I can’t stop thinking about that James Bridle article. The excerpt in the post makes it sound like standard parental handwringing, but the true premise of the article is more like, “There is an algorithmic nonsense machine churning out a surreal hellscape of videos where nobody knows if they’re made by humans or machines anymore, and the automated mixture and adjustment of content means that the videos are getting more and more unsettling and potentially traumatizing without even meaning to. Today’s kids are a couple of clicks away from unending playlists reminiscent of compilations of ‘disturbing’ clips from childhood films, except these playlists are in many cases generated entirely by accident.”
The results of these algorithms’ observation of viewer behavior and keyword trends is like the video equivalent of content-aware scaling, where the person’s face gets more and more distorted as the algorithm focuses on the parts of the face the viewer is paying attention to the most. They’re like what would happen if a Portuguese bootleg studio that made crappy 3D animated knockoffs of popular kids’ movies to cash in on inattentive grandparents asked David Lynch to direct their next feature.
I didn’t even know what those videos were when I saw them pop up in unrelated searches due to their word salad titles designed to cast the widest net. I thought they were just weird fetish videos because the thumbnails focused on tooth pulling or feet, but it turns out they actually were intended for children. You can’t even tell which ones are “originals” and which ones are parodies by looking at them alone, because a subset of folks were so baffled by the existence of these videos that they naturally just had to make “parodies” of them.
If you told me about the surreal rabbit hole that involved the making of these videos a month ago, I wouldn’t have believed you.
