A common myth in software development is that more output automatically leads to better results. The idea is that if we just code faster or produce more lines, we’ll create something valuable. But that’s not how it works. Gergely Orosz, a well-known writer at The Pragmatic Engineer, recently pointed out that working long hours or rushing to produce code doesn’t necessarily lead to innovation or quality.

He highlighted the “996” work culture—working from 9 a.m. to 9 p.m., six days a week—popular among some Chinese tech companies. Orosz argued that most of these companies aren’t creating anything truly original. Instead, they often just copy or rehash existing ideas. The intense schedule doesn’t boost creativity; it just burns people out. More work doesn’t mean better work. It leads to more code, but not smarter, more useful code.

The Hidden Costs of Churning Out Code

The more we produce, the more problems we tend to create. Orosz’s point ties into a recent analysis by GitClear, which looked at 153 million lines of code. It found that “code churn”—changing or deleting lines of code within two weeks—has gone up. This means developers are rewriting or fixing their code more often. The trend is toward copy-pasting and less refactoring, which is a sign of sloppy or rushed work.

AI tools, like those that help generate code, can speed up development by as much as 55%. But faster doesn’t mean better. Instead of building smarter, we’re just making more. The result? More bloated, complex codebases that are hard to maintain, secure, and understand. Having lots of code isn’t a sign of progress; it’s a liability.

The Real Driver of Good Software: Clarity and Thought

Software development isn’t a race to produce the most lines. It’s about making smart decisions. Experienced developers know that it’s less about typing and more about thinking. They focus on what code is truly needed, not just what can be quickly written. Every line of code adds complexity and potential problems down the line.

Charity Majors, CTO of Honeycomb, emphasizes that senior engineers excel at understanding and managing large systems. They translate business needs into effective technical solutions. More code means more security risks and more debugging. When AI is used to push out code quickly, it often results in a bigger mess. It’s like building a house with more walls but no plan—eventually, it falls apart.

Orosz’s critique about 996 companies producing copies is also about having no space to think. When developers are overwhelmed reviewing endless AI-generated pull requests, they lose the chance to step back and consider whether they’re building the right thing. Instead of architects, they become janitors, cleaning up after an endless stream of low-quality code.

Using AI as a Tool, Not a Shortcut

This isn’t to say AI is bad. On the contrary, it’s a powerful tool. Harvard professor Karim Lakhani points out that AI will not replace humans, but people working with AI will replace those without it. The key is how we use AI. If we treat it as a way to do the boring work—writing tests, generating boilerplate, updating docs—it can free up time for more important tasks.

The goal shouldn’t be to add more features or write more code. Instead, we should focus on clarity, simplicity, and quality. Developers need to spend time framing the right problems, not just rushing to solve the wrong ones. They need to be ruthless about deleting unnecessary code and making systems easier to understand. When things break—and they will—developers must understand their systems deeply enough to fix them quickly.

Training is also critical. Less experienced developers must learn to use AI effectively without blindly trusting it. The aim isn’t to fight against automation but to use it in a way that enhances human judgment. The danger of turning AI into a shortcut is that it produces a mountain of brittle, forgettable code. That’s the 996 mindset transferred into machines: more hours, more code, less thought.

In the end, better software doesn’t come from more lines of code. It comes from smarter, clearer, more thoughtful work by humans who have the space to innovate. AI can handle the repetitive tasks, freeing developers to focus on what truly matters—solving problems, designing elegant solutions, and understanding their systems inside out. That’s how we build software that lasts.

Inspired by

• https://www.infoworld.com/article/4094801/software-development-has-a-996-problem.html

Sources

• gmpg.org
• wired.com
• networkworld.com
• gitclear.com
• hbr.org