Neural networks are already being used to create concepts and illustrations, and some of the results are very good – see what we do done halfway through the trip. However, the possibilities of AI don’t end there. It can be used to create music, lyrics, for work tasks. And in the future, with the help of AI, it will be possible to make full-fledged games on their own, much faster than it is possible for individual developers to do now.
An example of this is Nvidia’s GET3D AI, which allows you to create three-dimensional objects and characters, and the speed is impressive – around 20 objects per second.
Researchers trained this AI on synthetic 2D images of 3D figures from various angles. It took about two days to upload a million images to GET3D and process them on the A100.
Now the model can generate not only high-quality complex models, but also textures for them.
AI users will be able to quickly import items into game engines and can also be used by 3D artists and other industry professionals. What does this mean for video games? The fact that virtual worlds can be filled much faster with objects, animals and people.
And with the help GAN-NADA style you can apply different styles using text commands. From a model house, you can turn it into an abandoned haunted mansion, turn animals into monsters, and give objects “character”.
So, if you’re thinking of getting into game development, this is a great time to start.
–
