Agentic Coding Risks Skipping the 'Reps' That Build Engineering Judgment

Addy Osmani cites a 2026 Anthropic study: junior engineers using AI assistants scored 50% on a follow-up quiz vs. 67% working by hand — but those who asked conceptual questions (vs. treating the model as a 'code vending machine') did better. Frames deep domain expertise and applied judgment as the two abilities agent-based work still requires, and recommends capturing lessons as tests/lint rules/memory rather than letting them disappear with the chat session.