Midjourney + ChatGPT-4 = INSANE Prompts and Images!
We just had two big releases this week, with GPT-4 and now Mid-Journey V5. So, in today’s video, I want to explore if this can be the future of photography, and I will also show you how you can prime GPT for getting into the role of creating prompts for Mid-Journey V5. So, let’s just get started.
The first thing I’m priming ChatGPT-4 for Mid-Journey V5 prompts. I want to show you is how I prime GPT-4 into understanding how Mid-Journey works and how I want it to perform. So, basically, I just started here with, “Hello, today I want to create images with a diffusion model. I’m gonna feed you some information about it, okay?”
So, that is good. The first thing I did was I just went to Mid-Journey and copied all of this data on how it worked, but I didn’t want any response, so I just said, “But wait, I have more info, just answer with ‘read’, right?” So we get to read back. Then I fed it all the information I can find on the newest model, that is version 5, right? I just found this on Mid-Journey’s homepage. So, I put all of this into GPT-4, right? And again, just answer with ‘read’, so I got the response. And I followed up with how you can set quality chaos seed, stop, all the parameters you can use in Mid-Journey so the ChatGPT knows that again, I just want the real answer. I’ve added some compatibility’s information, read.
So, now then, I followed up. Now, I will give you some examples of prompt used in Mid-Journey V5. Okay, so I went ahead and found prompts I liked on the Mid-Journey homepage focused on photography since that is our main focus today. So, I just pasted in a whole bunch of examples here. I think it was five. And I followed up with, “Would you say now that you understand how Mid-Journey works, yes or no?”
“Yes, I understand how Mid-Journey works now.”
Okay, so that’s good. Then I followed up with five more prompts. All of this has some kind of photography-related prompts because that’s what I want to do in this video. And again, I think I gave it six more different prompts. I just want to fill it with a lot of good examples of things I like. So, I think that’s pretty important. And then I’m ready to just prime it. So, basically, now I just go, “Great! Now I want you to act as a professional photographer. You will use a rich and descriptive language when describing your photo prompts. Include camera setups. The first prompt I want you to create is a photo of a 1930s female influencer. Take inspiration from the format and formatting from the example prompts. Don’t copy them but use the same format.”
Okay, so that is basically how I prime GPT-4 to come up with Mid-Journey prompts I am happy with. If you want the whole prompt I just used to prime GPT-4 to get ready for Mid-Journey, follow the link in the description and sign up for my newsletter. You will find a PDF there with the whole prompt inside. And now, it’s basically just I just hit submit, and this is the first prompt we got.
So, basically, let’s just copy all of this and head over to Mid-Journey and check out our first image. So, over here at Mid-GPT-4 + Mid-Journey V5 Photos Journey, I’m just gonna paste the prompt as it came out from GPT-4. I’m not gonna do any changes. So now, it’s just to hit enter.
Okay, so we got our first result back. Let’s take a closer look here. So, you can see these were in the 4×5 format, right? Wow, this is amazing. The fingers are still messed up; you can see she has one extra finger here. That’s a bit funny, but other than that, look at this.
Remember, this is a 1930s female influencer, so the camera is probably influenced by that, and I like that it’s an old camera. That’s great, and this looks very realistic, right? Also, this pose here, they’re all posing for a camera, I think. Yeah, this looks awesome. Very happy with this.
So, let’s take a look at something I wanted to show you here. So, I just created this quick photo list of photos I want to do today. So, this is the 1930s female influencer. Next, I want to do something with Vikings, as you probably saw in the thumbnail. That was a Viking prompt, so let’s just do that.
So, I think the first Viking prompt I want to do is: “Now generate a prompt of a photorealistic portrait of a screaming female Viking.” Let’s try that.
Okay, so that was the prompt. Let’s just take a deeper look into our first prompt. I’m not gonna do this for everyone, but you can see a powerful, vividly realistic portrait of a fierce female Viking warrior mid battle cry. Her face contorted with raw emotion and determination as she fearlessly leads her comrades into battle. That’s a great sentence, right? And you can see all of these camera settings that we asked for. We have the aperture, we have the lenses, we have the ISO, we have the shutter speed, and it also adds the aspect ratios, the quality, right? That is probably from our examples, so I think that’s very important too. The shallow depth of field skillfully isolates the warrior from the chaotic battlefield in the background. Wow, I’m excited for this prompt.
So you can see, I’m leaving out V5 because that’s automatic. So, I’m gonna copy all of this and head over to Mid Journey. Wow, this is insane. Look at this. This is crazy. So where was this? Wow, look at this. Uh, I might remember my camera is in the way. Let me move this. This is insane. Look at the intensity in the eyes, and look at the details. Oh my God, look at this. I’m kind of scared looking at it. Damn, that’s impressive. I’m speechless right now. This was—I didn’t expect this kind of quality. Amazing, but we gotta move on. So let’s head over to our next photo.
Now, I wanted to do a prompt of the perfect pasta dish, so let me just run that.
Okay, so here we have the prompt. I just wanted to point out one thing. So you can see we don’t mention anything of a photo here, but it still understands from the examples I gave it with the priming, right, that it should be a photograph. And it also picks the camera aperture, lenses, shutter speed, and everything we need in a photo. So that is very good. So let’s just go ahead and run this.
Okay, so here we have it. Here is our pasta dish, and damn, that looks good. I gotta say, that looks very yummy. Let’s just blow that up. Look at the details here. You can see the tomato; you can see the Parmesan. I can’t really not understand that this can—this has to be able to replace some kind of food photograph photography. Isn’t that really expensive to hire people to capture food? So what if you just didn’t—you just took an image of your food on your restaurant and fed this into the image-to-image, and then you get like an expanded or better version of it? Look at the details here; it’s just crazy, with the flower or what it is on the side here. You can see some kind of Parmesan has fallen off the plate.
I’m just amazed by this. This is just insane. This is, I have to say, this looks really Next Level. Very impressive. So, what I thought we can do now instead of me going through each prompt, what I’m gonna do now is just, I’m gonna go to my photo list here, hope you can see this, and I’m just going to collect a whole lot of different prompts. But before I do that, I just wanted to show you one more thing, and that is, uh, I want to show you something I tried that worked pretty well, and that is changing up the input prompt here to now generate a prompt of your interpretation of, and then you just add something at the end here. I’m just gonna run that and I’m gonna collect a whole lot of prompts from this list and then we’re just gonna look at all of them in mid-Journey, okay? So now we have every single image I wanted to collect. So let’s just go over them one by one.
So, the first one we have is the prompt was a Viking shape sharpening his blade. So let’s take a look at this. Look how realistic this is. This is just amazing. I kind of can’t get over how good the hair looks, right? That’s just, it looks perfect. I guess the weapon and the blades are a bit strange, but it’s so realistic. Look, the headband here is a bit off, but other than that, and look at the background, it’s blurry. This is just, again, look at that hair. That scar might be a bit off, but look at that hair, how good it looks. The fur is amazing. I’m just shocked how good this is. The fingers, at least it has four fingers, but the skin looks a bit off, but I’m kind of impressed by this. Wow, it’s gonna be a lot of fun playing around with this. Okay, so let’s move on to our next image, and that is going to be a gladiator hunting in the forest. So let’s blow this up again. You can see it’s a bit off. I kind of like the ground here, but the face is kind of messed up. This was quite cool, but the sword, again, it’s messed up. The shield has a blade, but I like the forest, that’s good. This may be a bit off again, that looks good. The shield is maybe a bit off and there’s something wrong on his head, so it’s not perfect, but quite good. I think this is the best one. It’s black and white, though, but if it wasn’t for that bug with the blade, I think this will be perfect, kind of weird with the hat and stuff, so if I wanted the perfect image of this, I kind of had to do some reruns, I think.
Okay, so let’s have a look at the next one. This is a good one. This is a perfect moment when a bee lands on a flower, and look at these details. What can I say? Look at the pollen here and look at the details inside the bee’s eyes here and the pollen and the fur. You can really see the fur here, that’s just, I can’t really believe this is AI-generated. Look how good the fur looks here. This is a pure macro photo. Oh, that’s good. I can’t believe how good the fur looks, amazing. Okay, let’s move on. And this was that GPT-4’s interpretation of dreaming, so let’s blow it up. I gotta say, I really like this, it’s just so wholesome, it’s so peaceful. There’s one quite gross thing here, just three feet, so let’s just move on from that. But look at this one, look how peaceful this is, this too, the hands are a bit weird and there’s something here, but this, I gotta say, was my favorite. That’s just great. Remember, this was just the prompt was, “Give me your interpretation of dreaming,” right?
Okay, so this was a good one. This was your interpretation of artificial intelligence ruling over humans. This image is cool, right? There’s like a big army of, is it humans or is it AIs? I think it’s AIs or some robots. This was great, this looks like from a movie, right? You see that? This one is guarding over all these humans here, this was quite cool, I guess there are humans in the background, it’s a bit strange, but I like the cyborg robot here. This is also very, it’s a bit creepy, but still, this I just did this image, I didn’t redo any, so that’s cool. Okay, so move on to this was a 1950s female police officer, so let’s blow this up. Perfect, how good is this? You can see the eye is a bit off, maybe other than that, there’s really nothing to, there’s no flaws here, is it? I can’t really see anything else that’s flawed here, the lips are perfect, the eyes are good, just looking that way, the hat is very good, of course, the text over here is a bit messy, but yeah, this is awesome.
Okay, so let’s check our next prompt. This was a mysterious cabin by the Fjord in Norway, so I can kind of relate to this. I think this one was the best one because this is usually the style of cabins we have here. So let me take a look at this one, yeah, this too is quite good. It’s not a typical Fjord setup, I guess it kinda is still, but this cabin is rarely seen this, but it’s still good. I like that three are green but on the other side, they are snowy, that’s a bit strange. Maybe this one is good, the wall is a bit strange, but other than that perfect, the mountains just exactly as we have it here and I like the reflection, right? The reflection is perfect down in the water on all images, that’s very good, nice.
So, this was Star Wars-inspired living room interior. Remember, I did a video on this with me Journey 4. Uh, I don’t know if this is improved, I think there are more details here, and it looks great. Just think of how many architects or interior designers can take advantage of this now. I know I’ve done a few videos on this, but look at this. This was just the only prompt I ran and it’s just awesome, crazy. Okay, so the final one we have on this video is an Iron Man-inspired sports car. Let’s blow this up. I like the one in the right corner here, the colors are spot on, the on the left top corner too, that’s very good. I think I read somewhere that some car designers now are using AI to get the inspiration and you can kind of see why because you can create hundreds of this just to get some inspiration, wow, crazy. This was a bit strange that the headlights are in the back and the front, that’s a bit strange. Other than that, I think this one was the coolest, awesome.
So, just a quick word on my first impression of Mid Journey Five, uh, I gotta say I’m really impressed. I can’t get over how good it looks. I think my favorite image has to be the Viking girl. I can’t get over how good this looks.
It’s so realistic; it’s almost scary. I guess the tongue here is a bit off, but yeah, my first impression is very good, and I can’t wait to get around to playing more with it. And I think combining this with GPT-4 can get you some really good prompts, so I suggest you try it out.
So, like I said in the middle of the video, if you want the prompts I use, the Prime GPT-4, just sign up for my newsletter. You can also check out my membership in the description below. Also, check out this video here if you enjoyed this one. I think you’re gonna like that. And as always, thank you for tuning in. Have a great day, and I’ll see you in the next one.