One of my first gigs out of school was writing about movies for a small independent paper in Denver. Around the same time I was also working the graveyard shift as an editor for a wire service. Press screenings were usually scheduled in the morning, and I’d show up for a 10am viewing of the new Meg Ryan and Russel Crowe vehicle, Proof of Life, having been up for 18 hours.
That movie was highly forgettable, but I do remember walking into the lobby with another critic after it ended and being quizzed by the PR agent, “What did you guys think?” While I tried to gather my thoughts, Pam threw out the perfect barb: “No dice.” Decades later, her words echo in my mind as more and more slop fills the content troughs. Neill Blomkamp’s short zombie AI thingy? No dice. Random people “recreating” a fight scene from The Matrix? No dice. Fake people in already bad commercials? No dice. Tilly Norwood? No dice.
There’s an overwhelming amount of awful generated content in the world, and the ever-present spectre of more. This makes it hard to process the odd slices of it that are compelling. As a designer who is often thinking about what motivates certain behaviors, this feels like a crevasse worth exploring. People will inevitably be confronted with oceans of slop, can we find proof of life somewhere in the deluge?
Gossip Goblin, (Swedish-based filmmaker and digital artist Zack London) has a 21-min bit called The Patchwright that isn’t totally no dice. It was written and voiced by humans and its dystopian world where most living creatures have been enhanced with embedded technology reads like satire. It still feels off in a way that I’ve been struggling to pin down, but in terms of being an affront on cinema itself, London’s work feels closer to James Cameron world building than someone with half an idea prompting an LLM to write a screenplay.
As I was scrolling Gossip Goblin’s shorts, I got pulled into “Dopamine Resistance (Twelfth Cycle of Humanity)” and right away started sweating. “You’re drowning your species in dopamine!” Shouted a severe looking cyber monk with hollowed cheeks and a chip stuck in his hairline. “You’re gorging on noise to block out your own thoughts.”
A nervous pit opened in my belly and a familiar fatigue pressed behind my eyes. It was like all those nights staring at a desktop screen and combing the nascent interweb for mp3s “...flinching from silence like it’s pain!” A pale drooling face distorted by harsh blue light, a close-up on a fervid eyeball stretching to take in more data “… and still you keep swiping at the void, begging for something brighter …” scenes of mushroom clouds and burned ruined cities, sickly deranged figures clamoring for more “… keep it alive because once it goes, no part of you will remain!”
Maybe it was a bit too on the nose, that’s why I felt light-headed and queasy. Something was up, but the experience was uncomfortable enough that I left the stone largely unturned. Then I was corresponding with Jonathan Larroquette (co-host of Uhh Yeah Dude, one of the world’s very first podcasts) and he pointed me toward Connor Clarke (aka kingcon2k11) another generative video artist.
Like Gossip Goblin, his work free dives deep into the darkest corners of techno-nightmare. More rooted in reimagining nostalgia, much of Clarke’s work looks like it was found on unmarked VHS tapes from a thrift store bin. There’s a music video for an unreleased Ween song that strokes my deep affinity for claymation while also filling me with freakish dread. Humans have always sought to improve the efficiency of their tools, but if someone had actually made this video with real fingers, brains, and clay, it would not be giving me anxiety.
Plenty of these videos seem to bask in the uncanny valley—body parts that aren't quite right, faces that read cold, animals moving with unnatural gaits. The rapid twisty nature of the kingcon2k11 fare clobbers my mind with oddities that are often gory or horrifying. Before I can process any of it, I'm up to my neck in the next bloodbath. Larroquette tells me that he also has a physical reaction to these videos and has talked to others who’ve reported the same.
Long ago, his UYD co-host, Seth Romatelli, created an oasis of chunked out '80s and 90s television on Vimeo called Seth's Clips. I watch them now and wonder if I’m looking at ancient ancestors of whatever is happening here and here. Seth’s Clips float by with little context, just a scene from Family Ties where Alex P. Keaton (Michael J. Fox) crashes out on speed running into documentary footage of Arnold Schwarzenegger going shirtless in a prison yard, tumbling into a tense interaction from an episode of Cops. Relics from an era when there were far fewer cultural touchstones waiting to be mashed up into oblivion.
Referenced clip · Seth’s Clips · Open on Vimeo ↗
Part of the mashup nature of these videos has to do with the limits of the technology. Models often top out at 15 seconds of output. So the hodgepodge nature of a lot of this content dances around those constraints. A newer Gossip Goblin video seems to be flexing prowess, with long extended shots. London starts with still images in Midjourney (refined in Nano Banana), which he animates with first-frame image-to-video. He maintains manual control rather than using more automated all-in-one generation tools. This explains why his work is more cinematic than most.
Clarke is part of a cohort of AI artists working under the banner of Slop Shop, which has done work for Xbox, Meta, Carhartt WIP, and Lil’ Uzi Vert. Online their work draws comments calling it out as a waste of resources and an affront to the stolen content fueling models. There are also those lavishing praise on these weird creations. A comment on one clip said, “Your work honestly inspires my drawings,” suggesting an interesting loop, AI art inspiring pen-to-paper art.
This opens a bigger discussion on models running off of ripped art inspiring new art, but also raises a pointed question about my physical discomfort. Is it a usurper thing? Am I subconsciously recognizing the power and prowess of a new, unfathomable intelligence? After being on top of the food chain for thousands of years we are getting twinges of feeling like prey? Some of these manic creations are like watching a toddler running around with a handgun that may or may not be loaded. The fear response limits my appetite, preventing me from bingeing the stuff that actually hits.
This is a good thing. My own disgust is doing something that infinite scroll was built to prevent—telling me when to stop. We don't typically build for that kind of limiter; we design it out. I have no appetite for the sloppiest of slop, but I am drawn to the examples that are running headlong into this uncanny valley of dystopian whatever soup. This stuff makes the most compelling argument for the existence of AI art, but I can only stomach so much. There’s space for whatever this new weirdness is, but will it ever become the main thing? I’m feeling “no dice.”
Josh Tyson is a contributing editor at UX Magazine and host of Invisible Machines with Robb Wilson.