I wanted to dedicate this edition to showcase various industries where an image generation application such as Midjourney can be applied.
» Background information: Midjourney is a generative artificial intelligence program that generates images from natural language descriptions (“prompts”). In order to use Midjourney, you’ll need to join their Discord server. If you don’t have a Discord account, you’ll be able to register for free when you join.
Here’s 4 places where I can see AI image generation being used:
Marketing & Advertising: Creating a template for ads
Journalism, Blogging, & Newsletter: Thumbnail creation
SaaS and Small Businesses: Creating a simple logo
Education: Exploration of combined art styles
As always, remember that AI should be used to support your work, not do the work on your behalf.
Case #1: Templates for ads
If your Photoshop skills are sub-par (like mine), you can leverage Midjourney to create a base image for you. Then using your “good enough to go on a resume” Photoshop skills (free version: Photopea), you’ll be able to perform minor touch-ups and add text.
Here’s one I generated a while back for an abstract art museum advertisement (fun fact: this was supposed to be an ad for programming!) with the minor touch-ups:
Midjourney generated template
Edited AI generated image
Prompt: Text: Font: Bold, clean, sans-serif font Colors: light green and celestial blue with a strong contrast between the text and the background for clarity with light yellow undertones Graphics: coding images, geometrical abstract with a semi - transparency overlay on a background image relating to programming.
Case #2: Creating Thumbnails
When I started my YouTube channel, I spent a lot of time creating eye-catching thumbnails - upwards of 2 or 3 hours!
Midjourney shortened this time to ~10 minutes. This gives me back A LOT of time to focus on my writing or do other things that needs my attention. Here’s this editions thumbnail with prompt:
Midjourney generated thumbnail
Prompt: papercut style, coffee shop with positive, relaxing vibes, colorful --ar 2:1
You may have noticed that I appended --ar 2:1 to the prompt. This is the aspect ratio option, which tells Midjourney the relationship between the length and width of the image.
Case #3: Creating a Logo
SaaS business owners, here’s one for you.
Prompt Midjourney to create a logo with your preferences (shape, style, color scheme) and it will generate 4 different logos. You can regenerate using the same parameters or with a new prompt to see which one(s) you like the best.
Did you know that the Bytes and Brew logo was created by Midjourney? It took me 4 re-generations and 3 different prompts.
Midjourney generated logo base
Prompt: A logo for a newsletter surrounding AI applications for programming and programming education. The main subject should be a robot in an isometric view with a cup of coffee and a news paper in its hand. The background should be white, and the subject should be happy to drink the cup of coffee. Enhance the shadows with high contrast for a captivating look.
Case #4: Art Education
Instead of just offering theoretical knowledge, educators can leverage Midjourney to provide a hands-on experience with blending art styles together. This, in turn, fosters a deeper understanding and appreciation for the style's features and nuances.
Prompt: Blend the swirling patterns and vibrant colors of Van Gogh's post-impressionism, fragmented forms of Futurism::1.5, elements of Surrealism.
Here, I pulled a little trick to tell Midjourney that I wanted a certain aspect of the image to be more heavily “weighted” towards one part of the prompt. This can be achieved with the “::” characters, followed by a number.
In the prompt, I wanted to have the futurism style stand out more, so I told Midjourney to weigh it heavier (
If I wrote a guide on prompting with Midjourney, would you be interested in reading it?
This guide would include best practices, prompting techniques, listing of art styles, and sample prompts with images.
Login or Subscribe to participate in polls.
Generative Adversarial Networks (GANs) made a significant breakthrough in 2014, which provided the ground work for AI image generation. However, very early work dates back to nearly 40 years earlier to the 1970’s.
Unfortunately, advances in this field during this time period were significantly capped due to the lack of the necessary technology. For example, there wasn’t enough computing power to run these models, coupled with a shortage of training data.