Meta makes its own AI to create images, but its purpose is very different from DALL-E.

DALL-E is truly the most popular artificial intelligence to date. Everyone wants to create their own works of art just by entering a few words. The possibilities are endless with this AI, and Meta seems to already know that. In fact, decided to make their own: go to the Make-A-Scene.

As he has become accustomed to working on his social networks, Meta is “inspired” by a new technology to launch his own project. However, it seems the company has other ideas in mind. According to their press release, this tool can help shape the deeper worlds of the metaverse, and can also help you create digital art, even if you don’t have the talent.

Also, unlike DALL-E, you can help Make-A-Scene mold the scenes using the sketch. That is, in addition to providing textual input, it will give you a window where you can specify the location and size of the composition elements. The latter is not complicated, as it is a simple sketch that will serve as a guide for the final outcome.

Make-A-Scene takes the layout of the scene to create nuanced input sketches. It can also create its own design using a text-only prompt, if the designer wishes. The model focuses on learning the important aspects of images that are probably most important to the creator, such as objects or animals.

The new AI in Meta and its variations on DALL-E

With its new AI, Meta doesn’t plan to achieve the photorealism we’ve already seen in technologies like DALL-E 2. In fact, on the company’s website it says his images are entirely intended to give a more artistic feel. They almost look like they were painted by hand. A good idea if looking to stand out from the competition.

Meta wants everyone to be an artist. This is it a test of artistic democracy as before. Even without “talent” or techniques, in a few words and a sketch, you can create a job you can be proud of.

According to AI artist Sofia Crespo, “As a visual artist, sometimes you just want to make a composition based on the hand, to draw a story that will follow the eye, and that’s it. »

Make-A-Scene, in addition, offers images with a resolution of 2048 x 2048 pixels. That’s no small feat, especially compared to the DALL-E mini’s awesome combination. Of course, this Meta AI is not equivalent to the portable version of OpenAI technology.

When can you try Make-A-Scene?

You’ve already seen what Make-A-Scene can do, now it’s time to figure out how to use it. It seems, for now, The AI ​​is in the closed testing phase What does that mean? This means that only those invited by Meta itself can use it.
Until there is a public trial, or an official launch of artificial intelligence, no one else can use it without the company’s permission. However, even if Meta is not a saint of our devotion, we must say that this AI offers an interesting vision of the future of these technologies.
Not just for the idea of ​​helping complete the metaverse. But also for the ability it gives everyone to communicate. More graphic and faster sketches for large architectural or audiovisual ideas, in a more effective communication between people with disabilities and their families or health care professionals. The possibilities are endless, and they are already there.

Leave a Comment