developer-tools software-development code-quality ai-impact

Lines of Code Are Back (and It's Worse Than Before)

thepragmaticcto.com

February 12, 2026

12 min read

🔥🔥🔥🔥🔥

52/100

Summary

Lines of code has resurfaced as a metric in the software industry, despite longstanding criticism of its effectiveness. Dijkstra previously condemned it as a costly measure that promotes poor coding practices.

Key Takeaways

The software industry has returned to measuring productivity by lines of code (LOC), despite long-standing consensus that it is a poor metric for assessing programming quality.
Major tech CEOs report increasing percentages of AI-generated code, with estimates ranging from 25% at Google to 90% at Anthropic.
The industry produced 256 billion lines of AI-written code in 2024, which is often celebrated as progress without consideration of the quality or effectiveness of that code.
Goodhart's Law illustrates that when LOC becomes a target for productivity, it leads to poor coding practices, a problem exacerbated by the rise of AI in software development.

Read original article

View story →

Coverage & discussions

Hacker News · thepragmaticcto.com

Lines of Code Are Back (and It's Worse Than Before) ↗

Community discussions

Hacker News discussion ↗

Historical count: 5 · 5 saved in sample · Historical sample

Analysis ready

Community Sentiment

Mixed

Positives

Lines of code can serve as a useful metric for estimating project complexity, helping developers gauge the scale of their work effectively.
Reframing lines of code as a measure of efficiency, such as achieving significant functionality in fewer lines, highlights impressive coding skills and can inspire better practices.

Concerns

AI-generated code often tends to be verbose and bloated, indicating a need for refinement and suggesting that initial outputs may lack quality.
The reliance on lines of code as a measure of productivity can be misleading, as it does not account for the quality or efficiency of the code produced.