Midjourney released a new version a few days before Christmas, and I spent the rest of the year trying it out. It’s an alpha version, meaning that it’s extremely experimental and could change at a moment’s notice.
I quickly realized that it’s very, very good at creating images in a photographic style, such as the snowflake above and the soap bubble with ice crystals below.
I’ve always liked the grunge aesthetic, so I tried to make a wall with peeling paint. Of course I also had to see what the new version would do with a Día de los Muertos theme. I was surprised to see the skull I’d requested protruding out of the wall instead of looking like a two-dimensional mural or painting … but I liked it quite a bit.
I wondered if a sun face with some Mayan or Aztec influences might also look good on an old wall.
Additionally, Midjourney’s new version seems to do fairly well when one requests an image “in the style of” a famous artist. For the image below, I requested a cat in the style of Pablo Picasso and specified the Fibonacci sequence as well.1
Right around the same time Midjourney released their new version, Microsoft’s search engine, Bing, began making free (though limited) AI image generation available via DALL-E. DALL-E’s been available to the public for just about as long as Midjourney but there were two factors way back then2 that influenced me in favor of Midjourney: 1) I could make as many images as I wanted for a flat fee per month and 2) they could be square or rectangular. Neither was true of DALL-E.
I tried using exactly the same prompt for the DALL-E image above as I used for the 3D skull further up: “stucco wall, peeling paint, distressed, dia de los muertos graffiti.” In many ways, the DALL-E image is closer to what I’d originally imagined. However, DALL-E still makes only square images (I cropped the image above).
DALL-E is available to anyone who’s interested via Bing.3 I’m encouraging you to go to www.bing.com/create and try it — I think they may give you 15 free prompts every 24 hours. And, since you get four images for each prompt, you might very well have a decent library of AI images by this time tomorrow.
Can’t wait to hear about your AI image generation adventures.
I’m fascinated by how the Fibonacci sequence is visually represented.
Way back then = August of 2022, almost a full year and a half ago.
If you’re on a Windows computer, it works best to use the Edge browser. If not, I know for sure it’ll work in Chrome — although it’s definitely slower. There’s also a Bing app available for mobile devices.