How far can mankind develop when technology is disclosed for free? Stable Diffusion, an AI image generation program, was released as a complete open source on August 22. Anyone can use this technology and make new modifications in their area.

Less than two months after the release of the Stable Diffusion, it was enough to show the power of the “open source.” On the 18th, Stability AI, the company that developed the program, received investment worth $111 million (about 145.3 billion won) and became a unicorn company (startup with a corporate value of more than $1 billion). More than 200,000 developers downloaded the Stable Diffusion model, and more than 1 million general users.

The power of open source is ‘expandability’. There are several AI programs that convert letters into images, but Stability AI’s stable diffusion is the only one that has released all the source codes. Since the release of Stable Diffusion Source, various attempts have been made around the world, ranging from personal-tasting image production to medical image processing.

The most famous example in Korea is ‘Nobel AI’. Nobel AI, which was released on the 3rd, is a program specialized in illustrations such as cartoons and animations, and is showing great interest in the subculture industry by creating high-quality paintings without any sense of incompatibility. Nobel AI learns and generates cartoon images by applying the stable diffusion model, and provides subscription-type services. Users can see the picture they want by simply typing in the words “Girl with Cat Ears.”

It also saves designers’ trouble by linking stable diffusion to existing programs. Christian Cantrell, a former Adobe developer, introduced how to link Photoshop and Stable Diffusion on his Twitter account on the 9th of last month. Just turn on the Photoshop window and write down the image you want in letters, and you can immediately create an image and work on it.

By writer

답글 남기기