💡AI Has Made Code Cheaper to Produce
Your old metrics are suddenly outdated
TL;DR
The cost of producing code has plummeted, but traditional metrics like velocity and pull requests no longer accurately reflect engineering output. Good engineers have always said to measure outcomes over effort.
The cost of writing code has collapsed due to AI tooling, making old metrics like velocity tracking and pull request counts obsolete. However, the shift doesn't mean that every claim about faster development is true. Code review, documentation, and onboarding remain crucial processes. The key takeaway: practices resting on high coding costs should be re-evaluated as new metrics are needed to measure actual engineering output.
Key Points
The cost of producing code has dropped dramatically with AI tooling, challenging old assumptions about development speed and efficiency.
Velocity tracking, pull request counts, and ticket closures were never perfect measures of engineering output; they're even less relevant now.
Code review, documentation, and onboarding processes remain essential despite claims that these are becoming obsolete due to faster coding.
Mechanical verification is getting cheaper with AI, but semantic verification still requires human judgment for complex checks.
The boundary between machine-checkable and semantic verification has become a strategic variable in engineering practices.
Why It Matters
If you're relying on velocity tracking or pull request counts to measure your team's productivity, those metrics are now outdated. Focus instead on business outcomes and system health as good engineers have always advised.
Frequently Asked Questions
Why does this matter?
If you're relying on velocity tracking or pull request counts to measure your team's productivity, those metrics are now outdated. Focus instead on business outcomes and system health as good engineers have always advised.
What happened?
The cost of producing code has plummeted, but traditional metrics like velocity and pull requests no longer accurately reflect engineering output. Good engineers have always said to measure outcomes over effort.
Comments
Be the first to comment
Enjoyed this article?
Get it daily. 7am. Free. Reads in 5 minutes.
Join 2,279 builders reading daily.