SHOW TRANSCRIPT

This transcript is automatically generated by Descript.  Any errors or omissions are unintentional.

Jeff Sieh: Hello, folks. Welcome to Social Media News Live. I’m Jeff Sieh. And you’re not.

Conor Brown: And I’m Connor Brown, and this is the show that keeps you up to date on what’s happening in the world of social media and more.

Jeff Sieh: Have you ever wondered what it takes to create mind blowing AI generated images? Maybe you’re eager to discover the techniques behind crafting visuals that captivate and engage your audience. Or maybe you’re looking to bring your creative ideas to [00:01:00] life using the power of AI. If those questions resonate with you, then you are in the right place.

Today we’re excited to welcome a guest. Who’s done so much in this area, Mr. Dustin Stout, the visionary behind Magi. And he has taken his passion for, for creativity and AI to new heights. He’ll be sharing what he’s learned, his insights and top strategies for mastering AI creation. So sit back, clear your schedule, clear your mind and get ready for this week’s episode of Social Media News Live.

Dustin, thank you so much for coming back on the show today.

Dustin Stout: Thank you for being who you are and for consistently, no really, for consistently delivering such amazing value. I show up every week that I can and I just, I love watching the conversations. I love how you care so much about creators and the people who are doing the actual work. Um, and uh, I, I think it’s just an honor to be here, it’s an honor to be your friend, uh, I love you, um, and I’m just grateful to be here.

Jeff Sieh: Yeah, Dustin, [00:02:00] and we’ve hung out at conferences before, he’s, I mean, and he’s such a, he’s such a, a, he’s such a great designer, he cares about UI, so, uh, I just, I’m, I am proud to call him friend, so, uh, Connor, if people don’t know Dustin, tell them about him.

Conor Brown: I know you might know Dustin from the chat, but he’s one heck of a guy as well. I’m a big fan, so I’m so excited to have him on the show today. Dustin Stout is an entrepreneur, AI enthusiast, and founder at Magi. Since starting his blog in 2011, Dustin has built a successful career as a full time digital marketing consultant, speaker, And has created numerous products that help digital creators accomplish more, including his current focus, Magi, which brings together the world’s best AI tools all in one place.

Thanks again, Dustin, for being here.

Dustin Stout: Exactly as Magi wrote it. Thank you, Connor.

Jeff Sieh: There you go. Uh, as always, this show is brought to you by our folks and our friends over at Ecamm. You can find out more [00:03:00] about them at ecamm. com forward slash Jeff. Use code JEFF15 to save 15 percent on your first purchase. And once again, folks, don’t forget, stick around to the end, because we’re going to have the drawing for that free Creator Camp ticket.

You can, uh, still enter for your chance to win at jeffsieh. com forward slash

win. It’s like 700 value, so like it’s nothing to sneeze at. And by the way, we’ve got our friends already popping in. Uh, Gary, I’m surprised that he’s here. I saw that he was doing some sort of event for his company today.

Gary, if you’re, uh, uh, you know, burning the candle at both ends, I appreciate you, my friend. And Dustin has brought his, uh, fan club from over on YouTube. Uh, Dustin is the man from Camellio and,

Dustin Stout: Hey, Camero.

Jeff Sieh: and, uh, my friend, uh, Chris Stone, his episode is coming out that he just had with us. In a couple of days. Um, yeah, butt crack, new drinking game.

I don’t know if I drink to that game, uh, there, Chris, but he also

put,

Dustin Stout: most fun

Jeff Sieh: yeah. And he also, yes, I do. Jeff does it for the free consulting. See, he’s in, he knows my strategy. Uh, don’t let the beard fool you. So, okay. So [00:04:00] speaking of that, um, the free strategy, uh, this Dustin, I think, so I’m getting ready to speak at Momentum in about two weeks.

Uh, the last year was the first time. I’ve ever been able to create completely custom slides. You know, this from being a presenter is that we’ve always had to try, like to find stock images and try to force them into kind of what we want to do. And it was the first time that I was, I was so excited because they were all mine.

They were all original that I made with AI. And it felt like as a creator. I was finally able to take what was in my head and put it out to make the point that I wanted to bring across. And it was super exciting. Have you, have you found that as well? Because it feels like that to me that we’re in that kind of age that

Dustin Stout: Oh,

absolutely.

Jeff Sieh: So talk about that a little bit.

Dustin Stout: I obviously have a bit of a, an eye for design and am a bit picky. Um, and so over the years, you know, AI images just have not been quite there. There’s always, [00:05:00] you know, it’s, it’s the, the typical things. You sit down, you have this great idea, this, this visual scene that you’ve built in your head.

And you type out the magic words and you hit the magic button and it comes out and it’s like everything is perfect except like that one thing. There’s that one thing in the middle of the screen that’s just like, oh, if I could just fix that one thing and it can never fix that one thing for you the way you want it to be fixed.

Um, but now, these image generators have gotten so good. And as long as you have a few principles and you’re prompting,

Conor Brown: so

Dustin Stout: typically get exactly what you want. Now text is, you know, the big revelation of recent iterations.

Conor Brown: joining us.

Dustin Stout: It’s one thing.

For most presenters, they want to put text on it, too. Up until recently, you wouldn’t be able to do that. But now, uh, I’m anxious to explore that for a presentation I’m doing next week in Dallas. I’ve, I’ve got my talk all [00:06:00] ready, um, but I haven’t done the slides yet because I know it’s going to take me 57 hours to, to make 14 slides.

Um, uh, but yeah, I’m excited to give it a try and see if the text can actually work. Now, the trouble is my, a lot of my talk is going to be, Pulling from pop culture, movies, AI and

Jeff Sieh: Mm hmm. Mm

Dustin Stout: going to talk heavily about Terminator 2, which was a big influence on my life. Um, so I’m going to be drawing from a lot of like real life visuals as well.

So, uh, it should be interesting. I won’t have to make everything out of AI, but I will have a few things in there.

Conor Brown: Well, I saw Jeff’s slides last year at Momentum and they were awesome. So it’s not like it was just an ease of use sort of thing. But I think what was also really, really cool is they came out. And once you see them, you’re like, that’s totally Jeff. Like that is, you know, so that’s what I think is really, really powerful.

It’s your own creativity within you and you can put that on the screen as well. Yeah.

Dustin Stout: Yes. Yeah. Your thoughts, your ideas, your style

Jeff Sieh: Yeah.

and then if you have any skills, um, you can mix them [00:07:00] together, like, I’m really, and we’ll talk about some of the, the different, you know, strengths and weaknesses from the different platforms, you know, text has been an issue, uh, with one of my favorite ones that I use all the time is MidJourney, but, uh, Ideogram is great.

I feel it’s, it’s the, the king of the, what, you know, the text generation. Yeah. If you want to have that with your,

Dustin Stout: really

Jeff Sieh: opinion. But, um, but we’ll talk, but first of all, I want to, I want to go back a little bit because, by the way, if you do, um, want to check out Magi, well, you can go to magi. co.

Justin has made a special deal for like it’s 30 percent off for the first three months. So guys go check that out if you haven’t already. And I’ll be plugging that all day because it is affiliate. But it’s an amazing platform. It is the best. The quote is on there right at the top. I think it’s the best bang for your buck if you’re trying to get in and decide and even learn AI image generation.

It is the best bang for your buck. So Magi. co. com forward slash Jeff. All right. [00:08:00] So, Justin. What inspired you to create Magi, and what is, makes it different from all the other competitors out there? Mm

Dustin Stout: Yeah. So when. When ChatGPT first hit the scene, I had already been working with AI. I had been using, you know, things like Jasper or copy AI. Like there was a million different AI copywriting tools that had hit the scene before ChatGPT was even a word people heard about. Um, but they all sucked. Right? Like, they were all pretty bad, and it’s because they were all predicated on, uh, basically a, uh, you know, an in out exchange.

You input some things to a predefined form or a predefined template. That somebody else made and it sends it to AI fills out that template that it’s been given, and hopefully it matches what your intended goal was. Many times it wasn’t very good, or it just didn’t quite match what you’re trying to accomplish.

You had to kind of massage it a lot to really get something [00:09:00] that was usable. But when ChatGPT hit the screen, the screens. And the scene. Uh, the paradigm was completely shattered. No longer did you have to go through, you know, search through a thousand templates and put your idea through somebody else’s lens.

You can now just talk to the AI. Just like you would a human, and it would give you exactly what you wanted, and if it wasn’t exactly what you wanted, you could talk with it some more, and it would make it better. Um, and that was mind blowing to me, and this is what caught, this is why AI caught on, because the barrier to entry, the, the, uh, ease of use was so natural that anybody could do it.

And anybody can start playing with it. And so, um, I was super excited about it at the time. I was still taking some clients. I was just, accepting the fact that the last two years of my life, building two other products, two other SaaS products was just, Utter failure. I, I was accepting the fact that these products were not gonna take off, but the way that I wanted, um, I was at a very low place.

I was very depressed. And, uh, just [00:10:00] not knowing what else I have to give the world because, um, if you’ve ever been an entrepreneur or founder, you’ve been in that place where like, all I see is failure in my, uh, rear view mirror. Uh, what else can I do? And it was a, just one of those moments, like all us entrepreneurs have.

We, we find something that, that is solving a problem. Right. And maybe we see some gaps we see like, Oh, but why doesn’t it do this? Or why is this missing? And I wish I could do this with that. So naturally the, the entrepreneur spirit was like, ah, I wish I could make a better way to use this and so like with ChatGPT in particular, I wish I could organize these chats.

I wish I could search and filter through them to find the thing that I made last week that was so good and now I can’t remember it. Um, I wish I could put these into folders. I wish I could share some of this easily with my team. I wish I could collaborate with my team. And so all of these little quality of life things, mostly business and professional work oriented stuff, workflows that I was used to, um, they, they just weren’t super easy with ChatGPT.

And it’s [00:11:00] in its current form back then. Um, even now, some of those things aren’t solved. So I was like, you know what? I heard that, that ChatGPT has an API, which means other third party apps can sort of build their own version and utilize the same technology. So I decided I had no resources. I had no funds to pay developers.

I just thought I have nothing else to do except to try and build this myself. And so I decided to build the, the app that, that would help me do more work faster and easier with, with AI. And, um, it really kind of ballooned from there. You know, once, once you, once you get going and you realize, okay, Oh, shoot.

There are other things out there. Maybe I can tie in some of these other AI tools as well and bring it all in one place so people don’t have to play subscription ping pong going from one subscription to the other, back and forth. They could just have them all in one place and I can make that possible.

Jeff Sieh: Yeah, that, I mean, that’s one of the geniuses and make is it, and you make it so easy. One of the things that I think is really cool, and we’ll get into the, well, I’m gonna [00:12:00] wait. I’m gonna save my thing. Connor, I know you had a question.

Conor Brown: Well, I know I just love being able to, because I’m a Magi user, being able to kind of put in the same prompt, but then ask different models the same thing and kind of pick and choose from what I’m getting from different models. And kind of almost Frankenstein this thing together. It’s kinda like you’re getting the best of all of these worlds.

So, Dustin, you know, magi has so many different AI models, whether it’s the conversational type, like, you know, chat, GPT, Claude, um, uh, Gemini Lama, those sorts of things. But also from a. Uh, an image perspective, you know, there’s Dolly, there’s Stable Diffusion, Flux, I know just got added, which is really, really cool.

So can you kind of explain just a little bit more for people that aren’t familiar with Magi, like how these models are integrated into the platform and how you utilize them?

Dustin Stout: Yeah, again, going back to that idea of, Like I realized that there was an arms race about to happen right at the beginning. I knew that, you know, every big tech company was going to create their own [00:13:00] AI model and want to kind of dominate the space. And then you had image generators popping up, video generators, audio generators, and they were all going to be these separate and, uh, siloed off tools.

And I wanted to bring them all in one place. Now, thankfully, like I said, Many of these tools have what’s called an API, which essentially means a developer can take that, that technology and bring it right into their own applications. And so I just, uh, you know, being a UI UX guy, I’m a little bit obsessed with, with design and making beautiful, enjoyable apps.

Um, I, I just found a way to kind of wrangle them together and put them all in one place. And so, That, that, uh, idea that not only do you get to only have to pay for one subscription instead of paying for multiple ones, but you also get to explore the different tools in the same interface, which, which is really important because experimentation is probably the best way of learning because these, these AI.

Models and tools. They’re very complex and very intimidating for a lot of [00:14:00] people. Um, still like they’re, you know, as prolific as chat GPT is in the, you know, the, the conversation everywhere, there are still millions of people who have no idea what to do with AI. And, uh, you know, let alone being able to go to all the different apps, explore how this one does it, what this one does, and that one does.

I, I really set out to create a space that was easy, enjoyable to use, and allowed you to sort of explore freely the different available options that were out there.

Jeff Sieh: Yeah. So one of the things, and I want to, uh, I’ll bring this up real quick is that, and I love what, you know, this is on the landing page. So if you go once again to this magi. co forward slash Jeff, you know, I said, if you wanted to start getting used to getting into AI, Magi is the best deal around. And I really think that is true, but Dustin, what are some other ways?

That, um, if people are new to AI image generation, why does Magi make it a great starting place?

Dustin Stout: Well, mostly because [00:15:00] you get multiple tools in one place. Now, um, that’s the biggest thing is because, you know, Each AI model is better at certain things, you know, just like different humans have different skill sets, you know, some people are better at, you know, designing apps, you know, some people are better growing beards, um, some people are, you know, some people are better at different things, all these AI models, they have their own personalities, and they have their own strengths, their own weaknesses.

And so being able to explore them in the same space, really It helps cut that learning curve, I think. Now, on top of that, we have also built some tools into Magi that make it super easy. Now, a lot of these AI tools out there, uh, God bless them, they’re built by developers. And, kudos to all my developers out there.

I love you, developers. But they have no idea what real humans are like. So, uh, they build things in a way that’s kind of difficult to understand because they understand it as they’re building. And I know this myself, because when I build things A lot of times I overlook some of the more complicated [00:16:00] stuff that normal people don’t understand.

So I’m not knocking the developers. You’re too smart for your own good. Um, so I’m, you know, I took the time to really, how can I simplify these tools and make it accessible to the average person who doesn’t know what, Inference steps means, or what guidance scale means, um, right. And what all these technical terms mean.

Um, so I, I’ve built them to be as simple, yet functional as possible. And on top of that, you know, one of the most difficult parts of image generation is prompt engineering. You know, what are the magic words I have to come up with? To get the perfect image, right? And most of us who are not designers, who are not artists, we don’t have the vocabulary that is necessary, nor do we have the understanding necessary to really paint the picture that the AI needs to take what’s in our head and put it into a visual reality.

So I built what’s called the prompt enhance button. Uh, so the idea here is you can type a simple prompt. And you can [00:17:00] hit the enhance button and that, that will trigger an AI model that I’ve trained to take your prompt and expand it and make it more descriptive so that it paints the entire scene the way that the AI can perfectly, uh, bring it to life.

And of course you can change it, you can make tweaks to it, but, uh, for those who don’t have all the, uh, you know, image prompt engineering knowledge and skills, you know, this, This is basically the easy button uh, for generating highly engineered image prompts.

Jeff Sieh: And

Conor Brown: And I love that button. It’s so great because I’m not that good at prompt engineering, you know, I’ll feed it a picture of me and I’ll say, make me look hot. And it just comes out with me on fire. So that’s not what I’m looking for.

Jeff Sieh: the rest, the rest of us are thinking that about you, but not,

Dustin Stout: Yeah. I mean, I’m, I know all the prompt engineering. I built the tool, but I still use the button. Why? Because it’s so much easier. It’s so much faster. uh, whether you’re a pro or a noob, you can really enjoy the shortcut that that, uh, prompt enhanced button offers.

Jeff Sieh: yeah, it’s very, very cool. And so we’ve got [00:18:00] some great comments. Tech Troublemaker says, Magi, it’s just plain fun, right? It’s just fun to do. Uh, and you have a track record because Dustin makes cool stuff. He says he still remembers social warfare. And to your, yeah, I use that a lot too, uh, Tech Troublemaker.

So also, um, Gary says, um, when we’re talking about diff, using different platforms, he goes, true, I found Claude to be better at some, but not all of what I need. And one of the things, because we just talked a little bit about, like, why it’s, Magi is great for image generation, is a lot of times, and this is kind of maybe some advanced prompting stuff, but it all I’ll know what I want, but I need to ask it to like, I want to get the focal length or use these prime lenses in my photo and stuff and not having to go switch to chat with GBT, but just go in.

It’s already there in Magi. And then I can, you know, paste it over into, you know, you know, creating that image that is really, really cool that it’s all baked in. Uh, and when you’re in a rush, having everything all baked in into one thing is really, really key. And it also helps you become a [00:19:00] better prompt engineer because.

You know what to ask, you’re starting to understand the, the lingo, and being able to, you know, move those across, between them is really, really key.

Dustin Stout: absolutely. Yep.

Conor Brown: Yeah, you know what I love Dustin is being able to, you know, not only follow what Magi is doing and the new features of that, but kind of your entrepreneurial journey on this. You’re great about posting about how successful it has become and what’s the next thing and those sorts of things. So I think it’s a really, really cool story.

Overall that I’ve loved following, but I also know you probably face some huge challenges in building the platform from the ground up. Challenges that are beyond just, well, what AI models am I going to put into this thing? Uh, I know you’re UX and UI kind of designer, but this takes some serious developing.

I would, I would assume in a lot of tech. And a lot of marketing and, and all this stuff. So what are some of the biggest challenges that, that you’ve faced in building Magi? Um, [00:20:00] especially when it comes to creating something for content creators.

Dustin Stout: You know, technology has a lot of challenges. I mean, the biggest challenge I think currently with AI is just keeping up, right? Like, OpenAI yesterday. I was literally about to jump into a mastermind call. And the, you know, the founder of OpenAI was like, Hey, guess what? New model. Um, and I was like, wonderful.

Um, something new, uh, which is great, but you know, keeping up with it can be hard because with new models, with new iterations come new technical challenges and things that you have to change. So I feel like I have to constantly feed into development as a, as a marketer for the last 15 years. It has been nearly impossible for me to actually do marketing for my product because I’ve been in the product every single day making improvements, making advancements, and trying to refine it and, you know, being a bit of a perfectionist doesn’t help any of that.

I’m seeking [00:21:00] help. If you know anyone, please let me know. But yeah, I mean, it’s super difficult. Now, on top of that, we’ve also had some, uh, some challenges with, with fraud, you know, AI is one of those industries that has attracted a lot of bad actors, and, uh, unfortunately, but unfortunately, it seems like Magi has gotten a bit of press and a bit of, you know, And so it’s been a big target of fraudsters and most recently, I had to unfortunately shut down the free trial because the free trial had become a target for, for abuse and was costing me a lot of money, um, in terms of racking up my API bills.

Uh, we got hit by, um, you know, what I can only imagine was a group of savvy, uh, guys. Coders who found out how to get access to our system and use it for free on an ongoing basis. And, um, um, yeah, put us in a, a bit of a financial spot for a good few months, but we were able to learn from it, fix it. Uh, thankfully I, we, MagiGri to the point back in last year in December, Where I could hire on [00:22:00] some development help.

And so my, my developers are really great. They’re really smart guys and they, they, they know their stuff and they work hard and they work fast. And so we were able to plug all the holes that the, the fraudsters found and, you know, stop them from continuing to rack up my, my ridiculous AI API bills. Um, but that did necessitate.

Having to get rid of free trial and it really sucked for me because I really want people to enjoy it and try it without having to, you know, pay for it upfront, you know, because there is a bit of a, you know, a hesitation there. I know I love a good free trial to make sure I can test it before I buy it.

Um, but the reality is that it just ended up costing way too much money for me to support. And, um, we do offer a 30 day, no questions asked refund though. And that’s kind of how we. justify like, you know, not giving you a free trial, but hey, if you don’t like it, we’ll give you your money back. I won’t even sad about it.

Jeff Sieh: So, was the, the biggest part of the fraud, uh, people getting it for free and, and racking up the, is that [00:23:00] what happened? Okay.

Dustin Stout: Yeah, yeah, they just, they found a loophole in the free trial system and managed to just, just send all kinds of stuff through our system and,

Jeff Sieh: gosh.

Dustin Stout: uh, OpenAI and Claude willingly took our money. Um,

Jeff Sieh: The scary thing is, as entrepreneurs, and I had another friend who has kind of a, a, uh, educational side of her business, and what happened was her, her Stripe account, they were, all these places were, they were running, um, stolen credit cards through them. And she was petrified that Stripe would shut down.

Her system, because that’s where she makes money. And so,

Dustin Stout: We went through that exact same thing.

Jeff Sieh: yeah, so it just rubs me the wrong way that people try that. I mean, it’s hard enough being an entrepreneur and then having to deal with that stuff on top of it. So anyway, okay.

Dustin Stout: Yeah. I mean, it would be great if we could just focus on making the product good, but instead, we have to split our time to, you know, preventing fraudsters from exploiting the system.

Jeff Sieh: So, I, I didn’t want to have, I mean, we, everybody here loves, um, [00:24:00] Magi. I didn’t want to make it a big, uh, commercial, but it is a commercial. Go to magi. co. org slash Jeff. Um, no, but it’s great. I, it really is, um, if, because the, the, the whole thing that you can get so many different, uh, uh, Uh, Uh, platforms and try them out and see what works and have the options.

Because the other thing is, they’re changing all the time, like one will get, it’s like an arms race. Like one will get better, then this will get better, and having access to, the one that gets better, is really, really cool. You’re, like you mentioned you’re not having to cancel and go and you know, wait to your month’s out, and then go to the next one.

It’s really cool having that all in one place. But, I don’t want to be a commercial all about that, because we want to talk about, Which Magi does, and it’s another great thing because you get to have access to all those different image creators. And I see in Dustin’s private Facebook group, as soon as there’s a new one, somebody asks to have it added to

Dustin Stout: magic going to happen?

Jeff Sieh: Yeah, it’s

Dustin Stout: is yes.

Jeff Sieh: yeah, and so Dustin is really on top of that. I mean, you had Flux really fast. I mean, you had Ideogram really quickly, so. It’s really cool. That’s another big [00:25:00] plus on that. You don’t have to wait around for stuff to happen. Dustin does the best job he can to get those as quickly as possible.

Um, but we’re talking about AI image creation. What are some essential tips, especially for just images, or prompt strategies that beginners can use to create those high quality AI generated images?

Dustin Stout: Yeah, no matter what image generator you’re using, typically there’s three essential pieces to your prompt that you need to include every single time. Uh, when you’re engineering your prompt, you, you need to include a subject. Whoever the subject of that image is, or whatever the subject is of that image.

This is like photography or visual design or painting one on one, right? What is the subject? Secondly, you need the setting. Where is this subject at? Are they in a dark room with studio lighting? Are they in a dystopian cityscape? Are they in a Small home office with Marvel artwork in the background. Uh, describe the [00:26:00] setting, uh, for that subject.

And the last thing that you always want to include is style. What style is this in? Is this a painting? Is it a photograph? Is it a, you know, abstract piece of, uh, 3D animation? Um, you really want to describe what style of image that you want. So, three keys, uh, write this down. Subject, Now, as long as you include those three things, you’re typically going to get a pretty good outcome.

Now, the more descriptive you can be of those things, the better. You want to use, you know, as many, um, words as necessary to describe each one of those things. And when it comes to the subject, you know, you just want to be hyper specific and the more specific you can be, the closer your outcome is going to be to what you are imagining in your head.

Now, when it comes to. Style, here’s where you, you can really start to pull in some inspiration from other places and use less words. Uh, because you can reference other styles that you’re familiar with, you know, anime, for example, [00:27:00] anime in the style of Dragon Ball Z, you know, something like that. You can reference other work, um, or, you know, in photography in particular, you, you mentioned like focal length, that, those are important details.

So one of the things that I built into our, our, Prompt engineer or a prompt enhance button was if the user uses the keyword photo realistic, we’re going to include a camera name, the name of a camera, and perhaps the lens, the focal length of the lens and any other depth of field information. So those details are super vital.

To really getting the highest quality output because these AI models have been fed all this information. So they know what a f2 focal length is, right? They know what a Nikon camera is in all the different models. And so including those details is, is always going to help enhance the outcome of the uh, end image.

So be as detailed as possible and make sure you include subject setting and style.

Jeff Sieh: One of the

Dustin Stout: And [00:28:00] that’ll work across any image. You know, I just want to make sure. That will work on any image generator. That is sort of the formula that you always want it to use, whether it’s Midjourney or Ideogram or, you know, DALI.

Now, the challenge here with DALI, I’ll just say this real quickly. Most people don’t know this about DALI, but when you throw an image prompt into DALI, it’s going to rewrite the prompt for you no matter what. So it’s going to interpret your prompt and say, I think I could do a better job at this, and then it’s going to rewrite it.

And then it’ll generate the image based on it’s rewritten prompt. Um, most times it’s pretty close to what you wrote though. So

Jeff Sieh: really quick, I’m getting ready, it’s going to come out on I think Wednesday, a episode with Stelzner’s new AI thing that I talk about, and then I’m doing another training later, Um, on Mid Journey. But one of the things that I found out the, the, the most helpful is learning artists. Like, you may, like your favorite comic book artists or whatever, and then combining them together.

And sometimes you may not even know, and it takes a little bit of [00:29:00] research. Like, I was doing one that I wanted to have, like, the Jetsons. Remember the Jetsons? Um, that’s how old I am.

Dustin Stout: we actually we have an iRobot vacuum and we named it Rosie

Jeff Sieh: Exactly.

Dustin Stout: because

Jeff Sieh: But I couldn’t, I didn’t know what that style was, and I did some research, and it’s called Googie.

It’s not Google, it’s Googie, and that’s the style. And so being able to input that architectural style into my prompt made a huge difference. And I’m able to do some really cool stuff with that. So learning those artists or those architectural styles or things like that. Even painting, painters, like if I’m not an art major or any of that stuff, but learning the different painters and the styles and then being able to combine those together is where you can get some really cool, original looking stuff.

So

Dustin Stout: yeah, one of the tricks that I use a lot of times I’m doing photorealistic stuff. Like that’s just what I tend to gravitate towards and I like to borrow from movies. I love movies and film. And so I’ll reference a movie style. https: [00:30:00] www. youtube. com or www. youtube. com or www. youtube. com https: www. youtube.

com

Jeff Sieh: Yeah.

Dustin Stout: or www. youtube. com

Blade Runner 2049, uh, and use those movie references, and it really does many times capture that, that style or that aesthetic that I’m going for.

Conor Brown: We have a lot of really good questions coming in about this. Um, so I want to take Abby’s first because I think it kind of dovetails into what I want to talk about as well. And Abby says, does the order matter? Right? So in her opinion, she gets better results when it’s at the beginning of the prompt, at least in mid journey, what she’s working for and kind of along those same lines.

How does Magi work? The interface kind of help users who are not necessarily with a technical background go through this structure. Like Dustin, you laid out exactly what you need for great image generation, but how can Magi help? And, and does the [00:31:00] order matter? Like what Abby’s saying?

Dustin Stout: It does. Well, in most image generators, it does. So the stuff at the beginning of the prompt will, in most cases, be given the highest emphasis. Uh, the stuff near the end of the prompt will, will not be given as much emphasis. You might notice a lot of, um, variety. With, uh, whatever you put at the end, but typically the stuff you put at the beginning is heavily weighted and very consistent in the image when you put it near the beginning.

So if there, you know, are crucial elements, do you wanna make sure it does not? Miss or neglect, then always try to shift those towards the beginning. But, you know, always remember if you try to emphasize everything, you end up emphasizing nothing. So really have an idea where you want that emphasis to be.

Um, and you may need to play a little bit with it. Some models, uh, like Midjourney, do tend to emphasize the beginning a little more than say a DALI, but of course, DALI is rewriting the prompt anyway, so that’s probably why it, um, You know, doesn’t quite put as much emphasis in the beginning, uh, something like, uh, [00:32:00] Leonardo’s models.

They, they tend to do very well with whatever’s at the beginning and what’s at the very end. So there’s sort of like a middle out that, whatever’s in the middle is sort of less emphasized, I’ve noticed. Um, but yeah, for, for the most part, the best practices, most important things go at the beginning, even if that’s, if, if the style of the, A particular image you want is, is super important.

You want to make sure it doesn’t miss that style. Put it at the beginning, you know, something in the style of Tim Burton, uh, blah, blah, blah, blah, blah.

Jeff Sieh: Yeah.

Dustin Stout: Uh, and, and your, your second question there, Connor, you know, how does MadGuy’s interface help with the easy button? that prompt enhance button will put that structure in place for you.

Even if you put something like here, an experiment that I love doing is I’ll just put like two keywords. In, in the, in the box and see what the AI does with it. One of my favorite is just put the words pure joy and hit that prompt enhance button and just to see what it comes up with, it’ll, it’ll put all that structure in their subject setting and style, and, um, it comes up with some pretty amazing things.

Conor Brown: [00:33:00] So I have a follow up for that because, you know, you have the prompt, but especially in Magi and the, the, the tools as well on their own platforms, um, you also have the dropdowns, right? So using Leonardo as an

Dustin Stout: yeah. Yeah.

Conor Brown: a model, you have the size and then you have the style and Leonardo kind of has these preset styles that you can select from.

But. What if you want your own specific style, you’re putting that in the prompt. If you then go and select a, a style that Leonardo also has, are that gonna kind of cancel it out or get confusing at all?

Dustin Stout: Um, it might. It really depends on the model. So one of the great things about Leonardo. ai as a company, which is recently acquired by Canva, um, Leonardo is not just one AI model. It is a bunch of fine tuned So what they’ve done is they’ve taken Stable Diffusion, which is the one of the industry’s most talked about and most advanced image generation models that’s open source, and they fine tuned it for very specific [00:34:00] styles.

And that’s when you go into Magi, select Leonardo. You’ll see that you can choose from a plethora of different, um, fine tunings of their models. And so that style will be the most predominant when you select that model. Even if you tell it to do a different style, this model is trained on a specific style and will probably lean towards that one.

Now some of the AI models inside of Magi have what’s called image to image, which means you can upload an image, And it will use that image as influence just like it uses your prompt as influence for the outcome. Um, so you can kind of partner an image and text to produce a, a new image. And that will help, it will in most cases try to adopt the style of the, the visual style of the image first before applying any other secondary styles to it.

So, um, yeah, I mean, different models have that and we do, we try to make it as easy as possible with drop downs because, you know, like, uh, with something like mid journey, you have to know all the keywords, you have to know all the magic, they have [00:35:00] a web interface now, which

I think makes it a little bit easier.

Jeff Sieh: Yeah.

Dustin Stout: Uh, yeah, it’s a super good web interface. I used it and I took some notes, um, cause it did some things really right. Um, but like previously you just had to know all the keywords. You had to know, like do the two dashes and

then do The

aspect ratio. Yeah, the, the additional parameters. And so uh, in Magi we really just try to make it stupid simple.

Uh, cause I only have so much brain power in the day and

Jeff Sieh: So,

so the question that I, I’m one, when you mentioned Leonardo, uh, talking about that, and I haven’t played with it much. I know it got acquired by Canva. Is there any worry that that will go away because they were acquired by Canva? Will they shut down the API? Or is there any, you know, is it going to stay open source for you to use as a Magi

Dustin Stout: Yeah, I’ll, uh, I’ll be a hundred percent honest. When I got that announcement, I was not happy about it.

Jeff Sieh: Yeah,

Dustin Stout: So, and here’s the reason because probably six months ago or so, Magi had, uh, two tools that were really useful. One of them was a background remover that was [00:36:00] extremely good. And the other one was a background replacer.

So you could use AI to not just remove the background, but simultaneously replace it with whatever generative item you wanted. The, uh, the company that provided this API for those two functions was acquired by one of my least favorite companies in the AI space. Uh, mentioned at some point earlier in this broadcast.

I won’t name them. Um, but they acquired this company. And within a week of acquiring the company, they jacked up the API prices like 50 times. And making it just completely unusable. Like there was no way we were going to be charged a dollar per background remove. It didn’t make any sense whatsoever. So that was the fear I had.

We, so we had to remove those from the app and find a replacement. And so it was a bit annoying and frustrating. Um, so like, that was what I thought when Canva came in now. I love Canva. I actually interviewed Melanie Perkins early in Canva’s [00:37:00] journey, uh, when, when Guy Kawasaki first came on with them, and, uh, you know, our friend Peg Fitzpatrick was, you know, helping, uh, get them off the ground with, uh, with Guy.

And so I love their journey. I love what they set out to do. But holy cow, does it scare me that they acquired one of the best

Jeff Sieh: They’re like the richest company in Australia. Yeah.

Conor Brown: Oh, yeah.

Dustin Stout: Right. Well, they say they’re going to let it operate as usual. And I’ve, I’ve had the chance to talk to some of the Leonardo team over the past year and a half or so, and they seem like really great people.

So I’m, I’m more positive, more optimistic about how it’s going to, you know, continue to be its own entity. And I think that would be the smartest move for Canva in all reality that they have a huge community behind them. And that community and their open source nature is what’s really propelling them to be one of the best sources for AI models, uh, image models, anyways.

Jeff Sieh: Yeah. One of the, and then I think it’s fascinating because it is an arms, and this is the worst it’s going to be like, right. Really? You know, it’s like, it’s, this is where we’re at now. It’s the [00:38:00] worst it’s going to be. And it’s going to continue to get, they’re going to be new models, new programs, uh, Dustin will bring those in.

I’m sure. So even if one goes, you’re like, there’s something that’s going to take its place. It’s probably going to be better. Yeah. Images by Gretchen says, SSS, I use it in so many artistic areas. I remember my high school photography teacher in 1972 talking about this. I used it when teaching writing to my six year old students, and it is a core to storytelling.

So that’s the same kind of process when it comes to image generation. Yeah. Um, one of the things that I wanted to ask you is, like, I know in Magi you can save prompts and custom personas, but, um, How can users leverage these tools to make better images?

Dustin Stout: one of the things that a lot of the artists on our platform have found is that they, they have go to prompts. That prompts that are combinations of words or styles that they use over and over again that always produce a really good result. And so they want to first and foremost, save those prompts in a place where they can easily [00:39:00] access them right inside their, their tool, rather than have to like flip screens, you know, between their notes app or Google docs and.

So we, from the beginning, like even before the, the image tools were, were super sophisticated. I wanted to save prompts just for the chat interface. I would have prompts that I would use over and over again. So we have the save prompts feature that allows you to do that in a single click. It’s copied to your clipboard so then you can paste it into uh, the image generator input and make whatever changes you need.

Um, and personas are mostly for the chat interface at this point. Um, because they, they’re essentially. OpenAI calls GPTs. They’re custom personalities or custom instructions that you give to these AI models to make them better at certain tasks. And we have one that is the image prompt engineer. That’s actually what’s powering the prompt enhanced button inside the image editor.

And we have about 40 something other custom personas inside the chat interface that users can access. But if you have [00:40:00] Specific styles that you always want to include or specific instructions that you wish, um, you could, uh, have a, an AI model inherit from you. You could create a custom persona and essentially create your own prompt engineer persona inside of Agile to, to come up with your style images, you know, based on your parameters that you want to give it and have it use every time.

Jeff Sieh: Very cool.

Conor Brown: Nice. I love that. And it’s all about making stuff easy, right? Easier for non techies, right? To utilize AI, which is great. Rich Graham asked a great question as well. Does Magi have character references so he can upload pictures of himself to generate thumbnails, etc? Um, I know some. Uh, tools do, but Dustin, overall, magi and character references.

How do those work?

Dustin Stout: Not yet. The models that we included, that we have in Magi, they, until recently, didn’t have that capability, at least not through their APIs, so we’re kind of limited to what these models allow [00:41:00] through in programmatic access. Um, they might have them in their own native interfaces, but maybe haven’t included in the API yet.

But, recently some of them have added those parameters, and I have it on my developer’s roadmap right now, to add character reference. Now the challenge here, for me, as always, is, How do I make it simple for people who don’t know what character reference means? Um, right. That’s, that’s somewhat of an advanced modality for these AI models.

And so, you know, as my developers work on the technical side of how do we implement this, I’m going to be spending a lot of time on the, the user experience side going, how can we make this It’s super intuitive and super easy to use so that you know, somebody who has never heard of character reference can also access it without any trouble.

But it is coming, it is high on our priority list because I know personally, I want to start generating images of myself with a Jeff Sieh beard. So I need it for my purposes,

Jeff Sieh: Did you see the one today for our show that we did?[00:42:00]

Dustin Stout: I did with

Jeff Sieh: Yeah, yeah. That’s a little hard to do because your mustache kept trying to get over the finger. So, Rich, one of the things that I think I know what you’re talking about, like doing thumbnails like that you see that we did today for ours and some of the other past ones.

You almost always have to use a third party app. So what, and this is from Mike Russell, and if you want to go back and watch or listen to Mike Russell, when I interviewed him, he’s got a great YouTube channel called Creator Magic, he’s also the guy who did Music Rated Creative, so he’s like super nerdy about all this stuff, and he has a free, uh, face swapping that he uses in there that’s really, really helpful.

So what you do, So, image reference is not going to get you the exact same face on your generated image. It’s just not. It won’t work that way. Um, it’ll get it close. So what I do is I create it so it’s, it makes a Dustin close. Like maybe if you kind of look sideways, it would might be Dustin. And then I use a third party software or that free one to swap the face.

I upload a face of Dustin and it’s only job [00:43:00] as a tool is to replace that face. Sometimes it works. Sometimes it doesn’t. Most of the time it’s getting pretty close now with some of the models that are there. PIXA AI is one that I’ve used before that works really well. Um, but that’s how you have to do it.

You have to get it close and then take it to another generation because none of the AI ones, not MidJourney, DALI, any of those will get you the exact same face. As far as I know. I may be wrong, Dustin, but I, I, my thing you have to use a third

Dustin Stout: Yeah, well, I think what he’s referring to in terms of character reference is a, it’s a newer image generation technology that will say if you generate an image of a, a character, um, character reference allows you

Jeff Sieh: To go

Dustin Stout: give the AI that character and say, create more images with this exact same character in it.

Now you could, in, in a sense, some of these, Uh, apps that I’ve seen, you can upload your image and you are the character and it does a fairly good job of, of maintaining your likeness across multiple image generations. And so this is, uh, a pretty advanced form of, um, not just face [00:44:00] swap, which, uh, has been around for a long time and is really good.

But this is actually teaching the AI to keep this subject. As, as identical as possible across different generations. And that’s, that is a super exciting new technology. Again, super high on my list.

Jeff Sieh: the thing that’s cool about Mid Journey is like I generated and even got Justin, Dustin’s red shirt that he always wears. And by using that character reference, I could move Dustin, the Dustin doppelganger, through multiple images of having him do the same thing. And that would keep that, like you said, The same across there, which is a big deal.

That’s been really hard for AI to do. The, the cool thing that MidJourney announced in office hours is they’re working on a model where you can upload your own image and then, you know, type in a prompt and change things, which will be really, really cool. I mean, their in painting is really good. But that’s another thing.

So I’m getting nerdy. Okay. Um, let’s talk about, um, Dustin, what are some common mistakes that users make when they’re prompt, they’re [00:45:00] crafting prompts for image generation and how they can avoid them. So like, what do you see, uh, and a little plug for Dustin’s Facebook group. He always posts the images he creates and also puts the prompt so you can learn from that.

So if you’re a Magi user and you’re not a member of that group, you need to be because he, he does some really good, uh, training and educating in there. So biggest mistakes, Dustin.

Dustin Stout: I think the biggest mistakes is, is, you know, people kind of assume things that the AI can do that it may not be capable of yet. Like at the beginning, people were asking it for text on an image, not understanding why it couldn’t do text because, you know, like it’s a computer, right? Computer’s no text. Of course, it should be able to do text, but it couldn’t do that at first.

And so like the biggest mistake was, you know, just trying to ask it for things that it couldn’t do and then getting disappointed, thinking, oh, either this thing is stupid, it doesn’t work, or I’m stupid and I just don’t know how to use it. Uh, when the reality is it just didn’t have that capability, for multiple subjects. And AI, even now, isn’t super great. At multiple [00:46:00] subjects at once, um, but has gotten much better. Like Midjourney is pretty good at it. Flux is super good at multiple subjects. Um, yeah, so, you know, asking sort of for too much at once, and again, going back to, you know, the SSS uh, technique.

Conor Brown: we hope

Dustin Stout: If you’re putting too much in there and each of those elements, sometimes it’s unable to focus on the right thing. You know, if your prompt is a mile and a half long, you got to know that not everything’s going to make it into that image. Um, and sometimes it’ll, you know, just ignore most of it. So, you know, not being descriptive enough, being, you know, Too descriptive and that, you know, everything kind of gets jumbled together and lost.

Uh, but then also just kind of knowing the limitations of the AI model, knowing that, you know, Leonardo not quite great at text yet, but Flux and Midjourney and even Dolly very good at text, you kind of play to the model that has the strengths that you need.

Jeff Sieh: [00:47:00] Yeah, that’s really true. One of the things, I’ll go back to Rich’s point with that face thing. Rich, when Dustin talked about not making it too complex is for each, cause I have Dustin, uh, Connor and myself on those images. I, I break it down and I do Dustin first and then I do, you know, Connor and then I take them into Adobe Photoshop and remove the background and then combine them inside of Photoshop for the thumbnail.

And I create the abstract background with Midjourney or whatever I’m using to create it too. So I use that to combine everything. I’m not trying to do too much. Like Dustin was saying, I was, I’m not trying to do it all in one prompt. I do try to get one thing, each element, and then combine them in a, in a different program.

So hope that clarifies a little

Dustin Stout: and that concept is actually AI modality wide. So this idea that if you want the AI to do a really good job at something, focus it. Focus it on one thing at a time so they could put all its, all You know, a [00:48:00] lot of people, the mistakes they make with large language models with chatbots is they ask it for a bunch of stuff at once, you know, write me an entire blog post, 2000 words, uh, about walking cats.

I don’t know. Um, And so the mistake there is you’re forcing the AI to spread its processing power over a big amount of work. Whereas if you would just focus it on one section at a time, maybe start with an outline and then maybe start with the intro paragraph. It has all its processing power to focus on that one small thing and it’s going to do a much, much better job at that.

The same is true with image generators. You know, if you’re having it focus on multiple subjects at a time. It’s giving, it’s got less time to focus on the details of each individual thing because it’s spread out across, you know, it’s being diluted across multiple subjects. So the more focused, the more specific, the better.

Jeff Sieh: Yeah. So real quick, uh, before we start wrapping up, gosh, this has gone so fast. This, I could talk to Dustin about this stuff all day. Uh, but something that is very [00:49:00] focused and knows what they’re doing and is in their lane is our friends over at Ecamm. They, uh, make this great program that we’re shooting this stuff out on.

You can make podcasts. I get separate tracks when I’m done with this. Uh, if you go to ecamm. com forward slash Jeff, uh, save 15 percent by using the code JEFF15 on your first purchase. they got so much cool stuff. You talk about Dustin has a great community. Ecamm’s got a great community. You notice these successful companies have a good community and, uh, I just think that there’s a lot of similarities to what Dustin is doing and what Ecamm is doing.

Um, focused, but also very, very community driven. So, uh, make sure you check both of those folks out today. Um, one of the things, oh yeah, bring up a question really quick from Deb and I think it kind of goes back to what we were talking about. And her, she wants advice on getting the featured person or animal to look the same across image to image.

Especially in Magi. So talk about if there’s a way to do that that you would tell Deb, Dustin.

Dustin Stout: Yeah, you really need something like a character reference to do a great job with that. But the models that do have image to image, you [00:50:00] can try just using that image to image functionality to say, you know, keep the subject the same but change the setting and, uh, you know, find the Find the most descriptive way to articulate the subject and it’ll be pretty much similar every time with models like Flux and models like Ideogram and not so much Dolly.

Dolly’s a little wild. Well, probably because you’re writing the prompt every time. But the models that have consistent outputs Such as Flux. Um, I would use the image to image feature to try and recreate that same subject every time until we have that character reference ability, uh, possible. Um, but again, the descriptions matter.

You know, detailed descriptions if it’s a young man. With a beard and a black rimmed glasses with long hair wearing a do rag and a red shirt. As long as you keep those same descriptive words in there, you’re gonna get a fairly similar one. Especially if you’re using like an image to image, influence.

Conor Brown: Yeah, you know, today we’ve been talking about [00:51:00] AI, artificial intelligence, not the other AI, Apple intelligence, which is just so annoying, that’s

Jeff Sieh: It’s very true.

Conor Brown: ridiculous.

Dustin Stout: Oh, I’m glad you said it.

Jeff Sieh: Oh,

Conor Brown: brings up a question, you know, Dustin, do you have any thoughts on Apple intelligence and if you think it might kind of evolve as a dominant AI right now?

He’s avoiding meta and X’s AI, but I mean, it’s something different when, you know, your phone’s right there. It’s got everything. It’s power in the palm of your hand. So what do you think about that? Let me

Dustin Stout: if anybody’s used Siri in the last 13 years, you probably don’t have huge hopes for Apple’s intelligence.

Conor Brown: Siri’s great at setting a timer. I’ll give her that.

She’s

Dustin Stout: I’ll tell you what, I’ll give her that, uh, don’t gender the robots, I’ll give it that. Um, yeah, so, I, I watched their presentation back in June, was it, at the developer conference, and, um, I was, I was both, I think it’s both optimistic and, [00:52:00] critical because a lot of the, the terminology they use, a lot of the, the way that they are marketing this technology I think is far less useful than it will actually be.

Uh, kudos to them for implementing it in a way that I think will be accessible. I think some of it is just a little bit, you know, I don’t want to get too complicated in how they executed it, but the accessibility part of it, you know, like, um, the examples they give when you’re text messaging someone and you need to find like an image or an event or have it check your calendar, like that stuff could potentially be super helpful.

I tend to focus on using AI for business purposes, for professional purposes, and having it have the full personal context of my text message history, which is, um, I don’t know how I feel about that. Um, I don’t think it would be very helpful. I actually know it will not be very helpful in professional circumstances.

Like if it’s using my text messages with my buddies that I hang out with on weekends to watch classic films, and it’s using [00:53:00] that to To pull from context wise and voice wise, when I’m writing a professional email update to my investors, I don’t want it doing that. Like, I don’t want it using the same context of, did I speak to my friends, did I speak to my investors, right?

So the, the real challenge there is going to be, You know, when people knowing when and how to use it and not getting, uh, you know, mixed signals for, um, the stuff that they do personally, the stuff they do professionally and business wise. Um, I, I hope they do a good job. Like I really do. I don’t buy I’m excited to see how they do it, but also I’m not buying most of the marketing that they’re putting out there around it.

Jeff Sieh: So As we wrap up, Dustin, what are some kind of like, maybe you can give us a, you know, open the kimono or whatever they say to tell us what’s coming up for Magi that you kind of got in the pipeline. I [00:54:00] know you’re always adding new ones that come, but is there anything else that you’re working on in the skunk works of Magi there?

Dustin Stout: Yeah, I mean, we’ve got a really aggressive roadmap. The challenge is always Time and resources, right? We’re still a small team. It’s still me, two developers and a, a customer support person and a, a virtual assistant. Um, so we’re, we’re a small ragtag group of, of people who just wanna build cool stuff. Um, so there’s a million things we wanna build.

Uh, most importantly is, you know, just making sure that we always have the best and latest technology for our users. Just yesterday, OpenAI decided they were going to announce a new model, and so now we’ve got to get that one, uh, put in there, um, but video generation is high on our list. We do have a couple video generators inside of Magi now, but we know we can, there’s more out there that we can utilize.

Um, audio generation, um, we want to do more with audio, so, uh, Trying to bring in all the different modalities. We just made a big overhaul to our chat interface so that now you can chat, uh, and do images in the same place and, uh, you know, there’s a beautiful editor interface now on desktop that you can do and, uh, [00:55:00] as of today, we’re making some final adjustments on the mobile Uh, image editor interface so that on mobile, it’s going to be a much more beautiful and enjoyable experience to create images and videos.

Um, so really just, building, build the best and, and build the, you know, the most enjoyable AI tools that there could possibly be and making sure that, uh, all modalities are included.

Jeff Sieh: And it looks great, too, by the way. It’s, it’s, Dustin does, uh, really try to get design, uh, nailed, and I think he does it. So, um, we’ve got a couple questions I want to just jump real on, on to real quick before we, we wrap up. Uh, so, Image Migration says, you guys tuck pricing. I missed that part. If you go to, uh, magi.

co forward slash Jeff, that’ll give you actually, if you want to look at it, 30, I think your first three months and the pricing is listed. If you go and click on it, I’m not sure off the top of my head what it is now, uh, but that’s a great way if you want to go there and look at stuff. I mean, it’s also a beautiful website.

Um, and then you can ask people like, if you have any questions, ask, there’s a lot of Magi users right here, uh, in this, wherever you’re watching from, [00:56:00] ask them, they’ll give you the low down as well. And another great question was from, uh, let’s see real quick here. Um, where did I put it up? It was, um, uh, Kristen asked, Do you list out all the tools you have mentioned somewhere?

Most everything, uh, is built into Magi that we’ve talked about today. If we mentioned any tools, Kristen, uh, you can actually go to my toolbox, JeffSieh. com forward slash toolbox. It’s got everything in there. It’s got the Magi, uh, link in there. It’s got all sorts of things that we use for image generation, AI.

Everything that runs our, my business is there. So JeffSieh. com forward slash tool. That’s that. Forward slash toolbox is where you can find it. I’m just so excited that Dustin was here. And so thank you guys for watching. Gary says, super interesting talk. Thank you for taking this question about Apple’s AI images by Gretchen, everybody, Rich, all the folks who stopped by today, uh, stick around because we’re going to do that drawing, but I want to give Dustin a chance to tell what he’s going on, where to find him, all that sort of thing.

Dustin Stout: Yeah, I mean, magi. co is where you can find all the AI things. I’m at Dustin W. Stout on all the social networks, so if you want to [00:57:00] find me there, say hi, I would love to hear from you.

Jeff Sieh: Awesome. Connor Brown, the unsinkable Connor Brown. Where can people find out about you?

Conor Brown: Vacationkingdoms. com or across social medias at vacationkingdoms, uh, all over the place too. Jeff, let’s pick a winner. I’m,

I’m

Jeff Sieh: are you ready? All right. So, uh, I’m going to go and hit this. This is for, remember the Ecamm Creator Camp free ticket. We got a drawing. So I’m going to hit the draw winners button and we’ll see who won. Let me go actually to the thing and here we go. Here we go. I think it’s going to give me a

Conor Brown: exciting. Oh,

Dustin Stout: Oh, this is so exciting.

Jeff Sieh: I know. Thanks 10 seconds.

Conor Brown: much longer than I was

Jeff Sieh: Yes, here we go. Let’s see who it is. There’s a lot of entries. For

Conor Brown: 2, 1,

Jeff Sieh: and watch my internet go out. Brandy Jo from Olympic. Uh,

Conor Brown: Olympia,

Dustin Stout: Olympia Watch.

Jeff Sieh: yes, you are the winner. Um, I will email you the way you can pick up your ticket. Thank you everybody for playing, uh, and entering for this.

Oh gosh. How long has it been, Connor? We’ve done that,

Conor Brown: Months. Let’s hope Brandi Jo is in the purple cabin with me. Let’s go team purple.

Jeff Sieh: she’s sitting with you. Okay. Brandi Jo, put her in the Connor cabin. So, uh, thank you guys for, for entering. Uh, we’ve got some more stuff coming up, uh, into this year and next year. So appreciate you guys watching the show today. Thank you, Dustin, for sticking around for so long. Yeah, this has been amazing.

We’ll see you guys all next week. See you guys later. Bye everybody.