Midjourney's latest AI art generator achieves the breakthrough of writing text in images

The new version of the Midjourney AI art generator, Midjourney v6 (Alpha), has been released. Midjourney is a small independent research lab that specializes in generative AI image models. The goal of the lab is to explore new mediums of thought and expand the imaginative powers of the human species. The competition in the AI image generator field is fierce, with OpenAI’s DALL-E 3 (Stylized DALL·E) challenging for the top spot. One of the key features that sets it apart is its ability to render clean and legible text within the generated images, known as text-in-image. This is a difficult task that requires the AI model to recreate the typeface consistently and from scratch every time. Previous AI image generators have struggled with this, but Midjourney’s latest version has made significant progress. The new version also includes improvements in prompt length, object permanence, natural language processing, and prompt understanding. It now has a conversational style and is better at ignoring filler words. Additionally, it can comprehend linguistic negatives and adhere to them in generated images. To make the most of the new version, users can enable it in Discord settings without having to manually specify the model in their prompts. While Midjourney v6 can produce legible words in generated images, it may still require a few tries to get it right. Overall, the latest version of Midjourney’s AI art generator brings significant advancements in text-in-image generation and offers a range of new features for users to explore.