Over the past few weeks, I have been constantly seeing in the news in the memories of the Midjourney neural network. This artificial intelligence creates stunning paintings based on textual sources, and the final result can easily be confused with the work of some professional artist.
The only difference is that the generation of downloadable images requires neural networks seconds.
Right now Midjourney is in beta and is just for fun. But a few days ago, a Japanese indie game developer showed off their new scrolling shooter, in which all the visuals, including spaceship models and backgrounds, were created by this neural network.
Not only that, Midjourney even came up with the name of this game.
Shun. The first game made by a man and a neural network
The game is a simple 2D shooter where you control a space fighter. During an endless flight in the sun, hundreds of different alien ships that need to be shot, dodging their projectiles along the way.
You can play Shoon right in your browser window, just follow this link. It took only two days for the developer with the nickname Nao_u to create the game.
And most of this time was spent programming. So when you play Shoon, remember that in a few seconds by text request of the creator of this shooter. The programmer only needs to assemble the finished images into a single whole.
“Starwars x-wing space fighter flies from left to right in black space, armored core for response, high quality”
According to the game developer, the greatest clarity was a clear and well-formulated text image of the request, which one would like to receive in the end. But if it succeeds, the neural network provides just an inexhaustible number of different variations. You just need to choose the most peeking option.
As he created visual formations, Nao_u ran into contemporary difficulties. In particular, artificial intelligence is not able to generate an image with a “top view”, therefore, initially, from the idea of \u200b\u200bcreating a vertical shooter, like Sky Force, I chose the choice.
And still, the result was worthy. Yes, the neural network does not know how to create live images, but this is only a matter of time.
In fact, artists and animators write off early, but the Nao_u experiment proved that artificial intelligence can already help game users, graphic developers, and all those who invent and draw well.
For example, right now the guy creates comics, Midjourney draws pictures for fans. It turns out very well:
Apparently, we are on the threshold of a new era, when artificial intelligence becomes a real assistant to a person who does not perform tasks on preliminary tasks that arise in a complex matter, like creativity. Writers and artists will look for inspiration in the generated images to surprise us ordinary users with amazing landscapes, places and locations.
And then – who knows what it will lead to. Maybe in a couple of years we will see the first iPhone, the design of the body or the body of the system that the neural network generates for us?
Source: Iphones RU