Skip to content
HN On Hacker News ↗

One reviewer per layer: splitting code review by the way code is read

▲ 31 points • 12 comments • by djordjep • 1w ago • HN discussion ↗

Pangram verdict · v3.3

We believe that this entire text is AI.

100 %

AI likelihood · overall

AI
0% human-written 100% AI-generated
SEGMENTS · HUMAN 0 of 4
SEGMENTS · AI 4 of 4
WORD COUNT 127
PEAK AI % 100% · §1
Analyzed
Oct 29
backend: pangram/v3.3
Segments scanned
4 windows
avg 32 words each
Distribution
0 / 100%
human / AI fraction
Verdict
AI
Pangram v3.3

Article text · 127 words · 4 segments analyzed

Human AI-generated
§1 AI · 100%

Addy Osmani shared an article on how the Claude Code team reviews code now that Claude writes most of it. The numbers explain why they had to change. In their monorepo, merged changes went from about 2,500 a day in May to more than 30,000 a day in September. Nobody reads that line by line. On the Claude Code repositories, merged pull requests more than doubled between June and September, and the share that had to be reverted fell by more than 40%. They did not stop reviewing.

§2 AI · 100%

They layered it.

§3 AI · 100%

A change passes six layers of checks, and a rollback mechanism sits behind the last one. Each layer catches some bugs and misses others. Few bugs get through all six.

§4 AI · 100%

I like this system a lot.