My First Time Using an Insanely Powerful AI Concept Art Tool

So only after a week following my application for the beta run of Midjourney AI I was granted access to a vibrant world of artistic possibilities. For those unaware Midjourney AI is a super powerful ai-driven art too. It uses very clever machine learning to create surreal art based on a wide aggregate of image sources being fed to it based on a growing network of current users and the wider web.

Users simply fire a random and descriptive string of prompts such as "Human Torch flying through space fighting off an army of yetis" then the AI quickly pulls a number of blended visual variations together based on those descriptions. The results can vary from extraordinarily creative all the way to the down-right bizarre.

I had initially expected to have to wait several weeks to be accepted due to the high volume of keen creators waiting to give it a spin. The hype is very real.

Here is a quick test run of my first time using the AI. Prompt: "80s action hero fighting off neon tigers in Drew Struzan painting style". The results below were astounding to say the least:

What I found most fascinating was a limitation in its process that I found right out the gate. When you ask it to separate two oposing things such as "man fights tiger" what you often get in the variations is a forced blending of the two concepts. Hence I was given given a few horrific half-man/tiger hybrids which would feel right at home in a John Carpenter film. However the coolest merging occured in the image above whereas our 80s action hero appears to be sporting a tiger jacket engaged in battle. I presume he is fighting tigers due to the deep massive claw wounds which surprisingly appear to be oozing neon instead of blood.

In my last article I really emphasized how this tool will be wonderful for early stage concept art because I hadn't intitally thought to have taken the concept in this uniquely cool direction.

Prompt:  "T-Rex evolving dramatically, drastically and violently into the everyday chicken"

Prompt: "Massive Olmec stone head flying through galaxy in 80's airbrush painting style"

From even the shortest and vaguest of descriptions or hyper-specific ones it does quite a fine job of articulating these concepts to a high resolution and accuracy. While this tool is still in Beta the team are certainly finding more ways to broaden its abilities beyond the realm of abstract styles of art.

So it would seem even its current limitations still present very awesome opportunties for unique ideas and the sheer speed of generating these concepts under 60 seconds is an absolute thrill. The Midjourney Discord group is a rapid stream of outlandish prompts producing wonderfully bizarre and beautiful concepts at the speed of imagination.

AI art is still in its very early days. I'm certain there's great potential here for creators and will assist a healthly democratization of art for independent users in the same way NFT's continue to disrupt the norm.

I'm keen to see what you guys create :)