EP 24: How To *Actually* Use ChatGPT

This is the Everyday AI show, the everyday podcast where we simplify AI and bring its power to your fingertips. Listen daily for practical advice to boost your career, business and everyday life.

Will Chat GPT be your new Google? That's one of the things that we're going to be talking about today on Everyday AI. This is your daily podcast, live stream and newsletter to help you learn everything that's going on in the world of AI and how to actually leverage those things that you're learning. So thank you for tuning in, whether you are joining us live on LinkedIn, some other social media, or you're listening to this on the podcast. So thank you for tuning in. So a little schedule change today we're supposed to talk a little bit about hiring, but instead we're going to talk about how to actually use ChatGPT. Very excited about this. So if you've listened on the show before, I've told you, I've tried to replace Google normal Google searches with Chat GPT and that's one of the things that we're going to be talking about today. So let's, before we get into that, let's talk about what's going on in the world of AI news. Because there's a lot, as with every day, you know, unless you're spending three, four, five hours a day, it is extremely hard to keep up with what's going on and how this affects our daily lives. Because whether you know it or not, you know, whether you are a content creator, a lawyer or a doctor, AI is currently impacting your industry. So we're spending so much time every day to tell you how. So let's jump into it and look at what's happening in AI news. So Google has officially launched a new search engine. That's right. The same Google that you know and love and that has been through us, through the ups and the downs of, of the past decades is changing. So Google first announced this a couple weeks ago at their conference, but they are officially starting to roll it out today. Today. So Google has launched what's called Search Generative Experience, or sge. So what that is, it is an experimental version of the future of search, really. And what it's going to do is it's going to start to blend together kind of traditional search results with more of an AI experience. So they did kind of preview this at their event. We will show you that preview in the newsletter and we'll also have a link to sign up. So you have to, you have to get on the wait list and you have to also enable this. So make sure you go to your everydayai.com Sign up for the newsletter. We're going to have those links and those instructions on if you want to start using this. Sge how to go ahead and do that. So make sure to check that out. We have two pieces of OpenAI and ChatGPT news. So one of the biggest things, and I've actually received a lot of Messages on link LinkedIn from people in Europe saying, hey, Jordan, show sounds cool and all, but I can't really use any of these tips and tactics that you're sharing because we don't have access to chat GPT or in some cases, you know, Google's new AI kind of chat experience called Bard. So a lot of countries in Europe, because the EU largely is regulating AI technology a little more, getting access to these tools is not as easy as it is for people like myself who live in the US so OpenAI CEO Sam Altman met with European politicians discussing like, hey, how can we make this work? How can we make sure your privacy concerns are met but still be able to use chat GPT in these countries? So what came of it is, you know, there's kind of some, some threatening, some back and forth like, all right, well, if you don't want us here, we'll leave. We won't come here. But it sounds like they're probably going to be working things out and they're, they're talking about something called the, the EU AI Act. A lot of Alphabet soup there. So the EU AI act, and it could be the first set of rules to globally govern AI, which is kind of crazy to think about it that way, that we could have our first kind of set of, you know, international rules that tells us, you know, how we're going to be using AI, you know, ethically, morally, responsibly, so many things. So definitely keep, keep an eye on that. Another piece of chat GBT news is, you know, obviously to no surprise, the ChatGPT, the official ChatGPT app in the App Store, quickly became one of the most popular and most downloaded ever. It's been out for just under a week and there's already been half a million downloads of the platform. So we did talk about this the day it was released, personally, and I'd love to hear you. If you're tuning in. Drop me a comment, let me know what you think. Personally, I'm not a huge fan of the ChatGPT app on the iPhone just yet. A couple reasons at least the last time I checked, which I think was yesterday. It still doesn't have access to plugins and it doesn't have access to browsing. So what that means. So a lot of people when they're using ChatGPT, they may or may not know that kind of its knowledge base ends at September 2021. So if, whether you're using it for researching purposes or writing purposes or whatever it may be, the cutoff for most large language models that use OpenAI's GPT is September 2021. So a lot of times people are getting inaccurate information or they're just getting met with these kind of error messages when they're using ChatGPT. So that's one of the reasons, personally I haven't used this yet very much, because you can use the normal Chat GPT on a browser, on your phone and get much more functionality. So at least for me personally, not a big fan of the Chat GPT app. If you are, let us know how you're using it. Love to talk. Okay, one more piece of news before we jump back into Chat GPT. I don't want this whole thing to be, you know, chat GPT, you know, it seems like it's already sometimes running our lives. So let's talk about something, something fun and unexpected. I did not see this coming, even though it seems obvious now that the writing is on the wall. So James Cameron, the famed director of the Terminator films, recently confirmed, yes, they are working on the script for the new Terminator and yes, it is going to be less of a sci fi film and it's going to really be based on what's going on right now with AI. So to me that's extremely fascinating because, you know, the Terminator series, the franchise first debuted about 30 years ago in the mid-80s and at that time, you know, when you have machines walking around with, with this super intelligence, it was a sci fi film, right? And now you look back at it and it's not sci fi. That's kind of today's daily life, right? So we've shared about this in our newsletter a couple times. A lot of big companies, you know, Tesla, there's another one, Phoenix, I believe was the of their humanoid robots. But you have these. Boston Dynamics is another one. But you have these walking, talking robots that can do everyday tasks that can process language. So the old Terminators from 30 years ago, they don't seem too science fiction, at least from today's aspect. So it is going to be interesting to see what this new Terminator looks like and it is going to be based much more on where AI is at today. All right, so with that, you all know I get, I get thirsty a lot. I'm going to take a, take a five second break and we're going to talk how to actually use Chat GPT. So let's get into this.

Are you still running in circles trying to figure out how to actually grow.

Your business with AI?

Maybe your company has been tinkering with large language models for a year or more, but can't really get traction to find ROI on gen AI. Hey, this is Jordan Wilson, host of this very podcast. Companies like Adobe, Microsoft and Nvidia have partnered with us because they trust our expertise in educating the masses around generative AI to get ahead. And some of the most innovative companies in the country hire us to help with their AI strategy and to train hundreds of their employees on how to use Gen AI. So whether you're looking for ChatGPT training for thousands or just need help building your front end AI strategy, you, you can partner with us too. Just like some of the biggest companies.

In the world do.

Go to your everydayai.com partner to get in contact with our team or you can just click on the partner section of our website. We'll help you stop running in those AI circles and help get your team ahead and build a straight path to ROI on gen.

And again, if you are watching live, feel free to drop a comment on anything that we're talking about and I'd love to, to go over it a little more. Whether it's one of the topics that we talked about in the news roundup or if you have questions on how to actually use AI, go ahead and drop a comment. So let's talk. Okay, so one of the biggest misconceptions about ChatGPT and just, you know, speaking with other AI chats. So you know, your Google Bard and your Bing Chat, those are kind of the three big ones. So you know, you have your product from Microsoft, Bing Chat, Google bard and then OpenAI's chat GPT. One of the biggest misconceptions, aside from what we just Talked about, the September 2021 cutoff, is that you can't just say what you want and expect it on the back end. Right? There's, as a former journalist, there was a saying that we used called garbage in, garbage out. So so many people, especially when this technology first came out and there weren't any courses on prompting and you know, every other post on Twitter and LinkedIn wasn't, you know, these 32 prompts are going to save Your life before all of that, people were kind of disappointed with, with these AI chats with Chat GPT. They're saying, all right, well look, it's a failure that' if you put garbage in, you will get garbage out. So that's what I want to talk about today. And if you give me another five to ten minutes of your time, I will guarantee you that you will be getting much better results on the back end again, whether you're using Chat GPT or anything else. So let's talk about the process that, you know, our team at Everyday AI uses when, when we are using Chat GPT or, or any of these others. So we use a three step process and if you like this, let me know, reply in the newsletter, let me know and we'll dive more into it. So it's called Prime Prompt Polish. So three P's. So remember this the next time that you're gonna log on to chat GPT to, to do a search. Prime, prompt, polish, three Ps. Let's dive in. Prime. Okay. This is the biggest mistakes that, that people are making. So many people just jump in and they just expect, you know, these AI chats to be able to deliver whatever they want. But you have to prime, you have to assign a role to the chatbot. I know that sounds weird, right? But these AI chat platforms are trained on, I think, billions of websites and so much information that we cannot even comprehend. So it is, they have the ability to play any role that we tell them, but we have to prime them and we have to say as an example, you are a world class marketer. So let's say you want some marketing copy, right? You don't just say give me marketing copy for this product. Let's use Everyday AI as an example. So if I wanted an intro to the show or if I wanted ideas on how to grow the Everyday AI podcast and I wanted to ask Chat gbt, I would prime it. I would say you are a world class marketer, copywriter and researcher. You've watched, listen and analyze the top 100 podcasts in the US on technology. These are some of the ones that you've spent the most time on. And I would give Chat GPT not just the names, but I would give them the website as well. Because if you have browsing enabled, it can go check out those websites. And if you give it very specifics in this priming phase, again, if, if you're using the right tools, it will go ahead and read those, read those web pages, it'll do research on its Own before you even give it a prompt. So I always, in the priming process, I assign it a role. I get very specific. I give examples. Here's, here's the trick. I hope you're, I hope you're listening to this one. After you prime at the end, say, what other questions do you have before we get started? Right? That's huge. That's huge because a lot of times, you know, even when you prime it. So put this at the end of your priming, say, what other questions do you have before we get started? That is the biggest tip and trick I can give you today on getting better outputs out of Chat GPT, Google Bard, Bing Chat. Ask before you start and after you've primed. So type in that, that big prime. Assign it a role, say what you want it to do, and then say, give examples and say, do you have any other questions before we get started? And I guarantee you. Which is weird, right? It's weird that they'll be smarter than us. This happens to me every day. They'll say, yes, this doesn't make sense to me. What about A, B and C? Right? So they will come back and say, hey, I'm not fully prepared yet. So that is step one, the priming process. Okay, Step two, the prompt. So this is something that we've all probably done. If you've used ChatGPT or if you've even used, you know, if you log on to, you know, Bing, you could have accidentally prompted because, you know, especially, you know, Bing, you know, Microsoft with their Bing chat is starting to blur the lines between traditional search and chat. So maybe you've even, you know, prompted without knowing it. But the prompt, once you've gone through the priming and answered any questions, the prompt should be very specific. You need to give it a task and say exactly what you want. So you don't just say, analyze this. You would say, analyze this and give it to me in a table form with four columns. Column one should be this, column two should be this, column three should be this. Right. Etc. So make sure you, in your prompt, you say as specifically as possible what you actually want. After, after priming, what do you actually want, right? So Whether you want 10 headline ideas, whether you want a 600 word article, if you want a vegan recipe. Right? Whatever you specifically want, you have to be as pinpoint accurate as possible, okay? Including what you want it as, right? So if you're using Chat GPT to help code, you know, don't just say, give me the code, you say Give this to me in HTML you say, give this to me in css, give this to me in JSON, in markup. Right? Whatever language. So be very specific in the prompting process. Okay? Third, Definitely not. This is last but not least, after you've primed, after you've prompt, make sure to polish. Okay, so what does that mean? Even if you do a great job in the priming and you do a very specific prompt with examples, chances are you're not going to get a perfect product on the back end on the first try, right? Does anyone ever get anything perfect on the first try? Absolutely not. And neither will, you know, Chat GPT even if you do a good job setting it up. Okay, so how do we polish? Well, it's pretty easy. You have to put your kind of your, your editor hat on or your teacher's hat on, and you have to say what didn't go, right? What, what could be better. Right? And I actually have a couple, a couple tips here. Okay, These are, these are next level tips, but we're going to do it anyways. One of the most overlooked parts of this process, and this is where the polishing comes into play, is after you get your first response, if you're not sure how to make it better, talk to Chat GPT like a human, right? So let's say you wanted a table to help you analyze your competitors. And you ask for certain things and some were good and some weren't right, and you don't know how to make the prompt better. You can say a couple ways you can, you know, polish that. Number one is you can say, hey, this column doesn't look like what I wanted. Could you please redo it and make this column more blank? Okay, that's one way to polish. Another way, kind of, kind of little cheat code. Be specific in what didn't work and then ask, how should I adjust my prompt to get these desired results right? So let's say you asked for Chat GPT to analyze your competitors and it didn't. It looked more regionally versus locally. Let's just use that as an example. So you say, hey, these are more regional competitors. I'd like more local competitors such as A, B and C. How can I adjust this prompt, the original prompt, to make sure to get these results right? So in the polishing process, once you've been using it for a while, it will become second nature to you. But one of the ways that you improved your, your prompting skills and your ability to use Chat GPT is to ask, hey, these results fell short in this way, this way, and this way. How can I improve this prompt to get better results? Right. All right, I have two more tips for you. So one, one quick, quick coffee break here again, please. If you have questions, let's go over it. I'm here. I'm working for you. Okay. Another huge oversight. So aside from people not realizing that kind of chat GPTs or large language models, a lot of time in general, they have a cutoff dates, right? So September 2021. The other thing is people aren't using kind of the tabs in Chat GPT to their fullest potential. What do you mean, Jordan? Well, every single search that you make in Chat GPT is saved in the left hand column. And it's a little different for Bard and for Bing Chat, but everything is saved, Right? So a lot of times when we're doing these repetitive tasks, if you're using ChatGPT, a biggest mistake that I see people make is they just start a new chat, you know, whether it's the next day, later in the day, a week later. Right. But if you're going through this process of priming, prompting, and publishing, it's, it's, it's a little bit of work. Right? But why would you redo it all? You know, if you're using, you know, this every week for a specific task that you're trying to automate a mundane research project that you don't want to do, put in the time to set it up once and then make sure you go back to that same chat, it is saved. So what I like to do is I like to label them. So Chat GPT as an example, we'll give it a label based on the question, I label it something else. You can go in, click the little edit button and call it something. Here's why I do it. And this is more of an advanced tip. So if, if you're new, just bear with me, it'll make sense. So in Chat GPT specifically, you have to choose a model, and then in that chat, if you want to refer back to it later, you can't change the model. So, you know, if you're on the free version of chat GPT, the default is what's called, you know, 3.5. So that's not as it's fast, but it's, it's not that as good as the. What's called GPT4. That's the golden standard here. So GPT4 has the best base GPT4 model. And then there's what's called GPT4 with plugins and then GPT4 with browsing. Okay, we talked about this in a, in a previous episode. We'll make sure to link that in the newsletter today. But once you choose. So let's say you want GPT plugins which allows you to essentially talk and have third party services work for you. So you know, let's say open table to do your reservations kayak to book your trip. You know, there's so many, so many now third party plugins that will work specifically for, for you. But if you choose plugins, you can't now say hey, chat GPT, look at this website because that's browsing. So I label, hey, this is research and I usually say browsing or research plugins. Right? So you might have a similar chat but you might want to split it into two so you can use plugins on one and browsing on the other. So what I like to do is copy and paste the same thing, put it in each chat. So if I need to reference it in the future, it is caught up. It's caught up and the platform knows everything. So that is the last kind of pro tip is make sure that you are labeling and saving your chats in the left hand sidebar and going back to those. This is great. Think of what are some of the most mundane processes in your day to day, in your week to week. Right? So it could be data entry, it could be, you know, just analyzing stats. It can be doing a lot of competitive research. Whatever it is, think what are those, those parts of your day to day job that in theory ChatGPT can automate. Because here's the thing, you might not think it can do your job or some of your job, but it definitely can. Professor Muhammad just left a comment here. Using chat GPT to write hybrid papers. Virtual human like chat GPT and real human author himself. Exactly. That's such a great way is, is, you know, using AI or chat GPT as a copilot. Here's, here's the other thing. Until you have a master level of working with ChatGPT, until you've used it for hundreds of hours, you know, it's, it's, it's never going to be as good as a human. But don't get me wrong, once you've used, especially once the plugins and the browsing, if you've used that for a couple hundred hours now, it's pretty much human level. I'd hate to say this guys, I was a journalist for six or seven years. Yeah, it can research and write better than me. So once you learn how to use it, it can get to that level. But I do agree if you aren't using this, you know, day to day, it is much better to have that hybrid model that Professor Mohammed just talked about of Chat GPT plus human. So I hope you all found this helpful and useful. So again, just to recap, I really want you the next time you prompt and let me know how it went, go through a three step process, prime, prompt and polish and drop me a note, let me know did that work? Was it better? All right, so as a reminder, you know, we kind of talked about the difference between, you know, free chat GPT and premium. So if you do want access to the plugins and the browsing, you do have to pay 20amonth. Or instead you could go to your everydayai.com sign up for our newsletter and we're going to tell you how to win a year of Chat GPT plus so you don't have to pay for it. We'll pay for it for you, so make sure to go check that out. Thank you for a couple of comments that came in. Thank you for watching. Thank you for listening on the podcast. So we are on your favorite podcast platforms as well. Apple, Spotify, Google, there's all these other ones that I've never heard of. We're on all of those. So maybe, you know, 7:30am Central Time every morning doesn't work out for you in the live stream. I know we, we have a lot of people listening on the podcast. Thank you for that. So make sure check that out. And hey, I won't be back tomorrow. Today's Friday, but I hope to see you Monday and every day after that on Everyday AI. Thank you guys.

And that's a wrap for today's edition of Everyday AI. Thanks for joining us. If you enjoyed this episode, please subscribe and leave us a rating. It helps keep us going for a little more AI magic. Visit your everyday AI.com and sign up to our daily newsletter so you don't get left behind. Go break some barriers and we'll see you next time.