MidJourney's SHOCKINGLY good new UI just released (no more Discord!)

Volo
26 Jan 202407:45

TLDRMidJourney, a leading AI image generation tool, has released a new web interface, eliminating the need to use Discord for image creation. The platform is initially available to a select few, including the video's narrator, who has generated over 5,000 images on the platform. The new interface features an explore page with community-generated art, and users can generate images by typing prompts without the need for specific commands. The interface allows for customization, such as aspect ratio, stylization, and weirdness levels. Users can also use images as prompts and remix existing images. The platform is missing some features like in-painting but is expected to evolve, with the web application likely to include all features currently available on Discord. The narrator is excited about the improved usability and believes the web interface is a significant upgrade from the Discord bot.

Takeaways

  • ๐ŸŽจ MidJourney has released a new web interface, allowing users to generate images without using Discord.
  • ๐Ÿš€ The new platform is currently in an early preview stage and accessible to a limited number of users.
  • ๐Ÿ–ผ๏ธ Users can view a Pinterest-like list of community-generated images on the explore page.
  • ๐Ÿ“ˆ MidJourney focuses on refining the quality of generated art, making it increasingly difficult to distinguish from real images.
  • โœ๏ธ The image generation process is streamlined; users can type a prompt directly without using a specific command.
  • ๐Ÿ–ฅ๏ธ Users have the option to generate images in portrait or landscape format, in addition to the default square format.
  • ๐ŸŽญ The 'stylization' setting allows users to control the artistic aesthetic of the generated images.
  • ๐Ÿ” The 'weirdness' and 'variety' flags let users experiment with the uniqueness and diversity of the generated images.
  • ๐Ÿ’ณ A subscription to MidJourney allows for a faster generation of a certain number of images and unlimited generation in relaxed mode.
  • ๐Ÿง™ Users can choose between different models, such as V6 for the most realistic images and 5.2 for fantasy art.
  • ๐Ÿ”„ The platform includes features to remix images and use them as prompts for new image generations.
  • ๐Ÿ” There is a search function that allows users to find similar images generated by others and see the prompts used for those images.
  • ๐Ÿšซ The platform does not currently support negative prompts or in-painting, which would allow for erasing and regenerating parts of an image.
  • ๐ŸŒ The web application is designed to be more user-friendly and is expected to evolve with more features from the Discord bot.

Q & A

  • What is MidJourney's new UI feature that has just been released?

    -MidJourney has released a web interface, which allows users to generate images without having to use the Discord bot.

  • How can users access the new MidJourney web application?

    -Users can access the new web application by going to alpha.midjourney.com.

  • What is the purpose of the 'Explore' page on the MidJourney web application?

    -The 'Explore' page allows users to view different art generated by the community, including trending and top art, as well as their personal likes.

  • How has MidJourney improved the quality of the art generated?

    -MidJourney has focused on refining the quality of the art they create, making it difficult to distinguish between images generated by MidJourney and real images.

  • What is the process for generating an image using the new MidJourney web interface?

    -To generate an image, users can type a prompt in the designated area, and the system will generate images based on that prompt without the need for a specific command like 'slash imagine'.

  • What are some of the new features and flags that users can set when generating images on the MidJourney web application?

    -Users can set features such as the aspect ratio (portrait or landscape), stylization level, weirdness, and variety. They can also choose the speed of generation and select the model (V6 for the most realistic images or 5.2 for fantasy art generation).

  • How does the MidJourney web application handle multiple concurrent image generation requests?

    -The web application allows users to input the same prompt multiple times, running concurrent image generation processes.

  • What is the 'Vary' feature in the MidJourney web application?

    -The 'Vary' feature allows users to generate four more images that are similar to a chosen image but with slight differences.

  • How does MidJourney handle prompts with specific elements, like a 'Pikachu jumping over a rock'?

    -MidJourney captures the artistic aesthetic well but may not always include all elements from the prompt in the composition. Users can edit the prompt to better suit their needs.

  • What is the 'Creative Upscale' feature in the MidJourney web application?

    -The 'Creative Upscale' feature allows users to enhance the quality and detail of an image, potentially making it look more refined or 'sweet' as described in the transcript.

  • Does the MidJourney web application support negative prompts?

    -While there isn't a dedicated UI component for negative prompts, users can include them in the prompt text, and the system will apply them to the image generation.

  • What feature is currently missing from the MidJourney web application that the user would like to see?

    -The user would like to see an 'in-painting' feature that would allow erasing a part of the image and regenerating that part.

Outlines

00:00

๐ŸŽจ Introduction to Mid Journey's New Web Interface

Mid Journey is recognized as one of the leading AI image generation tools, and it has recently launched a web interface, eliminating the need to use a Discord bot for image generation. The tool is in its early preview stage, accessible to a select few, including the speaker who has been a significant user, having generated over 5,000 images. The new platform (alpha.MidJourney.com) showcases community-created art and is rolling out features progressively. The interface includes an explore page with a Pinterest-like display of images, sorted by trends and personal likes, highlighting the high-quality art that Mid Journey is known for. The platform's focus on quality is evident, making it challenging to distinguish between generated and real images. Image generation has been streamlined, allowing users to input prompts directly without using specific commands. Users can also control various aspects of the image generation process, such as aspect ratio, stylization, weirdness, and variety, as well as select from different models for different styles of output. The platform also allows for the use of images as prompts for further generation and provides options for creative upscaling and remixing.

05:00

๐Ÿ” Exploring Advanced Features and User Experience

The speaker delves into the advanced features of Mid Journey's web interface, discussing the ability to generate images in portrait or landscape formats, control the level of stylization, and experiment with the 'weirdness' parameter for more abstract results. The platform's default setting generates square images, but users now have more flexibility. The speaker also talks about the 'variety' parameter, which determines the diversity of the initial set of four images generated. A subscription plan is mentioned, which allows for a certain number of quick generations and unlimited images in a relaxed mode. The platform's latest model, V6, is noted for producing highly realistic images, while model 5.2 is favored for fantasy art. The speaker also discusses the ability to use images as prompts and generate variations based on those images. However, they note that the platform struggles with following detailed instructions consistently. The lack of a dedicated UI component for negative prompts and the absence of an in-painting feature, which would allow for the regeneration of parts of an image, are mentioned as areas for potential improvement. The speaker concludes by expressing excitement about the platform's evolution and its focus on usability, inviting viewers to share their thoughts and subscribe for more updates.

Mindmap

Keywords

๐Ÿ’กMidJourney

MidJourney is an AI image generation tool that is considered one of the best in its field. It has recently released a new user interface (UI), which is a significant improvement over the previous method of using a Discord bot for image generation. The new web interface allows users to generate images more intuitively and efficiently, which is a core theme of the video.

๐Ÿ’กWeb Interface

The term 'web interface' refers to a graphical user interface that allows users to interact with a web application. In the context of the video, MidJourney's new web interface is highlighted as a major upgrade, providing a more user-friendly way to generate images compared to the previous Discord bot.

๐Ÿ’กDiscord Bot

A 'Discord Bot' is an application that operates within the Discord platform, performing various tasks such as messaging, image generation, and more. The video discusses how MidJourney has moved away from using a Discord bot for its image generation process, which was considered clunky, to a more streamlined web interface.

๐Ÿ’กImage Generation

Image generation is the process of creating images from textual descriptions using AI algorithms. The video demonstrates how MidJourney's new web interface allows users to input prompts and generate images based on those prompts, showcasing the tool's ability to create high-quality and artistic images.

๐Ÿ’กPrompt

A 'prompt' in the context of AI image generation is a text input that describes the desired image. The video script describes how users can now simply type their prompts into the new web interface without the need for specific commands, as was previously required with the Discord bot.

๐Ÿ’กStylization

Stylization refers to the degree to which the generated image embodies the unique aesthetic of the AI tool. In the video, it is mentioned that users can adjust the level of stylization, with lower values producing more literal interpretations of the prompt and higher values resulting in a more 'MidJourney' artistic style.

๐Ÿ’กWeirdness

Weirdness is a parameter in MidJourney's image generation process that controls the level of abstractness or unconventionality in the resulting image. The video explores the use of the 'weirdness' flag, indicating that higher values can lead to more unexpected and creative outcomes.

๐Ÿ’กVariety

Variety, in the context of the video, refers to the diversity of images generated from a single prompt. MidJourney's system generates four initial images and allows for variation among them, offering users a range of options to choose from or to further refine.

๐Ÿ’กModel

The 'model' in AI image generation denotes the specific version of the AI algorithm being used. The video mentions different models such as V6 and 5.2, with V6 being the latest and most realistic, while 5.2 is known for its reliability in generating fantasy art.

๐Ÿ’กUpscale

Upscaling is the process of increasing the resolution of an image while attempting to maintain or enhance its quality. In the video, the presenter upscales a generated sword image to a higher resolution, demonstrating the tool's capability to improve the detail of generated images.

๐Ÿ’กNegative Prompts

Negative prompts are text inputs that specify what should be avoided or excluded in the generated image. Although the video does not mention a dedicated UI component for negative prompts, it suggests that users can include them in the main prompt text for the AI to consider during image generation.

Highlights

MidJourney has released a new web interface, eliminating the need for using Discord for image generation.

Only a limited number of users are allowed an early preview of the new web application.

The new interface allows for image generation without the need to use the 'slash imagine' command.

Users can now generate images in portrait or landscape format, in addition to the default square.

The 'stylization' feature allows users to control the aesthetic strength of the generated images.

The 'weirdness' flag is a new feature that can be experimented with to create more unusual images.

The 'variety' setting determines how different the initial set of four generated images will be.

A subscription to MidJourney allows for a certain number of quick image generations and unlimited relaxed mode images.

The latest model, V6, is designed to generate the most realistic images.

Users can use images as prompts to generate new images with slight variations.

MidJourney's new interface allows for the editing of prompts and the rerunning of image generation.

The platform is creating interconnectivity between Discord and the web app.

Users can upload a file to generate an image based on that file.

The 'creative upscale' feature allows for the enhancement of generated images.

MidJourney has improved in generating images that follow the user's prompt more closely.

The platform allows users to search and remix other users' images using their prompts.

Negative prompts can be included in the prompt to guide the image generation process.

The ability to in-paint, or regenerate parts of an image, is a feature that is currently missing from the new interface.

The web application is expected to evolve and include more features from the Discord bot in the future.

The new web interface is seen as a significant improvement in usability over the Discord bot.