How is AI changing code review?
Most answers agree AI is automating or streamlining code review, shifting human focus to riskier or more complex changes, though some highlight rising AI-only reviews and improved review quality.
8 answers
Answers


Balaji Srinivasan
Senior Director Of Engineering · Linkedin
Risk Heat Map
Balaji Srinivasan says LinkedIn's next step for its AI reviewer is a risk heat map, an optional feature that scores each region of a PR's diff by predicted risk so engineers review the riskiest parts first.
ByteByteGo Newsletter · ByteByteGo Newsletter · Read

Addy Osmani
Member of Technical Staff · Anthropic
Shifting Review to Machines
Addy Osmani describes the Claude Code team's AI code review process as keeping review in place but moving most of the reading to machines, with humans pointed at the risky changes.
The AI Corner · The AI Corner · Read
Gautam Korlam
Principal Engineer · Sonar
First‑pass automated reviews
Gautam Korlam says AI is changing code review by enabling first-pass automated reviews, so humans aren't doing every single review; without that, reviewers become the bottleneck.
AI Engineer · Watch on YouTube

Katelyn Lesse
Head of API Engineering · Anthropic
AI-only PR reviews are rising
Katelyn Lesse says AI-only PR reviews are up, citing fresh Linear data that PRs reviewed solely by AI agents are a rising trend, and says she expects this to continue.
The Pragmatic Engineer · The Pragmatic Engineer · Read

Venkat Venkataramani
VP, App Infra · Openai
Agentic code reviews
Venkat Venkataramani says the way code review is done today makes less and less sense, and that agentic code reviews are now emerging to examine code changes through several lenses — making it possible for infrastructure and security engineers to review every change.
The Pragmatic Engineer · The Pragmatic Engineer · Read

Angie Jones
VP Engineering, AI Tools & Enablement · Block
Quality Has Improved
Angie Jones says the quality of code review has improved substantially from her personal experience, because AI reviewers filter out low-quality contributions before humans review them.
Quality Remarks - The Podcast · Quality Remarks - The Podcast · Listen at 30:10
Anders Hejlsberg
Technical Fellow · Microsoft
More ergonomic reviewing
Anders Hejlsberg says code review can be made more ergonomic than it is today, with AI potentially helping by explaining what the changes are.
Ryan Peterman · Watch on YouTube

Zach Lloyd
Founder / CEO · Warp
Shifting bottleneck upward
Zach Lloyd says that as agents produce code at volume, the bottleneck shifts to review—currently the biggest one—and he’s tackling it with computer-use verification and better specing so agents check spec adherence.
Dev Interrupted · Dev Interrupted · Read











