Slop Pipeline
Design the automation that could produce one slop archetype at scale, and make the economic case for why anyone would bother running it
Due: 19 April 2027 · Weight: 25%
The brief#
Design the machine that would make one slop archetype at scale, and argue for why it would be worth turning on.
The Slop Teardown asked you to explain why something exists. This one asks you to build the thing that would make more of it, cheaper, without a director in the loop for every unit. Pick one archetype — the AI history explainer with the affectless narrator, the product demo nobody filmed, the ambient “cozy” channel that never repeats a scene and never says anything — and design the pipeline that would produce it on a schedule, not the one clip that proves you can.
This is a making assessment, but the thing you are making is the logic, not a finished video. Treat it like an actual production system: what triggers a new unit, what generates the script or shot list, what generates the footage or voice, what checks it before it ships, and where a human still has to look at the output before it goes out. A pipeline with no human checkpoint at all is either a lie about how these systems really run, or an admission that nobody is watching quality — say which, if that is the position you are taking.
The economics half is not decoration. A pipeline that costs more to run than it earns in attention or ad revenue is a hobby, not the thing this assessment is asking for. Make the case with numbers, even rough ones: cost per render, renders per day, expected views or conversions per unit, and the point at which the pipeline pays for itself. A confident number you can defend beats a vague one you can’t.
What you submit#
Three linked pieces, submitted together:
- a workflow diagram (n8n-style, or any notation that shows nodes and the edges between them) covering trigger through to published output
- a prompt template library: the actual reusable templates the pipeline would run, not a description of what a template might contain
- a short written economics memo (400–600 words) covering per-unit cost, the scheduling cadence, and the break-even argument
The marking: block above weights the workflow logic and the economics memo
most heavily — a good diagram with no financial case, or a financial case with
no working diagram, is half the assessment.
The spec
The brief poses the problem and leaves room for a response. The spec is the fixed part: what a marker checks. Some lines can be checked mechanically; the rest call for judgement.
- names one specific slop archetype and holds to it across every artefact the pipeline is designed to produce
- the workflow diagram shows every node from trigger to published output, including the steps where a human would still have to intervene
- the prompt template library contains templates specific enough that a stranger running them would get output recognisably in the target archetype, not generic video
- the scheduling plan states a concrete cadence and volume (units per day or week), not "regularly" or "as needed"
- the economics memo gives real or realistically estimated numbers — cost per unit, time per unit, and an assumption about revenue or engagement — and states the break-even case plainly
- a reader who has not seen the brief can tell, from the diagram alone, where the pipeline would actually break
- submitted by the deadline, in the format named below
How it is marked
| Criterion | Weight |
|---|---|
| Workflow logic | 30% |
| Prompt template quality | 25% |
| Scheduling plan | 15% |
| Economics memo | 30% |