← Builds / BLD-17

Cutting a dovetail joint every day for a year

What happens when you hand-cut a dovetail joint every day for a year? From a failed first joint to a Japanese sunrise dovetail, and what 100 reps teach.

BuildWoodHand tools

How quickly can you get good at something? They say mastery takes 10,000 hours. This video tests that with one skill, hand-cut dovetail joints, practised every single day for a year. It ends with me handing my work to maker legend Jimmy DiResta.

Watch the full build

Pressing play loads the video from YouTube (privacy).

On 1 January 2022 I started with a pretty poor saw, blunt chisels, a pencil and a ruler, and a belief I'd quietly held for years that I'd never be able to do this. My first joint was the worst possible failure: I cut out the wrong parts and it wouldn't go together. That's the moment most people give up, deciding they're just not good at it.

Instead I kept going. I found time by getting up early and doing 20 minutes in the shed in my pyjamas before anyone was awake. Along the way I learned why sharp tools matter and how to sharpen them, built jigs to speed things up, upgraded to a proper dovetail saw, and discovered that my square wasn't actually square.

By around 100 joints and roughly 40 hours of practice, nowhere near 10,000, the results looked good, though I was still making mistakes. You never stop making them; they just get smaller and you learn to hide them. Then boredom hit. The Goldilocks rule of learning says motivation peaks when a task is neither too easy nor too hard, so I made it harder: different sizes, contrasting woods, angles, inlays, boxes, nested boxes and nested pyramid boxes, building up to the finale, the Japanese sunrise dovetail. It took four attempts, some spread over several mornings, and the last one alone was about three hours of work.

What I'd pass on: the lessons are in the failures, so get comfortable getting things wrong. Use the Goldilocks rule to stay motivated. And watch what you tell yourself. If you say you can't do something, you won't try, so say you can't do it yet. You don't need 10,000 hours to be better than most; for me, 100 reps got me pretty close.

Inspired in part by James Clear's Atomic Habits, with thanks to Jimmy DiResta, Matt Estlea, Jonathan Katz-Moses and Rob Cosman.

More from the workshop