Repeatability
Medium
The structural task — 500-word premise, near-future setting, memory editing theme — is consistent, but each instance demands unique creative choices in character, tone, and ethical framing. It's not a fill-in-the-blank template, so automation gains are partial.
Ambiguity Tolerance
Low
Success criteria are highly subjective: what counts as 'compelling,' 'ethically nuanced,' or 'well-motivated' is a matter of taste. There is no objective signal for when the premise is done or good enough, making it hard for an agent to self-evaluate reliably.
Data & Tool Availability
High
The agent needs only a language model and the task description — no external APIs, files, or permissions required. Everything needed to generate prose is self-contained.
Error Cost
Low
A weak or generic premise is easily discarded or revised; no irreversible harm results. The stakes are low and the output is fully reversible, which strongly favors automation.
Human Judgment Required
High
Crafting ethical nuance that feels genuine, avoiding clichéd sci-fi tropes, and creating character motivations that resonate emotionally all require the kind of literary taste and lived moral intuition that current AI consistently underdelivers on.