I have carried on experimenting with the wonderful Midjourney app on Discord. My previous post had some images and a degree of explanation, but as a reminder. This is one of the new generation of publicly available of AI based image generation from text. Just watching the flow of requests and the range of styles and images that appear constantly is fascinating. It generates 4 images for each text request which you can then curate and push for further generation of more similar to that one. It can also take a web image not generated by it as inspiration.
This however, made me chortle, then stop to ponder, then simply just say way. I said /Imagine Lego John Wick – that is all I asked
Look at the variety of styles from the top right and bottom left as actual minifies with shadows to a stylised top left looking as if he is paid in lego pieces not gold, and bottom right an angular representation but a cool image.
Then of course that got me going onto the obvious /imagine Lego predator
This just looks great, lots of predator variations, which as has happened with the films is pretty much in the storyline to have multiple tribes. (BTW Watch Prey the latest Predator movie on Disney+/Hulu it is really good !
Once again Midjourney hit me with its range of styles it delves into. I moved off the lego thing for a while and asked to see Laurel and Hardy breakdancing
This is a second generation image from a set of already very good images. Laurel and Hardy are of course well known, but there are no breakdancing scenes to refer to. Also, it chose black and white, presumably because most of the source content was black and white. The quirky style of this compared to the photo realistic lego minifigs and the stacks of other images show how amazingly versatile this approach is. You should be able to get a glimpse of the madness on my public facing page in midjourney (if that works )so I can stop posting here all the time saying “look at this: 🙂 You may need to change the second in from the right navigation drop down to grids or all, it seems to default to the large upscaled individual images.
Bring on virtual world/metaverse generation.