Latest News And Breaking Headlines

This AI art app is a glimpse into the future of synthetic media

If you’ve been hanging out on Twitter lately, you’ve probably noticed a plethora of AI-generated images sprouting across your timeline like weird, algorithmic visions. These photos were taken with a new app called Dream, which allows anyone to create “AI-powered paintings” by simply typing a short description of what they want to see. They are strange, often creepy things – and extremely fun.

The resulting artwork has its own distinct aesthetic, defined by swirling shapes and disjointed objects. The real magic, though, is that whatever you type, the app will generate something that’s visually appealing (at least until we get used to this toy too much) and that matches your prompt in often surprisingly appropriate ways.

For example, consider the image below: “Galactic archeology with metal-poor stars.” Not only did the app capture a photo that captures the mind-boggling galactic scale of a nebula, but the star-like highlights scattered in space are mostly blue — a hue that scientifically accurate in front of metal poor stars (because metallicity affects their color).

A few fast searches on Twitter reveal many more examples, but really you should play with the app yourself to understand it better. (If nothing else, the images it generates are just the right size to create a personalized wallpaper for your phone.)

This kind of AI-generated artwork isn’t new, but it’s getting better and more accessible. Past examples of these types of text-to-image models include research-oriented programs such as: DALL-E and VQGAN+CLIP, as well as more specialized commercial projects such as art breeder (which is especially good at making portraits of fictional creatures and people). With tools like these, the AI ​​art scene has exploded in recent years, with practitioners creating everything from real-life Roman emperors to infinite waifus.

The Dream app goes one step further with its speed, quality and accessibility. It is available on iOS, android, and the web and is the work of a Canadian startup called Wombo. The company previously made that AI-powered app that lets you input static images to create lip-synced renditions of memeable songs. What exactly powers Dream isn’t clear (we reached out to Wombo to find out), but much of the AI ​​art technology is open-source, meaning the company likely built on previous work to create the app.

In general, such programs are trained on vision data sets – huge libraries of images tagged based on objects and landscapes. The programs choose consistent patterns and themes in these images and use this information to try to generate something that matches the user’s prompt. We don’t know what dataset Dream’s algorithms are trained on, but based on the output, we can safely say that it contains a wide variety of images — able to generate images that match anime characters and video games.

The accessibility of Dream means that it is also used for new purposes. It has been used for viral games (such as entering your Title dissertation and sharing the result) and also for more targeted projects. In a great Twitter thread, the writer and illustrator Ursula Vernon (who publishes under the name T. Kingfisher) shared a short comic they had created using Dream. The characters of the comic are hand drawn, but the backgrounds are artificially generated, explaining the surreal, shifting quality of the images due to the setting: a dream library overseen by the Egyptian god of writing, Thoth. .

Vernon tweeted about her experience, noting that she had to do a significant amount of work to prepare the images and that the program’s inability to create landscapes from a space with consistent architecture created its own challenges.

“In conclusion: does it work visually? I think the answer is ‘sort of’ tweeted Vernon. “Of course, as an artist, I am well aware of the idiosyncrasies. Like a dream sequence, the screwed up architecture works for a bit, but how long can you get away with it? Sooner or later, the reader will probably notice that nothing in the same scene takes place from a different angle.”

Despite its obvious limitations, Dream gives us a glimpse into the future of synthetic or AI-generated media. For evangelists in this space, the promise of technology is one of infinite variety. In the future, they say, games, comics, movies, and books will all be generated on the fly in response to all our requests and whims. And though we have a long, long away from such media that match the quality of human output, limited, hybrid applications will come faster than you think – appearing like something you first saw in a dream.

This website uses cookies to improve your experience. We'll assume you're ok with this, but you can opt-out if you wish. Accept Read More