Why reading is your analog superpower in the age of AI? (mattilehtinen.com)

As AI automates more routine coding, this piece argues reading—especially deep, book-based reading—is an “analog superpower” that gives developers a durable edge. Evidence shows fewer developers learn from books (Stack Overflow: 54.5% in 2022 → 50.3% in 2024), creating a gap that readers can exploit. Many tech leaders (Gates, Nadella, Pichai, Bezos, Zuckerberg) read voraciously, and small habits—like one book a year—can already outpace many peers. Behavioral data reinforce the risk of passive consumption: average social‑media use is 2h23m/day and reported attention spans fell from ~150s in 2004 to ~47s today. Technically, deep reading builds the mental models, vocabulary and patience that matter when AI produces code but not judgment. Reading strengthens working memory, comprehension, and cross‑domain analogies (studies: weekly readers far less likely to suffer age‑related memory decline; 30 minutes of reading reduces stress comparable to yoga), and improves the precision needed for architecture decisions, code communication, pull requests and prompt engineering. Practical advice: don’t replace practice—code regularly—but pair it with focused reading (margin notes, reflective questioning) to sharpen problem‑solving, design thinking and your “communication API” with both humans and AI.