Why measuring AI-assisted coding by lines of code is a flawed metric
By
David Curlewis
Toasted golden, schmeared with insight. Top of the rack.
Summary
The article argues against using vanity metrics like lines of code, PR counts, and commit frequency to measure developer productivity, especially in the context of AI-assisted coding. It draws a parallel between the industry's past lessons (learning that LOC is a poor metric for human developers) and the current trend of applying similar flawed metrics to AI-generated code. The author advocates for outcome-based measurements—what actually ships, what it does for customers, revenue, and reliability—rather than volume-based vanity metrics.
Key quotes
· 3 pulledLines of code, PR counts… we spent a couple of decades learning these are stereotypically bad ways to measure a developer.
One of them writes 40% more lines of code than the other. Is that developer better? More impactful for the business?
You'd want to know what actually shipped. What it did for customers, for revenue, for reliability.
You might also wanna read
Why measuring developer productivity by AI token usage is a flawed metric
The article criticizes the trend of managers using AI token usage as a metric to measure software developer productivity. It argues that thi
dailykos.com·10d ago
AI-Powered Code Review: A Framework for Agentic Workflows in Software Development
This paper examines the evolution of code review practices and proposes a vision for AI-powered, agentic code review workflows. It argues th
arxiv.org·5d ago
AI's Impact on Software Engineering: Evolution or Replacement?
The article explores the complex relationship between AI tools like ChatGPT and software engineering, examining whether AI represents the en
The Verge·9mo ago
AI-Assisted Development Demands a Rethink of Testing Strategy, Not Just Test Automation
The article explores the strategic decision teams face after merging AI-assisted code: whether to add traditional browser tests, use AI tool
dev.to·8d ago
AI Won't Replace Developers, But Will Make Poor Thinking Costly
The article argues that AI-assisted coding is becoming the new standard in software development, but the real competitive advantage will shi