David Holz, the founder and CEO of Midjourney, recently shared updates on the company's current projects and answered questions from the community during the "office hours" on Discord. This interaction revealed that Midjourney is undergoing various technological innovations to keep up with the increasingly competitive AI image generation field.
New Version Delayed But Richer in Features
Despite rapid progress in a short time, Midjourney has seen a significant slowdown in the release frequency of new models. The company's latest version 6.1 was released in July, after which users experienced a relatively long waiting period. The originally planned version 7, set to be released before the summer, now has a new timeline: it is expected to be ready within one to two months.
Holz stated that the development team is not rushing to release new models but is focusing on enhancing the accessibility of the technology and the value of the tools for daily professional use. For instance, many users initially found it uncomfortable to use Discord, but now the company has provided a web interface for all users, including those who want to try AI image generation for free.
Performance Improvements and New Features
Midjourney plans to introduce several new features to enhance user experience:
Multiple Image Generation: Users will be able to generate up to 8 images at once, an improvement over the current limit of 4.
Image Editor: The company is developing a tool that allows users to edit external images (such as photos), including local repainting (inpainting) and extended painting (outpainting) functions.
3D System: A more ambitious project is the development of a 3D system that allows users to virtually immerse themselves in AI-generated images. This system will use a new "NeRF-like" format, rather than traditional polygon rendering, NeRF, or Gaussian scattering technologies.
Personalization: Based on user feedback, Midjourney aims to learn more about user preferences to provide more personalized results. This feature has recently been enabled in the Niji model, which specializes in generating anime characters.
Video Generation: In addition to static images, Midjourney also plans to venture into the video generation field.
Hardware Development: The company is also investing in hardware development alongside software.
Facing rapid development from competitors like OpenAI, Midjourney has chosen a more stable development path. By introducing more practical features and enhancing user experience, rather than just pursuing model iteration speed, Midjourney aims to maintain its unique advantage in the competitive AI image generation field.
Product Entry: https://top.aibase.com/tool/midjourneywangyeban