Artificial intelligence developed by OpenAI DALL-E 2Although it has not yet opened to the public, it has managed to come to the fore with its talent many times over. The artificial intelligence tool, which most realistically transforms what is represented in writing into a visual image, is a hot topic on social media these days with a different ability.

give access to the early access version of DALL-E Kevin Parry, has included new works that are mesmerizing in his posts on Twitter. What made the aforementioned visuals special from other visuals, this time was not the conversion of written definitions into visuals, but the perfect harmonization of different visuals in one frame.

DALL-E expands the image based on the context and theme of the image:

According to Parry’s post, DALL-E automatically expands any image to the size you want. This extension is not an enlargement of the image, but a continuation of the same image. like a re-creation It happens – just like fill the paragraph questions in exams.

In addition, DALL-E can achieve this not only by starting from one image, but also by ensuring that these two images based on two images are compatible with each other. This feature is also what DALL-Es display in text can also be combined with the ability to translate to images† The result is images like this: