Cultivating Instinct
summary
Argument: train the perceptual side of expertise with short classification trials plus feedback using huge noisy datasets so brains learn which features matter. Evidence: 3 hours of perceptual training raises pilots from ~50% to ~80% accuracy, chick-sexing program yields 97% accuracy at 1,200 chicks/hour, and concrete engineering wins include a test-helper patch that reduced DB calls by 40%. Apply to coding by targeting specific perceptual skills (race conditions, caching impact, debugging fingerprints).
AI-written summary · about
chapters (9)
- 0:00Intro
- 3:23Expertise
- 4:10Intuition
- 5:50How do you teach intuition?
- 8:05Science
- 10:27How we extract information
- 17:51Perceptual learning in code
- 25:05An amateur's guide
- 30:15Outro
