← Back to archive

Why lines of code are a meaningless productivity metric

27 June 2025 · 48 views · 0 likes · 0 reposts · 104 blog views

AI coding history product productivity

The timeless absurdity of lines of code as a metric.

Back in 1982, Apple's Lisa team tried to track developer productivity by counting lines of code written each week. Bill Atkinson, the brilliant mind behind QuickDraw, found this metric ridiculous. When he optimized the region calculation engine - making it 6x faster while eliminating 2,000 lines of code, he submitted his weekly report with "-2000" in the lines of code field.

The managers stopped asking him to fill out the form....

I witnessed a similar obsession with code quantity during my early days as a software developer. Management loved the idea of measuring output through line counts, completely missing the point that good code is often about doing more with less. The best developers I knew were the ones who could solve complex problems with elegant, concise solutions, not those who churned out verbose, repetitive code.

Fast forward to today's AI era ... with AI assistants generating verbose, often duplicate code, hitting those old-school KPIs would be trivially easy - perhaps 10x easier than before. An AI can effortlessly produce hundreds of lines of boilerplate/redundant functions/over-engineered solutions that would have made any line-counting manager ECSTATIC.

The irony is that the metric that was always meaningless has become even more meaningless in an age where quantity of code has never been easier to generate while QUALITY and EFFICIENCY remain as challenging as ever.

Why lines of code are a meaningless productivity metric

Originally posted on X

PS: I'm pretty active on 𝕏 if you'd like to follow along

Related posts

Archive

2026

2025

2024

2018