Amid concerns that AI coding tools like Claude Code might replace junior engineers, Anthropic's Cat Wu argues they can actually help less-experienced developers deepen their understanding. Speaking at AI Codecon, Wu outlined practical ways juniors can use these tools to accelerate learning.
"Think of Claude Code as your intern that you're trying to level up."
Wu suggests starting by asking Claude Code all the "dumb" questions. Unlike coworkers, the AI is always available and never judges. If a response seems wrong, take a screenshot and discuss it with a senior engineer to grasp the nuances. Then, use that feedback to update the CLAUDE.md file, improving the codebase for everyone.
Once comfortable, juniors can fix low-risk "paper cut" problems to learn the tool's strengths and weaknesses. Finally, they can support senior engineers by using Claude Code to thoroughly test AI-generated pull requests and teach it to test code during creation.
Wu emphasizes that these tools augment rather than replace junior engineers, turning them into more effective contributors.