Over the last decade, AI’s reality-bending abilities have been kept increasingly busy. We’ve seen computer vision-powered 3D dioramas, trend-setting style transfer, viral photorealistic selfie-tuning, selfie-retouching, face swaps, and deepfakes, and plenty of frivolous (and hilarious) selfie filter fun in between.
AI-powered visual remixing has repeatedly demonstrated its ability to capture attention. However, keeping “eyes on” after the novelty of an AI-generated effect wears off can be difficult. (Selfie retouching apps, on the other hand, have a constant demand for machine learning as a reality enhancer.)
What’s most notable about AI-enabled synthetic media developments over this time period is how quickly these visual effects have picked up speed, aided by ever more powerful mobile processing hardware.
As soon as a new piece of artwork appears, the app attempts to sell it by displaying a “buy print” option that links to its web shop and appears to be a clever way to turn a visual trick into actual revenue. (It sells “Custom Wombo Dream Print[s]” beginning at $20 for a matte poster and $45 for a framed print.)
If the startup can convert 20 seconds of processing time into $20+ in revenue, it could create a nice little money pipeline.
There will also be a slew of these artistic AIs, each with its own “flavors” and “characters” of visual output derived from its training data. Or, if you prefer, different “styles” of art AIs. (However, “specialisms” may be closer to the coded mark.)
There are a number of other GAN-based image generation AI tools available — and I confess to being a big fan of Pixray’s system (the pixel art outputs are especially cute), despite the fact that its processing speeds are much, much slower — but Wombo appears to have been the first to appify and monetize this technology.
The next decade of reality-altering machine learning will be an adventure.