Consistent Style
Midjourney released a new version and several new features that make AI generative art more scalable. That “scalable” part is a bit literal — as the service can now infer what an image would look like zoomed out. However, the farther out you go, the more likelihood for discontinuity.
The service also announced the release of style consistency, which means you can save a consistent style from one image to another. This is a huge time savings and will also allow for the scalability of generation jobs. Ultimately, this will lead to better content and eventually, videos that are completely generated by and with AI.
What we may find is that image generation actually has the potential to unlock many non-image related tasks. Many math and language problems can be expressed through images, so maintaining consistency could be a game changer for reworking these tools.
