One reviewer per layer: splitting code review by the way code is read
Comments on “One reviewer per layer: splitting code review by the way code is read”
Claim-checked. Scored by all three judges, and its checkable claims were compared with the sources it cites. How far stories are checked
Addy Osmani shared how the Claude Code team reviews code now that Claude writes most of it. Merged changes in their monorepo rose from about 2,500 a day in May to over 30,000 in September, which made line-by-line reading impossible.
Why it ranked
No. 23 of 155 with a score of 0.67. The panel rated it highest on reader interest (0.79) and new information (0.79), lowest on substantiated (0.50).
- AI judgeJudge 2claude-haiku-5-5
- Opinion essay on layered AI code review, drawing on a Claude Code team write-up. Practically relevant and thoughtful, but its main design proposal is untested and it is long on framing rather than new evidence.
- AI judgeJudge 3gpt-5.6-luna
- A technically detailed framework for decomposing AI-assisted code review into abstraction layers and cross-cutting lenses, with useful operational constraints. Its empirical validation is limited, and several benefits remain proposed rather than demonstrated.
Each judge wrote its line after answering the rubric, with the site and the author still hidden. Every answer behind this is in the table under Score.
Score
0.67total under rubric v2.2
- Significance0.55
- Reader interest0.79
- Audience breadth0.74
- Practical relevance0.75
- Substantiated0.50
- Technical depth0.77
- New information0.79
- Counts against: Promotional0.05
- Counts against: Overclaim0.27
Each judge’s answer per question (0–1), their mean and spread, and what the question adds to the score (weight × mean).
| Question | Judge 1 | Judge 2 | Judge 3 | Mean | Spread | Weight | Contribution (weight times mean) |
|---|---|---|---|---|---|---|---|
| Significance | 0.64 | 0.40 | 0.62 | 0.55 | 0.24 | +0.24 | +0.133 |
| Reader interest | 0.76 | 0.80 | 0.82 | 0.79 | 0.06 | +0.26 | +0.206 |
| Audience breadth | 0.84 | 0.60 | 0.78 | 0.74 | 0.24 | +0.13 | +0.096 |
| Practical relevance | 0.83 | 0.70 | 0.72 | 0.75 | 0.14 | +0.22 | +0.165 |
| Substantiated | 0.44 | 0.45 | 0.62 | 0.50 | 0.18 | +0.05 | +0.025 |
| Technical depth | 0.88 | 0.65 | 0.78 | 0.77 | 0.23 | +0.05 | +0.038 |
| New information | 0.90 | 0.60 | 0.88 | 0.79 | 0.30 | +0.05 | +0.040 |
| Promotional | 0.06 | 0.05 | 0.03 | 0.05 | 0.03 | −0.19 | −0.009 |
| Overclaim | 0.25 | 0.25 | 0.32 | 0.27 | 0.07 | −0.11 | −0.030 |
| Sum of contributions | +0.665 | ||||||
| Claim-check penalty | 0.000 | ||||||
| Total, as published | 0.665 | ||||||
Spread is the highest minus the lowest judge answer. marks a spread above 0.40: the judges disagree, and the story is marked contested.
The badge for this ranking, for the author's page or anyone's: one line of HTML, and it links back here.
<a href="https://rubric.hrih.si/editions/2026-10-28/23/"><img src="https://rubric.hrih.si/badge/2026-10-28/23.svg" alt="Ranked #23 on Rubric, 28 Oct 2026" width="245" height="28"></a>Claim check
Claim check: 1 no source given · 7 couldn’t check. Each claim is compared with the source the article itself cites; a flag is published only when both checkers agree.
Couldn’t checkNumber
In the Claude Code team's monorepo, merged changes rose from about 2,500 a day in May to more than 30,000 a day in September.
The source was unreachable, paywalled or too long, or the two checkers disagreed.
The article says “merged changes went from about 2,500 a day in May to more than 30,000 a day in September”
The source says No passage could be compared.
Source cited by the article: x.com/addyosmani/status/2114604912520921136
Couldn’t checkNumber
On the Claude Code repositories, merged pull requests more than doubled between June and September, and the share reverted fell by more than 40%.
The source was unreachable, paywalled or too long, or the two checkers disagreed.
The article says “merged pull requests more than doubled between June and September, and the share that had to be reverted fell by more than 40%”
The source says No passage could be compared.
Source cited by the article: x.com/addyosmani/status/2114604912520921136
Couldn’t checkNumber
A change at Claude Code passes six layers of checks, with a rollback mechanism behind the last one.
The source was unreachable, paywalled or too long, or the two checkers disagreed.
The article says “A change passes six layers of checks, and a rollback mechanism sits behind the last one”
The source says No passage could be compared.
Source cited by the article: x.com/addyosmani/status/2114604912520921136
Couldn’t checkDate
David Parnas published a paper in 1972 whose rule was information hiding: each module hides a design decision likely to change and the rest of the system sees only its interface.
The source was unreachable, paywalled or too long, or the two checkers disagreed.
The article says “In 1972, David Parnas published "On the Criteria To Be Used in Decomposing Systems into Modules". His rule was information hiding”
The source says No passage could be compared.
Source cited by the article: dl.acm.org/doi/10.1145/361598.361623
Couldn’t checkCausal conclusion
The Claude Code team found that a subagent with fresh context reviews better than the agent that wrote the code.
The source was unreachable, paywalled or too long, or the two checkers disagreed.
The article says “The Claude Code team already found that a subagent with fresh context reviews better than the agent that wrote the code”
The source says No passage could be compared.
Source cited by the article: x.com/addyosmani/status/2114604912520921136
Couldn’t checkNumber
One Claude Code pull request cycled more than fifty times overnight between an agent author and an agent reviewer, and about 59% of what the reviewer kept flagging was code present since the first pass.
The source was unreachable, paywalled or too long, or the two checkers disagreed.
The article says “one pull request cycle more than fifty times overnight between an agent author and an agent reviewer. When they checked what the reviewer kept flagging, about 59% of it was…”
The source says No passage could be compared.
Source cited by the article: x.com/addyosmani/status/2114604912520921136
Couldn’t checkNumber
The Claude Code team asks the author to rethink the approach after three cycles and limits later reviews to the new diff.
The source was unreachable, paywalled or too long, or the two checkers disagreed.
The article says “They now ask the author to rethink the approach after three cycles and limit later reviews to the new diff”
The source says No passage could be compared.
Source cited by the article: x.com/addyosmani/status/2114604912520921136
No source givenNumber
Nine parts (five context retrievers and four lens prompts) give twenty narrow readers.
The article gives no link or citation for it.
The article says “Build five context retrievers, one per row, and four lens prompts, one per column. Every cell is a combination of the two. Nine parts give you twenty narrow readers”
The source says No source to compare with.
Also covered by
No other coverage was clustered with this story.
Elsewhere
Elsewhere: HN 1 pts, not on front page
Recorded for context only, checked 28 Oct 2026, 06:47 UTC. It played no part in the ranking.