Goodmorning, Normal People!

Yesterday we took a quick look at the quickly closing gap between real and generated images.

I mentioned that my belief is the best way to understand these models is to use them for yourself, and today we’ll do just that. As promised

📓 A Normal Guide to creating your own AI-generated images with Midjourney:

1. Join Midjourney's Discord Server: For now, this model is technically only available in a ‘beta’ form through a Discord server. To get an invite, head to the Midjourney website and click the “Join the Beta” button at the bottom right. If you don’t have a Discord account already, you’ll be walked through setting that up as well. Welcome to the club.

2. Get Your Bearings: Once you’re in the Discord, you’ll need to access the right channel before you can begin generating your own images. For now, navigate to any of the open #newbie channels on the lefthand side of your screen.

3. Command the Bot: The way this works is that a bot is assigned to each channel, and you generate an image by sending a message in that channel. But in order for the bot to recognize your text as a prompt, you have to start it with a command – in this case, the command is /imagine <–– literally type that into the beginning of your text box, but don’t send yet!

4. Write Your Prompt: Now for the fun part. After your /imagine prompt, you can type just about anything. There are a million parameters to tweak and hacks for manipulating the image, but for your first time I would encourage you just to start typing and see what happens.

5. Patience, Young Padawan: After you hit enter, the bot will pass that prompt along to the Midjourney model and begin to create four options for your image. This will take a minute or two, but once it is done you will be tagged and have access to your image.

6. Voila!: You’ve done it, your first official AI artwork! You can save the grid of four images right to your device! If you want to keep iterating, be sure to type the /imagine command again before you try another prompt.

7. UV: The eight buttons below refer to Upscale and Variations. Say the second image is exactly what you wanted – tap “U2” and that image will appear as a single, higher res image in seconds. Or maybe the fourth image is almost what you wanted, but you want to keep exploring that direction. Tap “V4” to get four images with the same text prompt but with that image as a starting point visually.

As you begin to learn more about the model try searching for styles and prompts to generate wildly different kinds of images. For some scary realistic results, start your prompt with “iphone photo” and end them with “—v 6.0 —style raw” (that’s two dashes each time). Happy generating. 🤖 🖼️

Onward,

Brady Fowlkes

Subscriber count 👉 95

We’re trying to get to 250 by end of January, and we need your help to get the word out! 🗣️

Do you know any other normal people?!
Share this sign-up link with them today!