Text-to-image revolution explained
How programmers transformed the internet into an artistic tool. DALL-E 2, Midjourney, Imagen, explained.
Subscribe to our channel and enable notifications: http://goo.gl/0bsAjO
In January 2021, AI research has produced a multitude of deep-learning algorithms capable of creating original images using simple text prompts. This effectively extends the human imagination. OpenAI, Google and Facebook researchers have created text-to image tools they haven’t yet made public. Similar models are available online and in open-source software.
These tools are a major cultural shift, as they eliminate the need for technical work in the image-making process. They select instead for creativity, skillful language use, and curatorial tastes. It is difficult to predict the ultimate effects, but these algorithms, like the inventions of the digital camera and the camera itself, herald a new democratized expression which will lead to a further explosion of images produced by humans. Like other automated systems that are trained using historical data or internet images, these algorithms also carry risks.
This video is an introduction to how we arrived at this point, what the technology does, and its implications. And for an extended discussion about what this means for human artists, designers, and illustrators, check out this bonus video: https://youtu.be/sFBfrZ-N3G4