Copy bravely, then mutate the parts you can’t reproduce

Imitation is how humans learn complex skills. From handwriting to guitar, we copy, then we adapt. The mistake is thinking that copying is the end goal. It’s not. Copying reveals structure, structure reveals decisions, and decisions teach you what to do when you’re on your own. Good copying is reverse engineering, not forgery.

In skill research, this sits under deliberate practice and cognitive modeling. You observe an expert performance, break it into parts, and measure your attempts against that standard. A jazz student copies a solo note for note to feel the timing in their hands. On the second pass, they change tempo and the limits of their technique show up like highlighter marks. A developer rewrites a small open-source function, then imposes a constraint like no external libraries. The missing comfort forces new solutions.

The gap between what you tried to do and what you could do today is valuable data. That mismatch is not a failure, it’s a fingerprint. One writer I worked with copied the structure of a film review to write about a board game. Their sentences refused to stay as short and clipped as the original. We leaned into their longer rhythm and used subheads to keep pace. The piece sounded like them and still did the job.

Underlying mechanisms include variability of practice, which improves transfer, and error-based learning, where noticing discrepancies accelerates skill formation. There’s also identity work at play. When you purposefully deviate, you make a choice that reflects your taste. I might be wrong, but trying to “be original” first often produces vague work. Copy first to learn, then mutate your mismatches into style.

Choose one short exemplar and outline its structure so you can see how and why each part works. Rebuild it by hand once to feel the mechanics, then do it again while forcing one constraint like tempo or palette to flush out your natural differences. List every mismatch you notice and pick one to amplify into a variation that solves the same problem in your voice. Schedule this as a focused sprint so you finish both passes in one or two sessions and end with a small piece you can share or refine. Put the first session on your calendar now.