Neural networks that draw by words are one of the biggest trends of 2022. It all started in April when OpenAI introduced the DALL-E 2 neural network, which is capable of creating high-quality images based on text descriptions. The images it created were so popular with a society that one of them even made the cover of Cosmopolitan magazine. Unfortunately, the DALL-E 2 neural network is only available to a limited number of people, but that is not a problem, because there is a Midjourney neural network for the general public. With it, anyone can create a real work of art, just by telling the neural network what to do.
Let’s find out how to use Midjourney and what masterpieces have already been created with it.
How the Midjourney Neural Network Works
The Midjourney ai neural network exists in the form of a bot for the Discord messenger. The principle of its work is very simple: the user enters a description of the desired picture and gets a ready result.
At first, the neural network offers four options to choose from – a person can choose one of them and make a choice: accept the result, or continue editing with additional keywords.
How to use Midjourney – A Quick Start Guide
Since the neural network is only available in Discord, you must first register with this service
After that you need to do the next steps:
- go to the official Midjourney website and click “join the beta“;
- sign in to your Discord account and give the bot all the permissions;
- find the channels named “newbies” on the left side of the screen. There are many of them, you need to find a free one;
- at the bottom of the screen, find the text box, write “/imagine promt”, then enter a description of what you want to see in the picture.
As an example, you can write “/imagine promt red cat with blue eyes”. After that, the neural network will create four variants of the picture – the one you like can be selected with buttons V1, V2, V3 and V4. After that, you can improve the selected picture by writing more qualifying words. When the most expensive variants appear, you should press buttons U1, U2, U3, or U4 and download the best picture.
How to formulate Midjourney requests
It is better to write descriptions of pictures in English. It is important to understand that the neural network likes accuracy – as an example, nothing good will come out of a phrase like “an ordinary cat”, because the computer does not know what is “extraordinary” to you. It’s better to write clearly “red cat with blue eyes looking at the moon sideways view”. And all this in English.
It’s important to note that the Midjourney neural network can only be used 25 times, and it’s not about pictures, it’s about actions. Once you’ve reached your limit, the bot will offer to subscribe – the paid version of Midjourney costs $10 or $30.
Midjourney ai winning art competition painting
The Midjourney ai neural network has been in existence for several months. During that time, users have been able to create millions of different images. One of them was even talked about in the news because it helped the user gain fame.
In September 2022, the news that a drawing created by Mid journey neural network won an art contest was actively discussed on the Internet. User Jason Allen, nicknamed Sincarnate, printed out a picture created by the neural network and submitted it to the contest – the work won first place and made the author a celebrity.
It turned out that the judges had no idea that the picture was painted by a neural network. When the truth came out, people began to criticize Jason Allen – allegedly, he won unfairly. But the man assures that he spent at least 80 hours creating the image. To get such beauty, the man had to look through about 900 versions of the picture and choose the best of them. Then he enlarged the resolution of the image with Gigapixel AI and printed it on canvas.
Examples of Midjourney’s work
Many other images created by Midjourney’s neural network can be found online. Especially noteworthy are the works published in the official Midjourney community on Reddit. Below I offer a look at the ten best works since the group’s existence – it’s very young.
What is beauty according to the neural network?
The image below shows what the neural network thinks is beautiful. Well, Mid journey seems to have great taste!
Caveman selfies
And what would cavemen do if they found a modern smartphone and learned how to use it? That’s right – selfies, and that’s what it would look like.
“The Last Supper” in cyberpunk style
One Reddit user wondered: what would Leonardo da Vinci’s The Last Supper look like in cyberpunk style? Here’s the result.
A vintage map with monsters
A vintage map with sea monsters as the continents. It seems delusional, but the Midjourney neural network was able to visualize such a query.
Portal to the Past
The author of the following image assures that he hasn’t gone to work for three days – so much so that the neural network pulled him in.
Comic Book by Neural Network
One Reddit user is using Midjourney ai to create a comic book. Here’s one of the pages – note how he manages to keep the same drawing style.
The Hidden Gods
The next painting is called “Hidden Gods. I think it can be sent to exhibitions.
Generation Z on the French Revolution
What would a member of Generation Z (zoomers) have looked like during the French Revolution? Like this.
Portraits of people from a neural network
Midjourney’s neural network has cracked the sinister valley effect – people’s faces look like the real thing.
A photorealistic portrait of a girl
The following image shows how good Midjourney is at photorealism.
Have you tried the Midjourney ai yet? What do you think about it?
This article was inspired by the author