Everyday AI Podcast – An AI and ChatGPT Podcast - EP 46 - Midjourney Secrets Revealed: Keyword Prompt Method

Episode Date: June 27, 2023

Wondering what's the best way to use Midjourney, an AI image generator? Today David Barona, Founder of  Prompter Mag. joins us as we dive into the best methods for using Midjourney and how it can be prompted with keywords, emojis, and story prompts.For more detail, head to our episode page.Join the conversation and ask David any question you have here! Time Stamps:00:01:17 US presidential race uses AI; Google gaming; Baidu beats ChatGPT00:03:15 About David and Prompter Mag00:04:56 Using Midjourney to create a magazine00:07:00 How Midjourney works00:11:20 Prompting Midjourney00:17:07 How you can use Midjourney in your daily life00:19:09 AI and content creation Topics Covered in This Episode:- Versatility of Midjourney- AI image generators and experimentation platform- Satisfaction of completing design projects with AI- Ownership and copyright issues- Usage of AI in corporate content creation, legal concerns, and potential benefits- "Keyword prompt method" for identifying aesthetic elements- Deepfake technology and ethical considerations- Introduction of a new magazine focused on chalk art- AI in politics and Google's new AI strategy- Baidu's AI bot Ernie Bot- Everyday AI website features, including newsletters and podcasts- Guest speaker David Barona, founder of Prompter Mag.Keywords:Midjourney, AI image generator, emojis, single words, story prompts, tag style prompts, text to image, experimentation platform, designer satisfaction, refined craftsmanship, train models, legal concerns, open conversation, keyword prompt method, chalk art, deepfake technology, lawsuits, watermarks, Adobe, magazine, US election, Google, Baidu, Ernie Bot, Everyday AI website, David Barona, Prompter Mag.Send Everyday AI and Jordan a text message. (We can't reply back unless you leave contact info)

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. Meet Firefly AI Assistant, now live in Adobe Firefly, the All In One Creative AI Studio. Just describe what you want to create and the assistant handles the rest, orchestrating multi-step workflows across Photoshop, Premiere Express, and more in one conversational interface. You direct the outcome. The assistant accelerates execution. Can you create beautiful imagery with only text in a couple of seconds?

The answer is yes, with a little bit of practice and with some tips from our guest today on everyday AI. Before we get into that, and we're going to dive into Mid Journey and how to get the most out of this text to image generating powerhouse, before we do that, let me tell you, you are listening to Everyday AI. We are a daily live stream podcast and free newsletter, helping everyday people like you and me, not just understand and keep up with what's going on in the world of AI, but how we can actually use it in our day-to-day lives. So as a reminder, if you are joining us live, please,

this is an interactive show. Drop a comment, especially if you want to know more about Mid-Journey. But before we do that, let's talk about what's going on in the world of AI news. Some pretty big developments today. So let's quickly talk about them. So the 2024 presidential rates in the U.S. here has officially been dubbed the AI election. So there's a report that we're sharing about today in the newsletter, just talking about how in-depth, you know, the political parties are going in the U.S. using AI in the election

and in their political campaigns. The article is talking about how in-depth it goes. And it's wild. The second news piece to talk about today is Google's new AI strategy is actually tapping into gaming. So an article in PC Gamer today was talking about Google's kind of plan to beat or catch up with ChatGBT. And a lot of it has to go kind of old school with their gaming strategy. So more about that in the newsletter. Speaking of people trying to catch ChatGPT, so China's Bay-Due, their new AI

bot. It's called Ernie Bot. And kind of a new story came out today with some findings saying that they're passing chat GPT on some metrics already. So if you want to know if Bay-Doo's Ernie Bot is going to catch chat GPT and we have a lot of other news, make sure that you go to your everyday AI.com and sign up for our free daily newsletter. So that's right, not just a live stream, not just a podcast, free daily newsletter. But that's not what you're here for. You're probably here to learn about Mint Journey. So let's talk about that. So I'm going to bring on my guest today. So joining the Everyday AI live show today, we have David Barona, who is the founder of Promptor Mag. David, thank you for joining. Jordan, thank you, sir. I appreciate it, man. Thank you.

Thanks for having me. All right. Cool. We already have some people hanging out in the comments, Professor Muhammad, Lillian. They're excited. Let's go. Yes, let's go. So David, just tell us a little bit what is prompter mag and how did you get into this this whole mid-jorney thing? How did I get well? First of all, Prompter Mag is my publication that I've really dedicated to learning and teaching along the way how to use this platform and really how to get the most out of it. Honestly, I like making assets that I can use. I'm a graphic designer by trade and I'm in sales. I make presentations. I have to use. graphic assets all the time. So I really tried to develop a way that I could use mid-journey to do it

because obviously I can make them in about five seconds, studio quality, incredible stuff, and getting better with every version. And it was for my own as my own tool. You know, I really wanted to make assets that I could use and make money from. I love seeing the great results, but I also really wanted to monetize it somehow. So that's where it was born, man. And that's where it comes from is from me wanting to make beautiful stuff that I can use and not have to take a month to make something incredible, you know. Yeah. Speaking of things that are incredible. So if you're joining the live stream, your eyes are going to be blessed here.

If you're listening on the podcast, check the show notes. We're going to be sharing some of this imagery. And you can come back and watch the live stream here as well. But so we're showing a prompter mag. So this is this is an unreleased edition. So we're always trying to break. Yeah, we're always trying to break news here on everyday AI. But David, just tell us a bit about how the magazine is made up and kind of what the target or the goal is with the magazine.

Sure. So yeah. And this issue, this is chalk art here. This issue is dropping in about an hour and a half or so. so you'll get to save the PDF, et cetera, and look through it yourself. Look, I was always, as a kid, I was obsessed with magazines. You know, I went to, when my parents went to the grocery store, I would sit in the magazine now and just pull up whatever I could

and just, like, absorb these images. And this was, I have to, this was inspired by these two particular magazines. One of them is called Wizard, and the other one is called Toy Fair. And these magazines are defunct now by years. But they stuck in my head. And I just, they're about comic books and they're about collectible toys. And I, as you can see in the back, I'm a, you know, I'm a sneaker head. I'm a collectibles guy that's always been my thing.

I like to be surrounded by like all this crazy colorful stuff all the time. And that's sort of how that magazine looks too, you know. It sort of reflects what is around me and how I kind of. like to visually sort of take everything in. So super image heavy. And I, you know, forgive me to all those guys who hate when somebody splatters images over text. But I'm like, you know, I'm like a David Carson guy. I don't know if anybody here, you know, knows Ray Gunn and, and someone who just cares about things looking beautiful and not necessarily the perfect rule follower for design. I was always kind of a rule breaker and I like I like destructive design. I know that sounds crazy,

but I do. And yeah, that in a nutshell. Yeah. So let's quickly, actually before we dive in deeper, I do want to hit rewind here because not everyone listening or watching has used Mid Journey. So David, in the simplest way possible, explain kind of what Mid Journey is and how it works. Well, Mid Journey is an AI image generator, basically. And the way that this thing works, it's, you know, let's say you took an image and you shredded that image. Then you're dividing an image into many different but equal parts. You can pull those pieces apart and you can actually create multiple images that look

similar but are built from the same image. This is the basic concept. It just starts pulling. multiple options out of one or many other options. And these new options didn't exist before. And so in the end, it's just generating and predicting data, just like Chat GBT DBT does, except this time it's visual. You can go simple.

You could literally prompt, as I actually posted about not too long ago, you can prompt with an emoji, which is a little uncontrollable, but you can still do that. You can prompt with a single word. You can go with a story prompt where it's really kind of like a. a spoken a, a, anecdote almost.

You can also go, I do a lot of, my brain works in like an SEO type of a way. I'm like, you know, tag. So I do a lot of tag style prompts where I'm just sort of lining up terms

and separating them with commas to sort of organize it in my head. And yeah, it's, it's, to me, it's almost like an experimentation platform. It's,

it's so, I don't want to say user-friendly, because it may not be quite yet soon, hopefully. But it's so, like, intuitive when you wrap your head around verbal, you know, text to image. That's it, man. I mean, that's it. Yeah. And the intuitiveness, I think, has gotten better.

And, you know, if you are listening and this is catching your ear, you're like, hey, what is this? So, you know, kind of like what David said, it is just text to image. So think of chat GPT. but instead of text, you get beautiful imagery on the other side. So OpenAI has their Dolly kind of technology. There's stable diffusion. But Mid Journey is, I think, one of the more popular options. And it's, you know, as you can see here on the screen, if you're tuning in, it creates some amazing results.

David, the first time you use this with, you know, having a design background, obviously the quality has improved. But what was your thoughts when you could take? text and get beautiful images out. It was, it still is to this day. So, you know, as a designer,

we get really satisfied when we complete a project and it looks beautiful and you know you've been through all the details. And it may take you a week, two weeks a month. As an artist, it takes you potentially years to complete a piece. And you get this satisfaction where you kind of get this rush, right?

At least that happens to me. And I, with mid-journey, that satisfaction starts coming so, fast that I get, it's almost like I get a rush, and it's addicting, to be honest, with you. I get addicted and I have to remember how many, like, fast hours I have because otherwise I could just go down a rabbit hole of just getting this rush of like, ooh, look at that.

And like, literally, I feel like I get adrenaline as the image generating. I'm like, what's it going to be? So it's been life-changing, man. I got to tell you, it's definitely changed the way that I, I, I, I always said that AI is not going to replace a refined craftsman, but it's going to change the way that we refine our craft. And you can just tell from the second you start using this stuff. Yeah.

And just to follow up on David's point in the free daily newsletter today, so go to your everyday AI.com. We have done a couple quick five-minute tutorials on Mid Journey, covering some of the things that David's talking about here. But we do have a question from Audrey, and that kind of brings us to the whole topic of this episode. So saying, you know, how do you source the group of text? Is it through experiments or online threads? So, yeah, David, talk a little bit about that, the keyword prompt method that you're using to get these images that we are seeing on screen.

Audrey, great question. And for me and what you're talking about are where I call my keywords, you know, and a lot of people use that same term. but I called this the keyword prompt method because what I wanted to do is I wanted to figure out how to call out a certain aesthetic related to an art form. In this case, let's talk about chalk art, right? And so I'm not familiar with chalk art. I'm not a chalk artist. There is a whole set of technical terms and there's a vernacular around the art of chalk

masterpieces. And I don't know that stuff, right? And so as an artist and as a prompt engineer, I don't have enough of an arsenal to really be able to call out the aesthetic. So that's where the keyword prompt method comes in. And to answer your question, Audrey, I did all of this via few shot learning with chat GPT and a couple other tools like that. So I built a set of instructions and then I call it out. Now, you can't do things with AI without double checking your work. So I go through whenever I get my answers back, I have to go through an experiment with every single keyword.

Make sure it's not calling out some random stuff because I realized that if I, for example, if I had put chalk masterpiece in here or if I put chalk art too far up in the prompt or if I if it's too, if it's on its own, it starts calling. calling out a lot of pavement. It starts calling out a lot of sidewalks. And I didn't want that. I don't want the experiential images right now. I really want to show the art format. It's base. I want to show the blackboard, basically.

So that's kind of, hopefully that gives you an insight into how I built it. Fushot learning is extremely powerful. Get to know how AI really kind of works and how you can train these models. and then you'll get a lot more out of it. Yeah, that's a great point. And with these, especially mid-journey, you know, even myself personally, that's what I found myself doing early on is, you know, changing one word at a time, you know, moving the where that word appears to see how it impacts the end result, which is great.

And as a reminder, if you are listening on the podcast, you know, we will be sharing, you know, kind of sharing some of what David's talking about here. So on the screen, you can even see the images and the exact prompt that was used to create these images. So another great question here from Yaddy. And I think this is driving a lot of what's going to be in the news in the coming months. I think there's going to be a lot of lawsuits. You know, there already have been. But great question from Yaddy saying any potential rights issues, licensing fees with Mid Journey,

we'll need to consider in the future. Yeah. Yeah, a great question. And I think this is a question for all of us. I keep literally dropping hints here and there on some of my comments. I'm like, and I think I own this stuff because that's something else that I keep coming back to is now I have, and I was just telling Jordan, I've rolled at least 10,200. Maybe I'm in the 10,500 images right now.

And do I own them? You know, are they mine? Because I've downloaded them on my computer and they have my name. I mean, are they mine? I have no idea. What I do know is that I have purposely stayed away from dropping artists' names. And anything, look, I do a lot of times I will say Iron Man. I will do Optimus Prime.

I will do characters because if you think about it, those names are very unique. Snoop Dog doesn't exist. You know Mid Journey. It's only trained on Snoop Dog, right? Like it's going to call out Snoop Dog. So sometimes I want to use those likenesses. And I'm okay with letting those ones go, right? I'm okay if later on copyright comes for those and I have to let go of all my marble

images or all my whatever images. I'm okay with it. I don't think it's going to happen that way because I can go draw one and it's probably still mine because it's my original work. But that's my suggestion. It's just try to develop images that. do not include likenesses or artists' names in your prompts. And you see many images that are rolled with multiple artists' names in the prompts.

I just don't recommend it. But rather be safe than sorry, you know. Yeah. Yaddy, great question and David, great answer. I will say this, you know, a lot of, this is going to be a hot topic, not just for mid-jorney images, but even just for Chachy PT, right? There's already been literally countless lawsuits. And I do think you have to look at how these models were originally trained.

And I think when this technology first came out, people were getting results that had watermarks from the images they were trained from. So I do think that's something to keep in mind. And speaking of like Adobe Firefly, so Adobe has their kind of own version of this. It's a little further behind. But Adobe, Yaddy, to get to your question as an example, they trained their model with all images. is that they had copyright, that they had copyright rights too. So, you know, completely different. Another question I have for you, David, is, you know, probably a lot of people listening.

Maybe they're not artists. Maybe they're not graphic designers. How can, or maybe what are some ways that the everyday person who's not an artist or who isn't required to produce, you know, pieces of art, how can they use Mid-Journey in their daily lives? Well, you know, there's so many uses. And I think it comes down to thinking about, I have a son, I have a five-year-old son, okay? And I have to think about things like he has activities.

Now we're keeping him home for the summer. And he's an artist. He loves drawing. And my wife is not an artist like me, but she needs to hold down the fort. And she needs some tools. So I introduce her to Chad GPD and I introduced her to MidGertie. And now, for example, if he needs a page to color on, if he needs an empty coloring page, I will literally go to Mid Journey and I'll create whatever it is that he wants with lines

so that it's good right for coloring. I'll make mazes for him. Then I also use it for, I guess, in your daily life, if you're not a graphic designer, maybe this doesn't apply to you. But I just use it for inspiration and moodboarding. So I don't go to Pinterest anymore. You know, I don't really go to Google anymore. I really just go to Mid Journey to get my sort of daily fill of inspiration for my sort of visual explorations every day.

But a really good one is with kids. I got to have to bring. That one's the first one that comes to mind is making stuff that he can then, you know, I can print and then use. And I don't have to go to Google and struggle to find like, the right resolution and the, you know, yeah. Yeah, that's, that's such a good point because I think there's, especially with, with kids, you know, that, you know, the example of creating a custom coloring book. I even did this, you know, with one of my nephews who drew things and trying to turn his drawing into, you know, a piece of art. So I think there's tons of use cases. So we do have one

more question here. And again, this is a great area. But what's your quick take on this one, David? So Pedro is asking, what if we want to use these images in, a corporate project. Listen, that, to be honest with you, is up to your corporate client. You have to disclose this. And I've a shout out to Genesis Inc. here in Denver. They're an ad agency that I've been close to for a long time. And right now I'm consulting with them and sort of figuring out ways to put AI in place

responsibly, intentionally, and skillfully in order to really get yeses from the clients. Right now, AI's got a lot of sizzles. So you're going to be able to, in my opinion, if you can sell it well, you should be able to even develop a whole new sort of scaled up price point for content creation. The expectations are changing from clients now that AI is introduced, faster, quicker turnaround times. And so in a sense, you really have to feel out the client because if they're, if they are full. of lawyers and they just cannot move past the question of the legal problems, then no, you can't. But if they can be like, you know what, this project has an expiration date, we're going to put out a bunch of content and it's going to live for a month and then it's going to die anyway.

Let's use it, right? Like copyright is not coming for AI images in the next three months. So let's just push this out. Right. So in a sense, I do. I use them. Yeah. But I have to tell them, listen, this is the, this is the legal question. Are you cool with it? Because if so, let's go. And we can make a lot for, for potentially a little, right? So it's got to be an open, open, open conversation, brother. Yeah. Pedro, great, great question. David, great answer. It's, as with a lot of things in AI, there's a lot of great area right now. But, you know, those those people who are, are, you know, taking advantage in the interim are the people I think who are growing, growing their companies, growing their careers, growing, you know, their departments, whatever it may be. So, wow, we touched on a little bit of everything. We got into chalk art, you know, looking at prompter mag. David, so much good insights from you. Thank you so much for joining the everyday AI show.

Absolutely, man. I'm so happy to be here and answer questions. I hope I can do it again. I hope that, you know, mid-journey is, it becomes a tool for everybody. And, you know, you don't get replaced by AI if you're using it. It's just plain and simple. So start using it. If you're a designer, don't be a stick in the mud. Don't be a purist. You know, this happened before when computers came out.

You know, I talked to creative directors who were like, when computers came out, I didn't want to change. And then I realized it could make 10 things in the same amount of time that I could make one. You know, so don't be a stick in the mud. It'll just hurt you, you know, get into it, jump into it and do it now because the sizzle is still there. I believe it gets commoditized soon enough. Just enjoy it. Yeah, right. Love it.

Great advice from someone who is currently riding the wave. So thank you again. And as a reminder, if you haven't already, go to your everyday AI.com. Make sure to sign up for our free daily newsletter. We're dropping it at 11 a.m. Central Standard Time every single day. And we're going to have a lot more. If you like some of the things that David said, if you want to know more, we're going to be sharing all those things. So thank you for tuning in.

And we hope to see you back tomorrow and every day with Everyday AI. Thank you. Meet Firefly AI Assistant. Now live in Adobe Firefly, the Allman One Creative AI Studio. Just describe what you want to create in your. your own words and the assistant handles the rest, orchestrating multi-step workflows across Adobe Creative Cloud apps, including Photoshop, Premiere Express, and more in one conversational interface. You direct the outcome while the assistant accelerates execution.

Stand control with the ability to step in and refine at any time. See it today at firefly.adobie.com. 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.