Norvig's Antidote to '24-Hour' Programming Books: Give It Ten Years

Peter Norvig pushes back on the flood of “teach yourself X in 24 hours” books, arguing that mastery of programming—like chess, music, painting, or any serious craft—takes roughly ten years, or the 10,000 hours Malcolm Gladwell later popularized. A weekend of study can buy superficial syntax familiarity, enough to write Basic-style code in C++, but not the deeper understanding of what a language is genuinely good and bad at. Citing decades of expertise research (Bloom, Hayes, Ericsson) and examples from Mozart to the Beatles, he stresses that the mechanism isn’t mere repetition but deliberate practice: repeatedly tackling tasks just beyond your current ability, getting real feedback, and correcting mistakes.

His prescription for becoming a good programmer is concrete and social rather than book-driven. Program because it’s fun enough to sustain a decade of effort; learn by doing; and above all talk with other programmers and read their code—he rates that above any book or course. Work on projects where you’re the best (to learn leadership) and where you’re the worst (to learn from masters), and maintain code others wrote to understand what makes software survivable. He also recommends learning at least half a dozen languages spanning different paradigms—class-based, functional, syntactic, declarative, and parallel—so each reshapes how you think.

A formal computer science degree helps with credentials and depth but isn’t required; as Eric Raymond notes, studying brushes won’t make you a painter. The through-line is that there are no shortcuts to real skill, only sustained, feedback-driven work over years—a message aimed squarely at an industry perpetually looking for the fast path.