Google introduces text-to-video game tool Genie

Google's DeepMind division just introduced Genie, a tool to create video games using only images or text descriptions.

Genie was developed by Google DeepMind's Open-Endedness division and was trained on a dataset of 200,000 unlabeled videos, mainly from 2D gamers. Unlike traditional AI models, Genie learns by observing actions and interactions right from inside the video. It is this difference that helps it create video games from a single command or image.

Google introduces text-to-video game tool Genie Picture 1Google introduces text-to-video game tool Genie Picture 1

Google said that Genie is still in development so there are still some limitations in graphics quality, currently only creating games with a low frame rate of 1fps. Currently, Genie is still a research project within Google DeepMind for research purposes and is not applicable to the public.

News site LifeHacker commented that Genie's video games look quite bad, the movements are jagged, blurry, and there aren't too many things for gamers to manipulate.

Currently, the global gaming industry is experiencing new rounds of layoffs. On February 27, Sony Interactive Entertainment - the company that produces PlayStation - announced the layoff of 8% of its workforce, equivalent to 900 employees, to restructure its business in the face of an ever-changing economic environment. Meanwhile, the PlayStation studio in London (UK) will also be permanently closed.

A day earlier, game developer Supermassive Games also announced a series of layoffs. In January, Microsoft Gaming – one of Sony PlayStation's main competitors – said it would lay off about 1,900 employees.

In 2023 alone, about 6,500 employees in the video game industry will be laid off, but the actual number may be much higher.

4.5 ★ | 2 Vote