Midjourney, a leading AI image generation company, is working on a new feature that will completely change the way people interact with generated images Using a new form of 3D technology, users will be able to effectively enter into images

Announced at Midjourney's recent office hours on Discord, this new feature is just one of several new features currently in the works and could be announced between now and the end of the year Midjourney CEO David Holz also announced that by the end of the year v7 will be here, with a video model and, for the first time, an upgraded image editor that will allow users to edit external images, he said

The startup has been working on 3D technology for some time as part of a broader vision, like the Star Trek holodeck, of generating a world and using that world to make movies, games, or just something you can spend time with

According to Martin Nevelong, an expert in generative AI, Midjourney's 3D model will be an entirely new approach based on NeRF technology, which is widely used in game development NeRF is a It is a neural network

Few specific details have been released, but Holz has spoken several times in the past about his desire to build a virtual world in which everyone can interact and build At a recent office hour, he said he wanted to be able to easily convert Mid Journey images into a 3D environment

It would “allow for camera movement within a certain range” and have a camera path system that would allow for 60 frames per second rendering and video output This would be independent of the video model planned for the future, which would be similar to something like Runway or Sora

Other companies like Luma Labs, Adobe, and Meta also have text-to-3D models, but these tend to be object-based rather than environment-based; Roblox is working on AI-generated environments, and startup Cybever has an immersive text-to-3D world models have started a waiting list

According to a post on X by Alfonso Rosenberg, the main focus of the recent Midjourney office hours was more direct updates, including personalization For example, some modes will now be turned on by default, allowing users to more easily refine their results by simply selecting from the four images generated

There will also be an updated image editor that will allow external images to be edited within Mid Journey, but this will come with stricter moderation and privacy protections

A video model of Midjourney could be available by the end of the year, but it will likely excel at illustration styles rather than photography Finally, version 7 of the Midjourney model is in the training phase and is expected to be released within one to three months

Rosenberg, citing a Discord thread by JamesGriffing, said that two hardware projects are underway, a new explorer page is in the works and will be released this year to “build worlds, not just create images” He added that they are testing storytelling tools