Thoughtworks CTO: maybe we shouldn't be reviewing all this code anymore

rseroter · x · 2026-09-03

Thoughtworks CTO Rachel Laycock argues on martinfowler.com that code review has been overloaded for years — quality gate, security check, architecture review, mentoring, knowledge sharing — and AI-generated code at scale (Meta: +106% significant LOC per landed diff; DX: PR size +64%) forces teams to split those responsibilities rather than automate review wholesale. A thoughtful response to DX's Brian Houck in an ongoing debate about the future of code review.

Original post →

More from coding & agent

coding & agent channel →