Week 1: Naming the genre

Why "slop" is a genre term and not an insult, and what a semester of taking it seriously is supposed to produce

22 February 2027

Open the slides

The deck for this screening lays out the actual taxonomy — the five tells, the claim that a genre is defined by them rather than by its makers’ intentions, where the semester’s three assessments sit against that claim. This page covers ground the deck doesn’t: where the word came from, and why a course gets to treat it as a category worth studying rather than a complaint worth making once and moving on.

Where the word comes from#

“Slop” had circulated as niche developer slang since around 2022, a reaction to the first wave of image diffusion models — but it hardened into a specific, deliberate term of art and broke into mainstream tech commentary around 2024, applied to unwanted generative output the way “spam” was applied to unwanted email a generation earlier. Both words borrow from something produced cheaply, in volume, for consumption rather than for quality.

A hand-drawn ad with a spelling mistake is bad craft. A thousand AI-generated ads posted an hour, six-fingered hands included, are slop — not because the craft is worse (sometimes it isn’t) but because the volume-per-unit-of-attention the genre was built to extract is the point, and the tells are what’s left over when nobody needed to fix them to hit that number.

What “taking it seriously” means, mechanically#

Genre theory has done this move before. Melodrama spent decades dismissed as “women’s pictures” — over-plotted, over-scored, beneath serious criticism — before film scholars did the close-reading work of asking what the over-plotting and over-scoring were actually for, and found a coherent formal system underneath the condescension. This course makes the same bet about slop: that six-fingered hands and morphing logos aren’t just mistakes waiting to be fixed by a better model, but the visible seams of a specific economic and technical system, and that system has a formal grammar worth teaching the way melodrama’s is taught.

Concretely, that means three things every week of this course will ask of you, in rotation: name the tell precisely (this week and next), read the directorial choices — or their absence — behind it (week 3 onward), and follow the money that made producing it at volume worthwhile (week 4). None of those three moves is available to someone who’s still at the “this all just looks fake” stage. Getting past that stage is what week 1 is for.

Why a course, not a lecture#

A single lecture on “AI is making bad videos” would be accurate and useless — everyone already believes it, and belief isn’t a skill. What doesn’t yet exist, for most people looking at this content, is the vocabulary to say which bad, precisely, and the follow-through to ask what that specific bad is worth to whoever’s making it. That’s a semester’s worth of practice, not an afternoon’s worth of opinion, and it’s why the course is weighted the way it is: one assessment that asks you to diagnose (week 6), two that ask you to build the thing you’ve just learned to diagnose (weeks 9 and 12) — because diagnosis without production leaves you certain you’d never make this, which is a worse position for understanding it than actually trying.