of people are using the free tier. Actually, I use the paid tier. The, the plus tier is fine. It's $20 a month. But would love to know from you. So there is something definitely to be said about our dependence or developing dependence on ChatGPT and them being able to use that to generate revenue. Right. They can take away that free tier at some point, which I, I do worry about.
[00:05:01] Um, or they make the, they make the free tier really shitty. But we will see. Time will tell. But speaking of this dependence, the only reason that I knew about the outage was because I was at the store trying to compare elderberry syrup labels using CHAT GPT or trying to use CHAT GPT.
[00:05:16] I took a picture of it and I was like, hey, what are the difference between these? Like, I'm reading the back of the labels, they look the same by the different prices. Like, why does this one seem so much more potent but the other ones don't say, like, the actual, like, concentration. Like, what should I be looking at? Which one should I pick here? Because Lex is sick and she likes elderberry, helps her out. Um, but it didn't work. So, you know, I put the, I put it in and put it into ChatGPT and nothing happened. It was just like, you know, the equivalent of the circle of Death, the Wheel of Death. So I decided to just copy the whole prompt and try a different AI system. In this case, Gemini went to Google. And the cool thing is I just get that same exact prompt and I put it in. And Google did give me an answer with its AI summary tool, right? So it uses Gemini. And the problem to me is that the answer it gave is not as in depth, nor was it as specific. And I couldn't just like, look at the picture that I put. Like, I actually, I uploaded a picture as well, um, with. When I was using chat, trying to use Chat GPT. Um, but there is something to be said about the fact that it, that I could put that, that natural language prompt into Google, right? And I will say that the, the output is just not as good, right? So I, I did run the query later once I was already home, ChatGPT was working again. And like, just compared the two and it's just like night and day with what Chatgpt spit out versus what Gemini gave me, which was like just a basic, like overview summary. It was still good, but not like, didn't answer my question specifically. So two things that I actually want to point out with this is that one, you can use natural language to ask Google a question. That's huge, right? And I think that that in general is what is part of what is contributing to the rapid growth of LLM use IT usage. Right? ChatGPT usage. Right. You can talk to it like a human instead of a robot. And if you try and like think of the specific keywords and how do I phrase this? You could just like put anything in to an LLM and this kind of idea of if you're talking to someone and you go, you know what I mean? The LLM does sometimes, not all the times, but sometimes, uh, most of the time it does. And that's huge.
[00:07:17] That is huge.
[00:07:21] The second point, the second thing I want to point out here is how and why I think LLMs, um, will disrupt search. So that's part of the reason I think it's going to disrupt what we see as like, search in general and Google and things like that, because you can use natural language.
[00:07:35] But the second thing is that people want the answer to their questions. They don't want a list of links. I don't want a list of links. All right? This is not me saying that Chat GPT is going to cool. It was going to kill Google. Cool Google, wow. Is going to kill Google or kill search. I am not saying that. Everything evolves, everything changes. Everyone wants money, so they do whatever they can to stay relevant. That's why Google has AI embedded into it now at the top and gives you the AI summaries. Right?
[00:08:01] But the fact that people just want an answer is that is going to evolve how search engines work. And that's why people love using things like ChatGPT. I don't get a list of fucking links I gotta go through. It's just like, here is the answer. Yes, I can ask for links, I can ask for resources that I can go and check this answer against, but it gives me an answer to my specific question that is huge. So just something to think about.
[00:08:30] Let's get into the main topic of today's episode though, which is an introduction to using Chat GPT for content creation. So I titled this as an intro, mainly so that I have wiggle rooms down the road to make more episodes if as needed, that are about content creation. But this is going to be as comprehensive as I can Currently make it right. So right off the bat, let me give a little shout out to my girl, Stephanie Hine. She inspired this episode. Uh, she asked in the DMs, I put up a question box, I don't know, a week ago, whatever. And I was like, what do you guys have questions? What do you folks have questions about? What do you want to learn about with ChatGPT? And her question was about using it for content, creating content. And I was like, yes, great.
[00:09:11] So when it comes to content creation, right. When I talk about this in relation to social media, one of the things I always say, and I've always said is everything is content, because it is.
[00:09:23] I, uh, want to kind of flip that idea here and say that everything you create using ChatGPT can be considered content.
[00:09:33] And the reason I want to flip this is so that you don't think of using ChatGPT to generate content for social media as some radically different process than using it to create anything else that you've been using it to create. Right? Be it emails, lesson plans, I don't know, recipes, grocery lists, what else? Workout programs, business ideas, sales pages. It's all the same. It's all the same, right? So circling back here, using Chat GPT specifically for content creation for social media, do I think it's a good idea?
[00:10:04] Sure, if you're leaning into the strengths of the model and using it to generate content from content you've already created. Right. Uh, not just going there and be like, make a post for me. Like, about what? Based on what? Like what's your flavor? What's your tone like? No, we are, uh, you, we're going to put something in and say, based on this, can you help me create this? Right. Do I think it's a good idea if you're still editing it and putting your flavor on it. Absolutely.
[00:10:31] Right. Using Chat GPT for content creation, great idea. If you're still editing what it gives you, putting your flavor on it and making sure it sounds and feels like you.
[00:10:42] Lastly, ChatGPT, you know, using it for content creation, great idea. If you are still trusting your own taste more than anything else, a robot will never replace good taste. Okay.
[00:10:56] If you're just using ChatGPT to generate a shit ton of AI slop that you're gonna, you know, spam the interwebs with, then. No, please don't do that. It's not a good idea. Um, so let's get a little bit technical here, right? The technical how to, as it relates to using Chat GPT, to creating in this case, social media content. Now, admittedly, I haven't done a ton of this specific thing, you know, creating, creating social media in general. Like, that's like kind of my. I've done this since 2014. Like, that's, uh, a come second, you know, comes in that very natural to me. But I did play around with this just for, specifically for this episode. And I've also been toying around or toying with the idea of how I could use ChatGPT to create social media content for this podcast. And it's Instagram account, right? I. I do have an Instagram account for this podcast. It's called ChatGPT Curious Podcast. Um, that is the. I'll link that.
[00:11:54] Um, but y' all know, if you, if you've been in my ecosystem for a bit, you know, I don't like that 2 account life, but it just made more sense. And so I've been thinking like, hey, can I create content in an easy way, in a simple way, based on all the stuff that I already have? Because I'm already generating, you know, this, this podcast. Um, so what follows is what I did of what I put in there, what I prompted, what I was playing around with and how my brain was working, what I was thinking about, um, and what I would suggest to others, uh, if they are looking to use ChatGPT to create social media content. So first thing, you're gonna figure out what type of content you want to create. Is it carousels? Is it scripts for reals? Is it quote cards? Right? You could create multiple types of content, but just like, get an idea of what are you trying to create, right? And yes, you can chat with ChatGPT about this and be like, hey, what kind of content could I make? This is what I'm thinking. This is what I want to do. This is my feelings, whatever, ask it, have a conversation.
[00:12:52] From there, I suggest that you look to repurpose one type of content into another, right? So we're truly like, generating content we're not just creating out of thin air, right? We are looking to repurpose one type of content into another. This way you have a solid input to give ChatGPT, and it can generate something as opposed to just, again, making something out of thin air. So an example, you take a blog that you've written and you ask it to make a carousel for Instagram.
[00:13:19] If you don't know what carousel is, folks, it's like the. A post that has multiple slides on it. It's typically going to be. You can have a combination of video and just, you know, written. But in this case, we're thinking written, right? You go m from a blog to a carousel on Instagram. Or maybe you're going to go from a podcast and turn that. You want to turn that into a reel in some way. Or maybe you have a YouTube video transcript and you want to turn that into a carousel. Or maybe you have a presentation and you want to turn that into quote cards, right? Either way, understanding that we are taking something that you've already made and looking to repurpose it into a different type of content for social media, that is, in my opinion, the best way to go about it. Or you take old social media content and look to repurpose it into something else, right? Into another type of social media content that also works. Um, I know that we are talking about social media, right? But just another example would be taking a transcript from something and turning that into a blog post. You know, I basically do that for the podcast. I take the outlines, very, very comprehensive outlines for this podcast. And I have ChatGPT create my curious companion newsletter from those outlines. Like, yes, I have to edit it and make it mine and, like, you know, correct things. But, like, it is a just, you know, it gets me, like, I don't know, 85% of of the way there. So it's great.
[00:14:33] All right, so from there, right, that's step two. First is to figure out what type of content you want to create. Then two is look to repurpose one type of content into another. Step three is we're going to do what's called reverse prompting, or which is, in my opinion, the best way to get the best outcome. So this requires you to create the initial input material, which you already have, and the final output material. So in this case, if we're talking about, you know, going from a blog post to a carousel on Instagram, you'd have to create the blog post and you'd have to create the content or the copy that you'd want on each slide of the carousel. Yes, you could use ChatGPT to create that final output material. So in this case, the content that you want to be on the carousel, you can use, you know, use ChatGPT to help you out with that. But you still got to edit it and tweak it so it's exactly what you want. So it's like, yeah, if I was going to post this, this is exactly what would go on Instagram.
[00:15:28] Then you put both of those. You paste both of those things, right? The blog post and the Carousel content or in this case the input material and the output material into Chat GPT and you're going to ask it and you're going to say to it, hey, I'm going to paste X and Y in this case, that would be your input material and your output material. I'm going to paste the blog and the carousel cop copy. Please analyze both and generate a prompt that I can use consistently and reliably to get you to create Y. In this case, the output when I input X in this case. So that would be, I want you to, wow, get excited. In this case would be reliably get you to create the carousel when I input the blog post. Right. So I just said X and Y and Y and X so that you can put, you know, substitute whatever it is that you're working on. So one more time I'm going to paste X and Y, please analyze both and generate a prompt that I can use to consistently and reliably get you to create Y when I input X, then explain what makes it work and how I can adjust if needed.
[00:16:33] And voila, you, uh, got yourself a prompt that you can use over and over and over again. Yes. You're still going to have to tweak things at the end. You're still going to have to edit things and make it sound like you, but it will get you that most of the way. Right? You've got a repeatable system for creating a phenomenal starting point for content.
[00:16:50] For those of you that are listening are like, can I just like make this, like repeat this? I have to paste that every time. So number one, definitely paste that prompt into and save it in like your folders.
[00:17:00] Uh, excuse me, your, your notes. If uh, you're on iPhone, your notes. If you don't have an iPhone, go get one or you know, whatever you use for Droid. So like, I think Google keep is an option either way. Just put it somewhere that you can always have easy access to it. Right. But uh, other thing that you can do, and this requires the paid version of ChatGPT, uh, is one of my favorite features of ChatGPT called projects. So again, the paid version plus is fine. And uh, this is why I said earlier in the episode, I was like, maybe by the end of this episode or after listening to this, you might want to upgrade. Um, but projects, you can think of projects like folders, they are great for separating and organizing things that you are working on. One of the features of Projects is that you can input and save specific instructions. So you know, I have A bunch of projects. And one of them is for this podcast. I have one for, um, one for the newsletter that I send out this. The. My. For my, my Maestro mail. I have one for my mafia. I have a project for my Maestro musings, my other paid blog. Like, I have a lot of folders in here and it. That's just the stuff I'm going to be talking about with relation to that subject goes in that folder. But the cool part are the instructions. So when you go to create a new project, there's like a little, there's like the prompt box, but underneath it it says files and then next to it it says instructions. And you can input and save instructions in there so that every time you do a specific action within that, that project, it does the thing that you want it to do. So you could take that prompt that you have, right? And you could paste it into the instructions and say, hey, whenever I, uh, submit, whenever I, you know, paste a blog in here, please follow the, you know, do the instructions in the. In the following prompt.
[00:18:54] Amazing. And so now all you have to do is open that project. You will drop in your blog post or whatever it is, and it will, you'll hit enter and then it'll give you whatever output that you asked for. Right.
[00:19:09] Projects to me are a really cool feature and you know, they probably weren't an episode. And like me talking about the other parts of, of chatgpt that I like, the technical side, but I'm thinking maybe you can even hear it in this description I just gave you. It's really difficult to describe these things via, uh, a podcast. Like it's, I can describe it in two seconds when people can see it. And yes, I know if this was a video podcast, which is not because that is much harder I know do with my other, my other podcast.
[00:19:37] Uh, you know, if I could do it with the video side of things, much easier. But what uh, I'm going to suggest is go into ChatGPT again. It has to be the paid version and play around with projects. You'll be able to see it on the left side in that kind of navigation bar that's, that's on the left side of this, of the screen. One of them says new project. There's a little folder and it has a plus. Click it, play around. Right.
[00:19:59] Uh, but that is an introduction on how I use ChatGPT to create content. Right. So real quickly how I recently used Chat GPT and then we'll wrap it up. So this is actually a little bit of A throwback, but it ties into my story of early story or wow. My story from earlier today about trying to have ChatGPT help me in the store with the elderberry syrup. So, um, when Lex and I were in Gstad, Switzerland last month, uh, a lot of the stuff in the grocery store is in German and I cannot read German. Uh, so I took a picture and it was a package of meat because we're going to make a shark Gucci board. And I said, what does this say? Because I was like, what type of meat is this? And, uh, hit me back with an answer.
[00:20:41] So the skeptics in the. In the audience, I love this, uh, if you remember from episode one about the fact that LLMs are probabilistic. Like, yes, LLMs are probabilistic and they generate answers based on the patterns and data that they were trained on. But for things that are well established and well, and well represented in, in training, the models will almost always produce a. The fact correctly. Right. What will vary and leads to this idea of it being probabilistic is the style of delivery or the, like, the degree of elaboration. Right. So translation, a general conversation with simple words in a major language. You know, German. I'm just asking for what it says on this, like, very, like normal type of meat. Um, it's gonna. But there's good, very good, very high accuracy, like 90, 95 accuracy that it will translate it correctly for me. Right.
[00:21:38] Of note, translation and learning a language was one of the things that they demoed with the GPT5 rollout. So there's a lot of cool functionality to this thing. And yes, you are 100 correct to always be skeptical. That was last week's episode. Can I trust or should. What is it? Can you trust ChatGPT? So I love the skepticism. Um, and I also love that hopefully you understood what the I just said when I was talking about it being probabilistic.
[00:22:03] Uh, and the fact that what's going to vary then would be the style of delivery, but it will likely produce a factually correct answer with. With a high degree of certainty. Right.
[00:22:14] All right. That was exciting. I love talking about this stuff. Thank you guys for giving me this, this platform to. For listening. I hope you're listening. Maybe I don't know if you're listening, but for those of you that are, thank you. I really, really do enjoy talking about this stuff. So hopefully you found this episode helpful if you did consider leaving a rating or a review.
[00:22:34] Side note, a humongous, ginormous thank you to those of you who have left reviews already. I had no idea. Like, there's no, like, little thing that pops up and it's like you've gotten a review. I, um, actually just want to check and look at it before recording this episode and holy, there's reviews there. And just thank you, like, legitimately, folks. Thank you. That means a ton to me. I just really, y' all are the best, so thank you.
[00:23:02] Do not forget. Also, I have that companion newsletter, the Curious Companion, that drops every Thursday. And it's basically the podcast in the podcast episode in text format. So if you prefer to read or you just want a written record of things, join the newsletter fam. You can head to chat GPT curious.com forward/newsletter or check out the link in the show notes and it'll take you right there.
[00:23:28] As always, endlessly, endlessly endlessly appreciative for every single one of you, especially the folks who have left reviews and ratings. All the best.
[00:23:39] Until we chat again next Thursday, stay curious.
