When Josh Comeau wrote about AI having a multiplying effect on technical skills, he captured something developers feel daily. AI doesn't replace your ability to code. It amplifies what you already know.
Think of it like a senior engineer who reviews your PRs instantly, writes boilerplate for you, and suggests edge cases you missed. That's the multiplying effect. Your 10 years of React experience become 10 years plus an AI that can generate a complex component in seconds.
How the Multiplication Works
AI tools like GitHub Copilot, Cursor, and Claude operate on your existing knowledge. If you understand TypeScript generics, the AI can generate type-safe API clients faster than you can type them. If you know SQL, the AI can write complex joins from a natural language prompt.
The key insight: the AI is only as good as your ability to evaluate its output. A junior developer might accept a buggy suggestion. A senior developer spots the flaw, corrects it, and moves on. The multiplier is your skill, not the AI's.
Concrete Examples
Example 1: Debugging a Race Condition
You have a Node.js service with a race condition in an async loop. Instead of tracing through logs manually, you paste the code into Claude and ask: "Find the race condition and suggest a fix using Promise.allSettled." The AI returns a solution in 10 seconds. You review it, adjust the error handling, and deploy. Time saved: 30 minutes.
Example 2: Writing Unit Tests
You have a React component with 5 props, 3 states, and 2 side effects. You ask Copilot: "Write Jest tests covering all states, including loading and error." It generates 50 lines of tests. You modify the mock data and add one edge case. Time saved: 20 minutes.
Example 3: Refactoring Legacy Code
You inherit a 2000-line PHP file. You ask Cursor to "extract database queries into a repository class." It refactors the file, creates the class, and updates all references. You review for correctness and rename a few methods. Time saved: 2 hours.
