the tool image generation by AI, Halfway through the trip, predicts a year 2024 full of news. Although version 6 is still in the experimental “alpha” phase, the teams are already working on v7, promising some innovations, while evoking the arrival of the generation of videos and 3D environments before the end of the year. Summary of the latest information.AIO: The future of SEO positioning for artificial intelligence
v7 could arrive quickly, website too
In December 2023, Midjourney introduced the alpha version of the Sixth iteration of your model.. More than a quarter later, users are still waiting for a beta version, which the imaging teams are still working on. This v6 could have one last update before its final version, in particular with the aim of improving the generation of hands, bodies and the coherence of the texts, as well as the general speed of the tool.
However, the Midjourney teams seem to be primarily focused on version 7 of the AI. With a release date closer than one would expect. In fact, according to the consultant Julio Alithis could intervene in “more than a month, but less than three months”. Therefore, v7 is likely to be available in early summer. This version 7, which has become a priority for developers, promises a better understanding of the indications and an overall improvement of the model. “with a quality gap greater than between v5 and v6”.
Latest news expected by all Midjourney users: the Web platform Creation. This interface, which is already accessible to everyone, but only usable by members who have generated more than 1000 images, should allow creating images without having to go through a Discord server, as is currently the case. An eyelash Updates now allows you to stay up to date with the news from Midjourney, a page Leaderboard is under development to collect user feedback, while the platform is expected to open to more members very soon.Mastering Facebook Marketing: Strategies, Agencies, and Services
Midjourney works on custom models.
A great innovation is expected in the coming weeks, the functionality Model customization should be implemented in the tool soon. It should allow us to move away from the “Midjourney look” adopted by the images created. In fact, Midjourney has some generative biases, and the AI sometimes adds details that weren’t specified in the prompts. With Model customization, the tool will use your creative preferences to add details and enrich the requested images. His instructions will still take priority, but Midjourney will apply your preferences to items not specified in the query.
In order for Midjourney to understand what your preferences are, you can now access a special page from the web platform: https://www.midjourney.com/rank. In this interface, accessible by clicking Rate images In the left navigation panel, you will be offered two images (see image for one). Midjourney invites you to click on which of the two you prefer. The tool will use this information to understand your preferred style and improve your “custom model.” David Holz, founder of Midjourney, invites users to spend 10 to 15 minutes performing these actions: “Better early than late”, he wanted to clarify.
He Model customization It will not be integrated by default in v6, but it will be integrated for v7. The feature simply won’t be added to the Niji v6 model, which allows you to create images. in anime/manga style.
3D and video before the end of the year?
Finally, Midjourney leaked important information about its plans for 2024, without giving many details. One thing is (almost) certain: Midjourney should allow, before the end of the year, to generate videos and 3D environments. The idea behind these two new functions would be to convert images into 3D and video, instead of generating them through a text message. It is currently impossible to say which of these two innovations will arrive first.
In any case, it probably won’t be with v6, which hasn’t been tried. “good enough” by David Holz. Therefore, we will probably have to wait for the v7 rollout to get an overview of these new features and a more precise idea of their release date. The founder of Midjourney wants a video tool to be ready as soon as it is implemented: “There are ‘good’ things and ‘good enough’ things. I don’t think there’s anything right now that’s good enough. Even if OpenAI’s (soraEditor’s note) It’s good, it’s not good enough. We will continue to work on it until it is good enough. »
