3 Minutes
Linus Torvalds, the creator of Linux, has bluntly criticized a hiring and performance metric reportedly used by Elon Musk — counting lines of code. In a wide-ranging interview on the Linus Tech Tips YouTube channel, Torvalds dismissed the idea as not just misguided but outright "incompetent." His comments have reignited a debate across developer communities about what really measures software skill and productivity.
Why counting lines of code is a dangerous shortcut
During the interview, the host mentioned an old programming metric: total lines of code. Reports suggested that after Elon Musk took over Twitter (now X), engineers were asked to print out their source files — and those with fewer printed lines were reportedly at risk of being let go. Torvalds didn't mince words. He called using line counts to evaluate engineers "pure incompetence," adding, "Anyone who thinks like that is too stupid to work at a tech company."
That stinging assessment struck a chord. The underlying problem is simple: quantity is not the same as quality. Measuring productivity by how many lines someone writes encourages verbosity, fragile systems, and needless complexity. Well-designed software often does more with less — and that economy of code is a mark of craftsmanship, not laziness.
Developers react: a near-universal eye roll
Across Reddit and other developer forums, programmers supported Torvalds' stance. Comments ranged from bemused to furious: "Even a first-term computer science student knows that counting lines of code is the dumbest metric." The community argues that such a rule incentivizes bloated solutions and generates more bugs — the exact opposite outcome companies want when scaling products or stabilizing a platform.
Examples that make the point
- One concise algorithm of 100 lines can outperform a 1,000-line patch that duplicates logic and hides intent.
- Refactoring often reduces line count while increasing clarity and maintainability — would a strict LOC policy penalize that?
- Automated formatting or verbose logging can inflate line counts without improving functionality.
Imagine rewarding writers for the length of their essays rather than the clarity of their argument. The same faulty logic applies when you reward programmers for bulk instead of craft.
What companies should measure instead
So if not lines of code, what should managers track? Quality metrics such as code review feedback, bug rates, lead time for changes, test coverage, and the ability to deliver reliable features matter far more. Soft factors — collaboration, architectural thinking, and the knack for simplifying complex problems — also separate strong engineers from mediocre ones.
Torvalds' critique is a reminder for tech leaders: pick metrics that encourage maintainability and thoughtful design. Otherwise, you risk rewarding behavior that increases technical debt and undermines product stability — a costly mistake for any technology-driven company.
Leave a Comment