Unimpressive AI and the beauty of mystery
In the search for meaningful applications of generative AI, upstarts are scrambling to apply their new methods to any and all ideas. The most recent addition? Image generator Midjourney has added a feature called “Zoom Out”. The website description reads: “The Zoom Out option allows you to extend the canvas of an upscaled image beyond its original boundaries without changing the content of the original image.”
Sounds great right? Just expand any visual media beyond its original borders, to create an augmented piece “without changing the original”. This new feature has of course become an overnight sensation, and prompted (pun intended) internet users to try expanding a wide variety of content, from photos of loved ones, to world-famous paintings, in order to “see beyond the original boundaries”.
The feature, and its popularity, fits perfectly with the state of the internet and mainstream pop culture media. In our collective state of creative mind, there is an obsession with peeking behind the curtain: reboots galore as film studios “revisit” and “reimagine” legacy blockbusters in order to flesh out their stories, prequels and sequels frequent bookstore shelves as epic tales are extended and expanded, and every successful tv show needs a slew of spinoffs to “explore” the characters further. Even the gaming industry, a once immensely creative sector at the crossroads of media, entertainment and interactivity, is mostly dead set on re-releasing old games with updated graphics and “added content”.
Revisiting a story is not bad in itself. In fact, it can be rewarding for the audience in some instances. However, the reboot-and-spinoff-setup has a critical flaw: it does away with mystery, to the detriment of the storytelling and the allure of many characters. Not knowing the exact backstory of Han Solo is part of his draw, and what makes his arc so interesting; a smuggling scoundrel, turned hero at the time of need, as he discovers what real belonging feels like — and thereby finds his courage. I don’t need to know how he started out, or what the details behind his name are. His backstory is unclear, and that is what defines him.
The same principle applies to photos and art. If I see a close-up shot of a bee on a flower, I don’t need to “zoom out” the photo with Midjourney to see it surrounded by auto-generated plants in a fictive field. It’s not impressive, and it definitely isn’t exciting. The magic of that photo is in the intimacy of the moment that was captured — and what was excluded to make it happen.
There is no magic to auto-generated content. It’s just a process, in the same way the studios churn out regurgitated blockbusters to make a quick buck; neither come anywhere near something that would excite, let alone impress, me.
