Everyday AI Podcast – An AI and ChatGPT Podcast - EP 342: New AI Image Generators You Won’t Believe - Flux, Ideogram 2, and more

Episode Date: August 22, 2024

Win a free year of ChatGPT or other prizes! Find out how.Y'all won't believe these new AI image generators.... They're starting to blur the line between fact and fiction. (In both a good and a bad way.) And in the past few weeks, the entire game has changed. We go over what's new and what you need to know. Newsletter: Sign up for our free daily newsletterMore on this Episode: Episode PageJoin the discussion: Ask Jordan questions on AI image generatorsRelated Episodes: Ep 218: Winning the Probability Game in AI VisualsEp 198: Midjourney V6 – What’s new and producing powerful ad creativesUpcoming Episodes: Check out the upcoming Everyday AI Livestream lineupWebsite: YourEverydayAI.comEmail The Show: [e-mail] with Jordan on LinkedInTopics Covered in This Episode:1. Different image generators and their uses2. Importance of AI images3. AI image generation timeline4. Potential misuse of AI image generators5. AI image generation promptsTimestamps:02:50 Daily AI news06:20 AI image vs real image test09:12 AI creates images from text descriptions using AI.10:05 AI image generators use diffusion models, face copyright concerns.13:43 DALL-E and OpenAI in AI image game.18:14 Live demonstration, prompt, show, vote for best.21:30 Open source model used for image generators.27:49 Issues showing small URLs, presenting photo options.30:46 Interface is a love-hate experience with AI.34:47 Generate, observe real-time photo realistic beach image.38:25 Imprinting painting with quick color changes attempt.41:57 Results of AI image generator with social platform.42:41 Explore AI image generators for various industries.48:36 Former president spreads problematic AI-generated images.51:16 AI image generators creating videos are powerful.Keywords:AI tools, photorealistic images, ideogram, Imagine 3, Flux 1, image generation, Grok, Google DeepMind, Jordan Wilson, Midjourney, user control, AI image generators, misinformation and disinformation, open source models, copyright concerns, DALL E, diffusion models, AI-generated images, AI demonstration, Chat GPT course, webinar, audience vote, AI impactSend Everyday AI and Jordan a text message. (We can't reply back unless you leave contact info) Start Here ▶️Not sure where to start when it comes to AI? Start with our Start Here Series. You can listen to the first drop -- Episode 691 -- or get free access to our Inner Cricle community and all episodes: StartHereSeries.com Also, here's a link to the entire series on a Spotify playlist. 

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. In the past week alone, the AI image generator space has completely changed.

If I'm being honest, I've never seen such movement in such short of a time period in one area of artificial intelligence. The past week has been wild. So I'm going to be going over today live. the new AI image generators that you won't believe. So we're going to be talking a little bit about flux, ideogram, 2, and Imagine 3, and some updates to mid-Journey. All right, I'm excited for today's episode because I think AI images are going to impact us all much more than we think. And maybe when you first saw an AI image, maybe a year or two ago, you kind of laughed. You can't laugh at these anymore because, yes, AI images,

have great impact and great ability to, you know, really extend to companies' creative use. But there's also some huge downsides. All right. We're going to be talking about that today and more on Everyday AI. What's going on, y'all? My name is Jordan Wilson, and I'm the host of Everyday AI. And this is your daily live stream podcast and free daily newsletter, helping us all learn and leverage generative AI to grow our companies and to grow our careers. And that's one thing.

I think today's show is going to be a great. one. And if you are joining us live, like Michael and Fred and Harvey, Cecilia, Cyber, Colby, what up, Kobe. And yeah, I mean, this is one, if you're joining us live, I think this is going to be a show. I need my live audience for today. I'm going to be calling out on you for some suggestions to vote on some things, asking questions. So if you're listening on the podcast and this is something that you have a real interest in, this is one where you're going to want to check your show notes. You're going to want to go to Your EverydayAI.com, sign it for our free daily newsletter and go watch this one. You're going to need to watch this one.

I'm going to do my best to describe what's going on, but I think you can appreciate that this one might be a little bit hard to paint a picture multiple times with words on a podcast. But I'm going to do my best. But make sure while you're at your EverydayAI.com and while you're also going to look at the newsletter, make sure you sign up for our free contest. Y'all, we're giving it away. Thanks a million. So make sure you go check that out. All right.

So let's go ahead and get our correct one up here as we go over the AI news. All right. So Google and California have struck a $125 million deal to support local newsrooms amid AI concerns. So Google has partnered with California lawmakers to funnel a minimum of $125,000,000, million over five years into local newsrooms, marking a significant attempt to bolster journalism in the digital age. This agreement highlights growing tensions between obviously the technology sector and journalism and media companies, especially regarding the rise of artificial intelligence.

So Google, according to reports, Google will contribute $110 million, while California will add $70 million to a new transformation fund managed by UC Berkeley's Journalism School aimed at supporting local newsrooms. All right. So that one is pretty interesting. All right. Our next one, this one is pretty important, I would say, y'all. But new reports are showing that Microsoft is set to launch its controversial AI feature recall for testing in October. All right.

So this feature is going to be reportedly being rolled out to insiders first. You have to be a Windows insiders to start testing in October. The feature has sparked a lot of debate due to its ability to take screenshot of user screens, raising security concerns among researchers. The original release of this recall feature was delayed to address these issues. So if you didn't see anything about this originally during the Microsoft Build Conference, Think of it like that your operating system is going to have a photographic memory of almost everything that's on your screen, which obviously caused a lot of people to both say, oh, this is amazing and, oh, this is potentially terrifying. And then you can talk with the co-pilot rewind feature to essentially recall anything.

So the company has committed to implementing security enhancements for recall before its broader release on co-pilot plus CDs. Yeah, you have to have the co-pilot plus PC to take advantage of this, which has a better hardware. Also, the timing of recalls release is very strategic as Microsoft aims to attract consumers looking to pursue new computers during the holiday season. All right, our last piece of AI news, Neurlinks's second patient has used brain chip to play video games, showcasing new possibilities for gaming and independence. So Neurilink, the brain implant chip company co-founded by Elon Musk, has made headlines with the update on its newest patient, Alex, its second human patient equipped with a brain chip implant. This breakthrough not only enhances gaming experiences, but also significantly improves the lives of individuals with mobility challenges. So Alex lost control of his limbs due to a spinal cord injury and now uses the Neurlink brain chip to control the mouse on his last.

with his thoughts. So in a video demonstration that was released by Neurlink, Alex easily navigated and shot targets in a video game calling the experience quote unquote insane beyond gaming. He is also reportedly using the chip for creative tasks, including designing and 3D printing a custom mount for charging the implant. That's, that's wild, y'all. That's wild. All right. I want to talk about AI images and I wanted to start by showing you this, All right, which of these photos for our live stream audience, go ahead and type in, which one is real? It might be more than one, but at least one of these is real. So go ahead and type in your guess on which one of these four pictures is real.

At least I'll tell you this, at least one of them is maybe more than one. So I took one of these photos on a vacation, and it's a photo for our podcast audience. This is a photo of a leopard kind of walking around. You know, hey, if I ever get rich and retire early, this is what I'm going to do. I'm going to go on safari. I'm going to take people out, show them how to shoot photos, right? And that's also very important for today's episode because I've taken about more than a million photos, right?

So I used to do photography semi-professionally. I have like four DSLRs, you know, sitting here in my closet somewhere, full frame. Yeah, expensive hobby, right? So one of these, at least one of these, is real. All right. So I'm curious, you know, everyone gets your guess in. So, you know, Harvey says three. Cecilia says three and four.

Brian says three. It looks like the leader so far as people think number three is real. So I have four different photos. Some of them are AI generated. At least one of them is real. Fred says one and four. Harvey says three. Let's see. Cyber here says four. All right. So Colby says one, three and four. Cecilia says two. All right. So I'll tell you this, y'all. The one with the most votes, number three, is AI generated. All right. So I think the second most votes, you guys got it right?

But that just that just goes to show you, y'all. I mean, let's look. All right. So we have ideogram two, mid-journey, imagine three, and then the real one. So even our live stream audience, who I would say, I would say that you all are probably much more in tune with everything that's going on in the AI world than the average person out there, right? For our live stream audience, I just shouted you all out.

I know most of y'all have been around for a year plus, right? You've seen these tools. You've talked. You've been able to talk with the experts who are creating AI. And even you all were confused. If I'm being honest, I was up really late making this presentation last night. And even I was starting to get confused. I'm like, wait, which one did I take?

Right? I took the photo. And I forgot which one is real. All right. So, yeah. Are you guys, are you guys, uh, stumped there. So let's talk about first, what is an AI image?

Okay. Because that's extremely important. So AI image generators and AI images, well, you create images from text description using artificial intelligence. I want to start with a basic definition. Okay. So you described something. Like in this case, my prompt, I did something simple.

It was along the lines of a photo realistic or a photo of a leopard walking the South African safari with, you know, kind of like weeds or a prairie in the background, right? Simple text prompt. All right. And then what happens is essentially most AI image generators are examples of diffusion models. In diffusion models, that's the technique. behind AI image generators. So essentially, they start with random noise, okay? And then they slowly turn that random noise and random pixels into a clear image step by step.

And these diffusion models, these AI image generators, are generally trained on, I mean, no one's ever announced the number, but presumably it is tens of billions, hundreds of billions, potentially trillions of images, right? And yes, there's a lot of copyright concerns here. A lot of the bigger names in this space, such as Mid Journey, have gotten some lawsuits already for allegedly, you know, using copyrighted work. Let's just be honest. Most AI image generators and most AI companies in general train on the open Internet,

whether they have access to that information to train their models on or not. And I think, you know, I've talked about this plenty. There's going to be a lot of lawsuits settling out. What is copyright? What is IP, right? But that's not what today's show is. So that is how AI image generators or, you know, these kind of diffusion models work. So like mid journey, like ideogram, like Imagine 3, like Dahl-E, like stable diffusion.

Maybe you've heard of these names. Okay. So essentially, you put in a text, a text input. The diffusion model just kind of starts with noise and then it turns that into an image based on all of its training, right? All these billions of photos where a human starts labeling them. And they said, oh, this is a leopard. This is a cheetah. This is a lion, right?

This is a South African, you know, landscape. This is a Brazilian landscape. Right. So all of these things, your words mean a lot when it comes to working with these diffusion models, right? you can swap out one simple word and get very different photos. All right. And then you eventually get a final product. Okay.

And we're going to get into now the king of the hill. All right. So the king of the hill here is mid journey. Well, they were. So I started at the top of the show talking about how much things have changed this week. They've changed exponentially, okay? Like, let's just talk. Let's just talk here.

In the last week alone, I think Mid Journey has maybe lost its spot on the top of the hill. All right. But we're going to figure that out live here, right? As long as you guys stick with me through my tech issues. But Mid Journey was the king of the hill, but I don't think they necessarily are anymore. Okay? Because let's, but also we have to go ahead and acknowledge. how much AI image generators have even grown.

I talked about that at the top of the show. Maybe you saw something a year or two ago and you said, oh, that's cute. Oh, that's funny. That's entertaining, right? Like original models like Dali, right? Poor Dali, they've gotten lost in the full, right? Dali was actually the first technology in 2021. And then Mid Journey came along in 2022.

That's when, you know, my team and we started using Mid Journey is kind of a couple months after it first came out. And I do think just FYI that either Dolly or Open AI is not out of the AI image generating game because we've seen the capabilities of their AI video tool, SORA, which has yet to be released, can also produce images. And those images are amazing. Okay. So, but right now, kind of if we're looking at the timeline, Dahlie was the first on the scene. you know, before chat GPT, Open AI released Dolly, right, which a lot of people don't know. But Mid Journey came in, swooped up all the attention, all right? And a lot of people, their first interaction with AI images was seeing people share something

via mid-journey. But it's changed a lot. So, you know, one of the first versions we used was V2. But I just, as an example, I took a V3 image, ran the same thing back to back over the course of 18 months. All right. So for our live stream audience, this was a. a simple prompt. This was a simple prompt, something about, you know, a man in his mid-30s

wearing Nike gear, posing in the middle of downtown Cincinnati. All right. So the image on the left is from Mid Journey v3 about 18 months ago. And the image on the right was from last night. All right. And live, I mean, live stream audience. Can I get your reaction on this so I can read it to our podcast audience? Because I don't think people are going to believe. me, the first image, the one from 18 months ago, I mean, it looks, I don't know, it looks like, like Plato or, right? It doesn't, it doesn't look. Nothing in this photo looks remotely real, right? You're not going to fool anyone, right? You can't use, you couldn't use images, I don't think, in real campaigns. You couldn't use them for your business, for blog posts,

for even for mood boards, for ideas, you really couldn't, right? But look at this one. Look at this one, y'all. This one from last night in the newest version of Mid Journey. This looks extremely real. You know what? If I'm pixel picking, I can stare at this image for a very long time, a very long time. The only two dead giveaways, right?

The shoe here at the bottom, the right shoe is a little wider than the left. And the zipper kind of tag is maybe slightly longer than a zipper tag would be. And hey, I only know that because, you know, I've worked with Nike on and off for like 10 years. So I know what their clothes look like. But this is extremely impressive. in the growth that we've seen in AI image generators in mid-jury over the course of a year and a half. To me, it is mind-boggling. Mind-boggling. All right. So, woozy here says the first image looks like a screenshot from an early version of Grand Theft Auto.

Yeah, I agree. Right. It doesn't look real. All right. Let's take another look here. Here's another example. This is a simple text prompt asking for, an image, an overhead shot of a famous waterfall in Iceland. All right. The one on the left obviously struggled. It looks like, I don't know, a decent painting, right? I don't even know if you'd call it decent. The one on the right, wow, right?

It's beautiful. It is a very good looking photo, right? And that pains me to say. I mean, I went to Iceland. I took thousands. I don't know, probably tens of thousands of photos. I didn't get one that looked like that. Obviously, you know, if you go look at, you know, Google Maps,

this exact same instance of, you know, the water falling into, you know, into a rift in the land. This exact one doesn't exist, right? but it's a very aesthetically pleasing, extremely real looking photo. Y'all, live stream audience, can you believe the difference here in 18 months? It's wild.

All right. So here's what I need your help for live. All right. Give me a prompt. Adobe just introduced an entirely new way to create, bringing the power and precision of its creative suite into one conversational experience. Meet Firefly AI Assistant, now live in the Adobe Firefly app, the All In One Creative AI

Studio. Powered by Adobe's Creative Agent, Firefly AI Assistant lets you start with your vision, just describe what you want, and shape the outcome as it takes form with the Assistant. The Assistant orchestrates multi-step workflows, drawing on 60-plus pro-grade tools across Adobe Creative Cloud apps, including Photoshop, Illustrator, Premier, Lightroom Express, and more to help bring your ideas to life. You can also get started with creative skills, a growing library of pre-built workflows for common creative tasks, like batch editing photos, creating mood boards, portrait retouching,

and creating social variations. Every step the assistant takes is visible so you can refine, redirect, or take over at any time. You stay in the driver's seat as the creative director. Adobe Firefly AI assistant now in public beta. See it today at Firefly. We're going to do this live. All right.

Give me a prompt. And I'm going to go ahead. We're going to do this in mid-jurney. So give me something somewhat descriptive, all right? And I'm going to walk, I'm going to walk everyone through what I'm doing. But then I'm going to show everyone at the end, the examples. And then I'm going to want everyone to kind of just go ahead and vote and tell me which one is best. All right.

So as an example, a prompt might be something. like, you know, DSLR photo of a man in his mid-30s wearing a vintage suit standing in a bustling 1920 street, cinematic lighting, or something like realistic photo of the Chicago skyline, cinematic, bright and vivid colors at sunset. All right. So I want a couple. We'll probably just do one or two because I'm already, I'm already having some tech troubles here. So thanks a lot, Internet and AI is coming after. the Everyday AI show today.

So I need a couple from our audience. And then I'm going to walk through each one of these, each one of these platforms. And I'm going to talk a little bit as I do it live. But I'm not going to show you live. All right. So yeah, I know that's a little frustrating. I'm not going to show the process because I want everyone to see the results live.

All right. So I'm going to walk everyone through this. All right. Let's see. Someone else. All right, we got one from Harvey. All right. So thanks for that, Harvey. This one's a little long. Let's see, because I'm going to also have to copy, copy and paste this. So here's what Harvey said, a vibrant, colorful illustration of two hands forming a heart shape. The thumbs are positioned at the bottom. All right, that one is going to be a little, a little tough. Let's try and get something that you might see every day. So that's, that's kind of what I want. I can pick one out if our audience, you know, can't come up with anything. But here's what we're going to. be doing. We're going to be looking at, and it can be simple, y'all, it doesn't have to be something, you know, something very long. It can be a short sentence. All right. So we're going to be going over mid-jurney, ideogram. So ideogram just released version two. We're going to be going over

imagine, okay? And then we're going to be going over flux. All right. And I'm going to tell you a little bit about these image generators here as we go through them. All right. Let me just look one more time here, see if we have any good ones here. All right, here we go. I like, I like this one, Colby. Hey, Dr. Harvey, I liked yours. It was just, I wanted to get something that a lot of people maybe have seen, all right, have seen before. All right, so here we go. I'm going to use a version, I'm going to use a version of Colby's here. It's pretty, pretty long, pretty long, but we'll try it. All right, so it's create an image of a dynamic action scene featuring a surfer riding a massive ocean wave. The perspective should be from a DSLR camera with a 70 to 200 millimeter lens.

Love that lens, by the way. Capturing the subject in a crisp, detailed focus against a soft blurred background. All right. So I'm just going to use the first half of that, Colby, because the rest, it's, you know, to be able to read it over and over. It's going to be a lot. All right. So here's what I'm going to do, y'all.

I'm going to take that prompts. All right. So then I'm going to go into all of these different, all of these different image generators. and then I'm going to show you what they produce. And then I'm going to want you to go ahead and vote on which ones you think are best. All right. So as a reminder, I'm doing this all live, y'all. And I'm going to walk through each platform.

All right. So to let everyone know, Flux is actually an open source model. It's new. It's really, I think, kind of like almost broken, the internet, right? So I'm using all of these different tools, all right? So we're doing mid-journey, we're doing ideogram, we're doing imagine three, and then we're doing flux via grok, okay?

All right, so I'm saving all of these now. I'm kind of saving my favorite, and I think this is a pretty, a pretty good example here. All right. So we're going to do this one here first. All right. There we go. All right. So we have our mid-jurney image.

Now I'm going into ideogram. So like I talked about, ideogram just released yesterday an update. Okay? So now we have ideogram two. And ideogram was built by Google engineers. All right. So what I'm doing now is I'm saving all these image addresses.

Okay. because I don't want y'all to know which one is which. All right. There we go. Y'all, some of these, if I'm being honest, I'm very impressed by some of these. Very impressed. All right. Some of them, not so much.

And again, these are just like examples of zero shot, right? I'm not doing anything, you know, a lot of iterative prompting. All right. So thanks for sticking with me. All right. So now we're going to go ahead and I want you to say which one's better. And I'm going to say one, two, three, four. All right.

So I'm going to go ahead here again and share my screen. Hey, this is Jordan, the host of Everyday AI. I've spent more than a thousand hours inside chat GPD, and I'm sharing all of my secrets in our free prime, prompt, polish chat GPT course that's only available to loyal listeners like you. Check out what Mike, a freelance marketer, said about the PPP course. I just got out of Jordan's webinar. It was incredible.

huge value. It's live, so you get your questions answered. I'm pretty stoked on it. It's an incredible resource. Pretty much everything's free. I would gladly pay for a lot of the stuff that Jordan's putting out. So if you're wondering whether you should join the webinar, just make the time to do it. It's totally worth it. Everyone's prompting wrong. And the PPP course fixes that. If you want to access, go to podpppp.com. Again, that's podpppp.com. Sign it for the free course and start putting chat GPT to work for you. All right. Here we go. Here, here's our image generations. All right. Actually, don't, don't cheat. There's actually no way for me. Let me see here. I'm trying to get rid of. All right. I can't fully get rid of it. So don't look at the small

URL at the top. This is number one. Okay. This is number two. This is number three. Oh no, we lost one here. We lost one, y'all. Sorry, give me a second. All right. And I want everyone's vote. I want everyone's vote. Where do we miss one? Here we go. Okay, we missed one. There's two. There's three. There's four. All right. Sorry, y'all. All right. Ready? I need your votes. I need your vote. I wanted to do this live. This is one of those. This is one of those y'all where the tech issues are getting to me. All right. Here we go again. Ready. Here's one. Here's number one. Here's number two. Here's number three. And here's number. four. All right. So again, this prompt came from Kobe, so thanks for that. It was create an image of a dynamic action scene featuring a surfer riding a massive ocean wave. The perspective should be from a DSLR camera with a 70 to 200 millimeter lens capturing the subject in a crisp, detailed focus

against a soft blurred background. All right. So again, number one, all right, pretty good one here. We have a pretty monster wave. We have a surfer. Nothing, nothing in this one screams to me either amazing or terrible, if I'm just giving my straight feedback, but the wave looks extreme here, right? Here's number two. This one, it looks like much more picture-esque to me. It looks like, you know, we have a sunset. To me, this screams like some Hawaii or Tahitian vibes, much more visual, I would say. Here's number three. This one, this one, to me, me looks much more like action sport, right? It looks much more like something you might see out of a maybe a surfing magazine. All right. And here we go for number four. All right. So hey, my,

my thoughts on this one. Realism is a little down. However, the colors look pretty vivid. All right. So live stream audience, what did everyone say? It's unanimous. It's unanimous. All right. Oh, it's not unanimous. Joey, Joey just broke it. But for the most part, Most people are saying number three, number three, number three. We got one vote for two, actually. All right. So let's just go ahead. Number one here, this one is mid-journey.

Okay. So thanks for not cheating. I think you could probably see the URLs at the top, very small. There's no other way for me to show these photos easily since I'm having issues with my slides. All right. So number one, which no one voted for, was mid-journey. Number two, this one here, which I personally love this one, this one is ideogram. If I was voting, I probably would have said number two because I think number two is the one that probably did the best with prompt handling, with following the prompt very, very well.

And then so number three, which is what most people voted for, I think actually number three, When I look at it again now after seeing y'all voted for it, pretty good sense of realism, if I'm being honest, right? If you're scrolling social media, I don't think you're going to second guess that this is a real photo. It looks very good. I've seen no crazy AI artifacts. You know, there's no three legs, anything like that. It's a good wave. It did follow the prompt handling very well.

And then number four, this one was flux. So the new open source model. Okay. And again, like I said, this one, it was more of a head-on shot. So you get something. It looks just kind of almost like a painting. It's not, you know, very realistic. All right.

So I just want to put this out. I want to put this out there, y'all. This is not a scientific test, right? This is not a scientific test. I wanted to do something live and go over everything. But let's just go out now that we're sharing, you know, sharing everything here. I want to walk you quickly through each platform. All right.

And we're going to go through this quick. But mid journey. Let's talk mid journey first because there's actually a lot of new things inside of mid journey. One of the biggest things, and this was just yesterday. So this is fresh news, y'all. So now the mid journey web interface is open to everyone. And when I say everyone, I mean literally everyone. So mid journey for the longest time has never.

not had a free plan. You cannot try the platform. Also, up until this week, not everyone could access Mid Journey via the website. You had to go into Discord, which I personally didn't hate. But now a Mid Journey's been beta testing. They're essentially their new website platform where you can do all of this and work with their technology inside of the Mid Journey platform. And now that's available to everyone. So I don't know how long the free trial will be around. It's definitely worth going in there and trying it out. So I'm going to give everyone a very quick overview of the platform. So like I said, I think up until this week, it was mid-jurney number one and everyone else

wasn't even a close second. I don't necessarily think so anymore. So as we look through the interface here, I both hate it and love it. So now you have this kind of side panel on the right-hand side. You can go through and see all your image generation. So here I was last night, just running a couple of quick, generations of these leopards that we went through and some other photos. So it's a clean interface.

You have to get used to it. And here's the thing, y'all. You don't have to be able to speak AI image prompting anymore to get the most out of Mid Journey or other tools, right? I can literally just go up here and let's just maybe, I'm going to rerun this. A wheel A, sorry if I didn't get your name correctly. She gave me this one. so a photo realistic image of a beach at sunset, right?

All you do, you'd go up there, you click enter, and then you go under create. Okay, and I will say I only have a paid plan for Mid Journey right now. I'm on the free plan for everyone else. So then when I go into Create, I can click here and I can go and look at all, you know, we don't have a ton of photos, but I don't know, tens. We at least have thousands or maybe tens of thousands of photos inside Mid Journey. I can go look at all of those. But as it's updating here, it's.

done, right? So then I can go in and click inside these images in MidGourney, and I think they came out pretty good. So this is, you know, a little doesn't really look real necessarily. And then inside Mid Journey, you can do a lot of different things. So I can download this. I can open it in Discord. And there's so many parameters. We're not going to get into the parameters of Mid Journey, but Mid Journey does kind of have, you know, these, I'll call them like how there's coding, right? and you can put in like little lines of code if you're a developer that do certain things. Mid journeys like that. They have a lot of these what are called parameters.

We're not going to get into it. But it gives you very granular control. So a lot of the granular control that you can get out of mid journey, you can do so with parameters. But there's a lot of easy things that you can just click on, right? So I can just click on. I can do a remix. I can do a pan.

I can vary certain images. So there's a lot of very easy things to do. I can rerun it. You know, so Mid Journey, I think, is a great and easy to use platform. All right, that's enough with Mid Journey. I wanted to give everyone a quick overview. So now let's go to ideogram. Okay, so ideogram is, ideogram 2 is brand new.

So like I said, this is built by former Google employees with plenty of big name backing. And like I said, ideogram 2.0 was just released yesterday. And it does really well, particularly in creating images with complex text. such as logos or banners. That's something where Mid Journey has traditionally struggled. It's gotten better over the last couple of months with some updates like Mid Journey 6.1. But still, I think imagine or sorry, ideogram 2.0 is really positioned as a tool that can really work well with text. And I think I'm just about out of credits, but I'll show you how easy it is.

I'm going to do the same thing. Yep, I think I have one more here. So I did a full review of this new platform when it came out yesterday. This is, I think, a little more user friendly than mid-jorney. Because like I said, for a lot of the things to accomplish in mid-journey, you have to kind of know some parameters. You can still accomplish a lot without any parameters, right? Just through natural language.

But really to get some of the true powers, you have to kind of speak a little bit of mid-jurney. Idiogram, not so much the case, right? There's something called a magic prompt where it can make your prompt better. I don't need that. You can change the aspect ratio just by clicking. You can choose the model. I'm going to do 2.0. I love the color palette. So you can do a different color palette based on if you want a certain mood or vibe. I'm just going to click auto on this one. And then I'm going to generate. I think this is the last one. Also, with ideogram, I can click different styles. So general, realistic, design, 3D, anime. I'm actually going to click realistic here. Click generate. And then we're going to do our photo realistic image of a beach at sunset. So like I said, with ideogram, I'm on the free plan. It is somewhat limited. You have, I think it's like essentially 10 credits a day. And then you'll see in real time, you can kind of see this diffusion in action here, right?

So on my screen, it's starting with noise. And then these noise are slowly turning into pixels. And then the pixels are turning into blurry photos. And then the blurry photos now are like pretty good looking photos, right? So now I can go back, go to my images. And here we go. here's our photorealistic image of a beach at sunset. So like I said, ideogram is great with handling text.

You know, it's really, really good if I'm being honest. All right. Now let's go to Imagine 3. Okay. So this, y'all, I'm not going to lie. Imagine 3 might be my favorite. The results are great. And if you guys have tuned in to the show a lot, you know, sometimes I'm a little

harsh on Google, right? I think, if I'm being honest, I think Google has a sometimes disjointed AI strategy. It's so hard to find the tools, use the tools. They have like, you know, 10 different versions. You can't use them all on the front end of Google Gemini, their large language model. Google Gemini has a hard time using Google or searching the internet, which makes no sense. So I'm hard on Google, but I love, I love their Imagine 3. So again, in true Google fashion, it's a little hard to find.

you can just Google Imagine 3. We'll put the link in the newsletter, but you actually have to go into the AI test kitchen. Yeah. And then use what's called ImageFX. So the model is Imagine 3, but that's how you get to it. It's a little clunky if you're new here. All right.

So let's just go ahead. I'm going to do this same one. Oh, nope, this is not what I wanted. Let's go back to image. There we go. Try the same thing. Photorealistic image of a. beach at sunset. Let's go ahead and just click create. So one thing I like about Imagine 3,

and you saw that on the other one, but when you have a longer prompt, it actually gives you these drop downs, right? So I said a photorealistic image of a beach at sunset. And it made the words photorealistic beach and sunset essentially drop downs, right? So if I click photorealistic, now it's giving me different ideas. It says abstract, impressionist, Surrealist. Beach, I can click beach. It says, now it says forest, mountain desert. Sunset. Now it says sunrise night day. So it's just a really cool way to play around a little bit. Oh, FYI, I didn't get to this in ideogram, but you can do some remixing and upscaling when you have a paid plan. You can upscale, but I think you can remix for free. All right. So now going back to imagine, so I kind of showed

you things that you can do in body, right? Another great thing that I, like about Imagine from Google. So first of all, I mean, y'all, what do you guys think of these results? Right? I don't know. To me, I mean, this one is, you know, maybe a little too much vignette. I really like the results of Imagine 3. All right, very simple prompts.

So the other thing, an Imagine 3 right now is free to use and the limits are pretty good. So there's also some in painting, right? So as an example, right, this one might be. hard to do. So I'm going to say, you know, you can get a brush, essentially. That's what in painting is. So I'm trying, this isn't going to turn out well because I'm doing it fast or this isn't going to turn out good. I'm trying to paint some of the purple parts of the sky. And then you can essentially describe the change. And I'm going to say, you know, make this part of the sky blue, right? So this might look a little wild because we have some like crisp,

almost like autumn type colors in the sky here. You know, it's kind of going top to bottom, like purplish down to like fiery orange. So blue might not do the best, but I'm just trying to do something turning the top part of this image, which is more purplish to something more blue. I didn't do a good job filling in all of the details. And with the brush size, I was just doing something quick. So you see, it did a pretty decent job there, right? So now we actually have a nice gradient. It goes from more of this like indigo-blueish to a pinkish, too.

you know, a yellow orange. So it did a pretty good job there in the matter of, I don't know, a second. So I'm actually a huge fan and bullish on Imagine 3 from Google. I think this is one of those areas. It's like, oh, Google's actually crushing it here. All right. And then our last one here, we're going to have to go into Twitter, actually. And then we are going to have to use GROC. All right. So we're going to do the same thing. For this, I need to say generate an image of, All right. So there's other ways that you can use flux. So flux one, you're going to be hearing a lot about flux one.

And I'll tell you why here in a second. It's an open source model. So essentially you can work with it offline, right? Like you could any open source model. And you can fork it. You can build on top of it, et cetera. There's a lot of different platforms, including perplexity, including, you know, Twitter's GROC or Axis GROC, right?

their large language model. All right. So I'm just saying generate an image, sorry, generate a photorealistic image of a beach at sunset. All right. So now I have to put this into GROC and then it uses, GROC uses flux 1 in open source model to create this. So pretty good here, pretty good. The downside right now, depending on what platform that you use to access the flux 1 model, you don't really have a lot of editing capabilities, FYI, where in Mid Journey, you do have a good amount of editing capabilities. And then also you saw in an ideogram, you have some basic editing capabilities. I think you have a lot more on the front end before you generate. And then we saw on Google some of the capabilities to edit their live as well.

All right. So that's enough of doing this over live. Now, I kind of want to wrap this up because there's a couple big points that I do want to get to, all right, that I think are important. And also, was this helpful, y'all going through seeing these lives? All right. So like we said, the newish kids on the block are flux and more on flux here in a second. All right. But flux one, I think is going to be talked about a lot, all right?

number one, because it's probably the best open-sourced AI image generator out there. It came out of seemingly nowhere just this week, and everyone's talking about it. Then like I said, imagine three from DeepMind, okay, from Google DeepMind. For me, personally, I love the interface. The results, even if you're not great at prompting or describing something, I think the results are actually fantastic. Ideogram too. So ideogram is from former Google employees, and they have a lot of funding.

And ideogram gives me like very like Instagram-ish vibes, right? Because even your kind of like your account when you go into it, it is kind of set up to more be like almost like a social sharing platform. So you create something. It's an AI image generator, but it's almost like a social media platform as well, where you're sharing some of your creations. All right. And again, so all of these that I've talked about so far,

including now Mid Journey, have free plans. So you can go try these out and see if they're right for your next creative project, for your business, right? I think there's so many great use cases. So, you know, as an example, this one here, ideogram, it's showing something that looks like a, you know, nice interior design. Maybe you're an interior designer. Maybe you're a real estate group and you don't have, you know,

renders for a new project that you're working on and you don't need something final. You just need something to get the ball rolling. So many of these AI image generators, regardless of what type of business that you work in, there's great use cases now that I don't think existed 18 months ago. All right. Let me know live stream audience, but which AI image generator is the best? All right. I'm going to give you, I'm going to give you my take.

My take is I don't know, right? which for me, that's not normal. You asked me a week ago, two weeks ago, I would have said mid-journey, hands down, and it's not even close. But y'all, we've got over the last, imagine three from Google actually was just generally rolled out this month. And then just this week, we got flux and then we got ideogram two. So in the month of August, this is completely changed.

If I'm being honest, I don't think Mid Journey a month ago had an actual competitor. They do now. And here's the thing. I think if you're a beginner, I think you're actually going to get less out of mid-journey than let's say me, right? I've used Mid-Journey a lot. I've been using it for years. You know, we do a pretty decent job at it. We've done shows on it.

We've taught some people, some of the basics, right? I think for people who are quote unquote experts or who are creatives, right, if you are someone that in theory is going to be using AI images, you know, almost every day or on a weekly basis, I do still think Mid Journey is probably the best for most use cases, right? But for everyone else, I don't think so. Mid Journey, like I said, to really get the most out of it, it's not necessarily for beginners, although their new web interface helps. You know, each new update, you know, 6.2, 6.3, if we get mid-journey 7, it's going to help

beginners. But to really get the most out of it, you really have to work with those parameters, right? So they have dozens of different parameters. And you've got to kind of know them and use them correctly, all right, which are a little tricky. Where some of these other, you know, AI image generating tools, they're a little more user-friendly, maybe less robust, right? But you can click a drop-down, right?

You can just go back there and click something, right? We're mid-journey. It's not a lot of clicking, right? It's, you got to know, you got to know the parameters. You got to look at the site. You know, it's almost like you still got to speak a little code to get from good to great. Where I think now out of the box, some of these, like we just saw in our live demonstration,

I don't think mid-journey was the best, not even in the one I did of the leopards, not even in the two that we kind of did live here, the surfer in the sunset. I don't think Mid Journey was the best because, again, we were in those cases replicating a beginner, someone that goes in there with a basic prompt, not playing with parameters, just kind of simple out of the box. So I don't know which AI image generator is the best. All I know is this is going to lead to some crazy development. And here's why it's important.

Because there's pros and there's cons, right? So I talked a little bit about, oh, imagine what your business can do, right? imagine what new creative capabilities you have, right? As a former photographer, like I said, I've gotten paid. I've shot all kinds of, you know, events, you know, for, for nonprofits, weddings, you know, other family photos, right? I've done a lot. So I'm mixed on this, right?

On the pros and the cons, what are the pros and the cons of AI images? Well, it brings a new level of creativity to everyone. It levels the playing field, which I actually think doesn't necessarily, mean it's a good thing for everyone. But also, there's so much good that can be done, right? I even think of myself. I think of myself. I worked in, you know, a nonprofit for 10 years.

We essentially just became an activation and creative agency for Nike and Jordan brand. That's beside the point. But I remember so many nonprofits, so many nonprofits that did great work, right? But their websites were bad. They're flyers, which, right, flyers in the nonprofit world are still a big thing. everything's bad. Now that changes, right? So there's a huge upside to this, right? Because now all of a sudden people who are doing good work, small businesses that maybe can't afford a photographer, you know,

small to medium teams that, you know, need some new creativity. They have it now, right? So this really levels the playing field between what's possible or what used to be possible and what's possible now. But there's downsides. There's great downsides. And this is where, you know, we're talking about flux, right? Flux is an open source model, which means there's not really guardrails on it. So you've probably already seen, and if not, you're going to see a lot of presidential misinformation, disinformation. Y'all, don't send me hate mail. I'm just stating facts, right?

So as an example, you know, here in the U.S., we have Kamala Harris and Donald Trump competing to be. the president here and our election here in November. You know, the Donald Trump campaign has been using a lot of AI images, right? Which I think, you know, some people are like, oh, that's cool. It, you know, it resonates with, with young kids. And it's like, okay, yeah, sure. Also, it's extremely problematic, right? And I think the former president has caught a lot of flak online, right?

Because he created, we talked about this here on the show. he created some or his team or he's sharing a lot of these things that say like, oh, Taylor Swift is endorsing me, right? There's these AI, AI generated images of Taylor Swift saying, you know, vote for Donald Trump that look kind of real. I mean, not entirely realistic, but a little realistic, you don't know any better, right? If you don't live on the internet, if you don't follow AI every single day, you're going to think, oh, Taylor Swift is endorsing Donald Trump.

It's obviously not the case. So there's so much downside, right? Because flux, you know, you can create anything, right? These open source model. If you go in there in mid-journey, some of these other programs, ideogram, depending on the program, depending on what you're asking for, right? There's other things, you know, gore and violence, copyrighted images, right? So certain programs have guardrails and you can't generate anything.

But some of these open-source models, you can create anything, which is actually going to lead to a lot of misinformation, disinformation, not just in politics. I mean, geopolitical tensions, you know, people, you know, getting groups in trouble, right? You know, if you're a famous person, all of a sudden, you know, someone might be out there to get you. Or, you know, there's things that you can do in these, you know, image generators now where someone can take a photo of my face and essentially start using that in an image generator and make it look pretty real. So when I showed you kind of the side-by-side comparison of, hey, look how far this has come over 18 months. I think a lot of people, myself included, see the positives. Wow.

Right. This is amazing. Anyone can create anything. But then you have to think of the other side. Wow. This is terrifying. Anyone can create anything. So there's pros and there's cons.

And we're going to be seeing that and we're going to be talking about AI, because also here's the other reality. A lot of these other AI video tools, such as runway, Pika Labs, Luma Dream Machine, Kling, etc., right? As these photos get more and more realistic and it's hard to tell the difference, people are then using these AI images in these video platforms and these videos are getting even better. Right. And you're starting to see what look like videos, right?

If you see it small on social media, you might not know any better, right? Unless it's an account like, you know, an AI account like ours. And we're saying, oh, look at this, right? You might say, oh, that's AI. But you're already seeing it. A lot of misinformation and disinformation across all sides of the political aisles, across so many different realms of our world. And it's getting, it's getting hard, right?

So that's why I wanted to have this conversation. I wanted to show you both the power and the promise of these AI image, generators and how much better they're getting and even someone, if you're a beginner, you can go in there and get great results. But it's not just photos. These are being used to create videos. They're being used in great ways to now create video ad campaigns, right? That can do great things, but also they can be destructive.

All right, y'all, that's it. I know today's show, a little technical issues, but thank you for sticking around. I hope this was helpful. Make sure to sign up for our daily newsletter at your everyday AI.com. join our thanks a million giveaway. So it's super simple. Sign up takes 10 seconds. You're going to get a link.

Go send that link to your friends. When they sign up for our newsletter, you get some points. We're going to be giving away a year of chat GPT or your favorite large language model. It could even be a year of mid-jury, right? A year of whatever, right? It's on us for whoever wins this competition. So you get a link. Go share that with your friends, posted on social media, you know, send it in an email,

text it, whatever. So thank you for tuning in. Make sure to go to your everyday AI.com. And make sure to join us tomorrow and every day for more everyday AI. Thanks y'all. Meet Firefly AI assistant. Now live in Adobe Firefly, the Allman One Creative AI Studio. Just describe what you want to create in 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. Stay in 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 EverydayAI.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.