MidJourney
is a generative artificial intelligence program and service.
While it is an incredibly powerful tool that is constantly evolving while offering levels of creativity and opportunity like never before. It is without a doubt a disruptive technology. It has changed how we all think about what can and can’t be done.
*Note: Everything on this particular page was created with MidJourney AI v.6.0
Generative text-to-image Artificial Intelligence.
Midjourney AI is just one of a slew of generative art focused AI that creates images from text prompts. Having entered public open beta on July 12, 2022, it has since been joined by many others such as DALL-E and Stable Diffusion and with each new version AI improves its capacity for for more accurate and more realistic output.
Generative AI has made it practically impossible to tell the difference between what is and is not real in only two years, in another two years many if not most of the challenges AI struggles with such as creating realistic hands will be corrected.
Prompt writing and setting parameters are what allow MidJouney to create image. Take for example this prompt from Brady Lewis at Medium.com “A breathtaking aerial view of a sprawling futuristic cityscape, with towering skyscrapers piercing the sky, sleek hover cars zooming through futuristic highways, and vibrant holographic billboards illuminating the streets below.” will produce different images each time, but they all will be similar in theme and expression according to the prompt. Parameters such as “--s 100” or “--c 6” will tell MidJouney AI to add more levels of interpretation of style or variations to the final rendered image.
Make it stand out with descriptive language.
Contrary to popular belief, creating great looking images with AI still takes time. It is an iterative process filled with creative writing, problem solving and research. Why? It comes down to knowing what is needed to be described for MidJourney to create the best image. For example, the prompt writer needs to know the difference between “Boho-chic style flowy earthy tone dress with lots of patterns that is loose but still form fitting.” and “1980’s grunge outfit with stone washed blue ripped jeans and a black and red flannel shirt”. Descriptive language still takes some work, trial and error, and research to render. Precise and accurate descriptions allow MidJourney to more accurately convey the image in the mind of the writer.
Descriptive language extends to backgrounds, camera angle, architecture, photographic and cinematic elements, accessories, vehicles, and much more. The more descriptive a prompt is, the higher degree generative AI can compose an image that stands out. This is where research can make a huge difference, with the language used to describe an idea. The other huge factor is how the AI is trained to understand ideas, concepts and examples of space, photography styles, drawing and artistic styles, lighting and so forth.
