What would nations look like if they were supervillains according to the AI
For some time now, the phenomenon of generative artificial intelligence has increasingly conquered the community of artists and beyond. Lucas Freitas, in fact, content creator on Tik Tok , recently posted images generated with the Midjourney AI themed nations of the world transformed into super villains. In total 39 images depicting what are the peculiar characteristics of the various states if they were cinematic villains (see gallery below). Everyone, more or less, has tried this type of software to edit and make our photos artistic, for example. However, Midjourney's artificial intelligence is of the Text To Image type and generates images from text input.
Midjourney, the generative artificial intelligence
Midjourney is an independent research laboratory exploring new means of thinking and expanding the imaginative powers of the human species.
This is what is written on the official website of the platform by clicking on the item "Getting started", to then explain how to use it. In fact, the page goes on to say that it is possible to use the tool in two different ways. One is through the Midjourney Bot, which is used to generate images, the other through the web app of the site itself. At the same address it is possible to find a gallery with our works and those of other users.
Artificial intelligence services can also be accessed via the official Discord server or others set up for the purpose . In fact, via an invitation link, if you wish to use the Midjourney Bot, you can follow a specific automatic procedure. Obviously, in order to regulate the service, users must follow a certain code of conduct. The site itself advises against using the tool to create images that are likely to upset or cause drama. And by that we also mean gore and adult content.
The use of Midjourney and the tools behind generative artificial intelligence
Generating images With the Midjourney Bot it is quite simple and immediate. After registering with an account the site will redirect to the Discord server where you can join different conversations for "newbies". From here it will be enough to type in the text box what we want to generate, providing as many details as possible.
The command takes the form, /imagine prompt {PROMPT TEXT} , with users able to set various flags. By adding other settings you will get an image as specific as possible to the given description. The site offers a limited amount of free usage per month, after which you need to subscribe. Available plans are $10, $30, and $60 per month. But there is a discount if you choose to bill annually.
The technology behind the "social network for images", as Midjourney is defined, is the one developed by OpenAI. Indeed, its core has a neural network called CLIP that efficiently learns visual concepts from natural language processing . CLIP is trained on a wide variety of images capable of great natural language processing, abundantly available on the Internet.
This neural network model is designed to evaluate how well the generated images align with textual descriptions. Upon its release, artist Ryan Murdock (@advadnoun on Twitter) discovered that the process could be reversed. By providing text input then, image output can be achieved with the help of other AI models. Like the famous DALL-E 2, Midjourney is the result of large AI models trained on a large number of images. But unlike the latter some users say that Midjourney has its own distinctive style.
The mind behind the project
Service co-founder David Holz grew up in Florida and studied math and physics. He was working on a PhD in applied mathematics and took a leave of absence in 2008 to start the company Leap Motion. The following year, he spent a stint as a research student at the Max Planck Institute, followed by two years at NASA Langley Research Center as a research scientist working on LiDAR, Mars missions, and atmospheric sciences.
But all this did not satisfy him and the desire to work on interesting projects that he cared about grew stronger. Then he focused on Leap Motion by developing a hardware device to track hand movement. He ran the company for twelve years and employed around 100 people when he left.
Midjourney, she says, is pretty small right now. But Holz also says the technological aspect of AI and the extent to which it will improve is fairly easy to predict . To really understand what it is and what it should be, there really is a need to do a lot of experiments. Meanwhile, we enjoy what it has to offer, such as the cover image of the article and the gallery below. Obviously also created with generative artificial intelligence.
The article How nations would look like if they were supervillains according to AI was written on: Tech CuE | Close-up Engineering .