Rachel Laycock, CTO at Thoughtworks, takes the other side of the “what is code review even for” debate — a written response to DX’s Brian Houck after the two disagreed on a Code Remix panel. Her argument isn’t that AI broke code review. It’s that review was quietly doing five jobs at once, and the AI code-volume spike made that unsustainable.
The overloaded code review:
- Review carried quality gate, security check, architecture review, mentoring, knowledge-sharing, and ownership — all at once
- Meta’s human-landed diffs grew 106% in a year and median PR size is up 64% — no human review queue absorbs that
- Her fix is not “review faster”; it’s stop waiting until review to have the important conversations
Shift the judgment left:
- Alternative solutions: explore before implementing, not after
- Knowledge transfer: pair while someone is reasoning, not after they finish
- Junior learning: work alongside experienced engineers in design sessions before any code exists
- Collective ownership: build and operate together — pairing, mobbing, whiteboards — rather than reading diffs afterwards
- Architecture: design together, then encode constraints as executable fitness functions
- Formatting, linting, known security issues: automate — “we really shouldn’t still be arguing about whitespace in 2026”
Review by exception — the changes that still deserve experienced human eyes: a fundamental architectural change, a sensitive security boundary, a change with huge blast radius, an unfamiliar part of a critical system, or anywhere the team says “I’m not confident about this.”
An AI reviewer standing in for the human isn’t the answer either — that’s “automating the ceremony rather than questioning why the ceremony exists.” If an agent produces ten times the code but every line queues for a senior engineer, you’ve built a backlog and a new bottleneck, not a ten-times organisation.
She concedes Houck’s cognitive-debt worry is real but doesn’t think mandatory pull requests defend against it. The closing line is the thesis: “We need engineers to understand systems, not diffs” — which means designing, pairing, and operating together far earlier than review.