How to Practice Writing (and Write Stories With AI)

Deep-dive on deliberate practice for writers, using Anders Ericsson’s framework to isolate fiction sub-skills, plus a demonstration of how manual practice outputs become AI training data.

Key Ideas

Core argument: “Writing Fiction” is an umbrella term containing separable sub-skills (story seeds, outlining, worldbuilding, character development, pacing, narrative voice, word choice). Effective improvement requires unbundling, isolating one sub-skill, creating constrained exercises, and deploying volume. Demonstrates this with a Story Seed exercise using a 5-part template (Character Desire, Weakness, Danger, Action, Change) and shows three original fiction seeds. The leverage insight: manual exercise = AI prompt, manual reps = AI training data. Once you reach proficiency, you have both the framework and examples to automate that skill with AI forever.

RDCO Mapping

The deliberate-practice-as-prompt-engineering thesis is the strongest idea in the batch. Directly applicable to Sanity Check content about skill development in the AI age. The principle that you must be able to articulate a skill manually before you can automate it is a foundational insight for any AI-assisted workflow.