Mark FinkleAgentic coding creates opportunities and challenges to our existing ways of building software. Code reviews are one of the many topics to consider:
Agents write a lot of code -> People review that code
* Will we have capacity to review the code?
* Will the reviews catch defects?
Before we jump into finding ways to review all that code, let's consider the act of reviewing code — and what parts yield the most value and require people.
🌶️ It's not about finding defects
https://starkravingfinkle.org/posts/2026/03/coding-agents-who-will-review-all-that-code/
Agentic Coding: Who Will Review All That Code?
Code reviews were never really about catching bugs. The research is clear. As agentic coding floods our repos with pull requests, we need to lean into what reviews are actually good at, and get better at what happens before the code gets written.
Dave Townsend