The article critiques the resurgence of lines of code (LOC) as a productivity metric in software development, particularly with the rise of AI in coding. Historically, LOC has been deemed a poor measure of software quality and progress, with notable figures like Dijkstra and Gates condemning it. By 2023, LOC was largely considered an "input metric" with little value, yet the advent of AI has revived its usage. Tech leaders tout increasing percentages of AI-generated code—Google claims 25% and Meta predicts 50%—but there’s little emphasis on the quality or effectiveness of this code. The focus remains on volume rather than meaningful metrics, suggesting a return to problematic practices. The article outlines how AI makes LOC even more susceptible to manipulation. Unlike human developers, AI can churn out vast amounts of code effortlessly, making the cost of generating lines virtually zero. This leads to a scenario where the quantity of code becomes meaningless. Goodhart's Law applies here: as LOC becomes the target for AI tools, it fails to provide any real insight into productivity or code quality. Data supports this concern, revealing a rise in duplicate code and a drop in refactoring efforts. For instance, copy-pasted code increased from 8.3% to 12.3% in just a few years, indicating a decline in code quality. Moreover, the acceptance rate of AI-generated code is now monitored, but even this metric inherits flaws from LOC. Acceptance doesn't guarantee quality; developers may accept suggestions that aren't genuinely useful. Studies show that while developers feel faster with AI, they actually take longer to complete tasks. Trust in AI's accuracy has declined significantly, with many developers spending more time correcting flawed AI-generated code than they save in initial writing. The article paints a grim picture of a software industry that's prioritizing volume over quality, resulting in more code that is poorly understood and laden with flaws.