The Review Tax Index — Ambera Forge<br>The Index<br>The Review Tax Index<br>The claim you will hear is that AI-written code costs more review everywhere. The data says something more useful: the cost is a split — it lands hard on some repositories and not at all on others. This page measures where.
Repositories read<br>5,388
Merged PRs analyzed<br>444,225
Zero attribution<br>56%<br>no detectable agent authorship at all
Agent-native repos<br>15%<br>of attributed repos, 20%+ of merged work carries attribution
The split<br>The measures run over the 2,353 repositories with detectable attribution and at least 10 analyzed PRs — a review-cost ratio needs an attributed side to exist. Within each, attributed work is measured against the rest of the same repository — same reviewers, same conventions, same era. A multiple above 1x means the attributed side took more; below means it took less.<br>Time to merge, attributed / rest<br>354<br>repos above 1x
874<br>at or below 1x
0.5x<br>median
29% of comparable repos pay more review on attributed work; 71% pay the same or less. Middle half of the population: 0.2x to 1.2x.
Review rounds, attributed / rest<br>repos above 1x
13<br>at or below 1x
0.6x<br>median
32% of comparable repos pay more review on attributed work; 68% pay the same or less. Middle half of the population: 0.2x to 1.2x.
Where the tax lands<br>Median detected share across the attributed repositories is 3% — and that is a floor, not a share: attribution comes only from marks a tool leaves on a commit, so inline-completion work is invisible by construction. Across the full population the median is zero — most repositories read carry no attribution at all.
By primary language<br>SegmentReposMedian detected shareTime to mergeReview roundsTypeScript4423%0.4x—Python3614%0.5x—Go3023%0.5x—Rust2372%0.7x—JavaScript2172%0.4x—C++1703%0.4x—Java1312%0.3x—C#1296%0.3x—C982%0.4x—PHP872%0.3x—Swift683%0.5x—Kotlin532%0.4x—Ruby493%0.5x—<br>A dash means the segment has too few comparable repos for that measure to publish — a median over a handful of ratios is noise wearing a number.
By project size<br>SegmentReposMedian detected shareTime to mergeReview rounds50k+ stars2384%0.4x—20k-50k stars6243%0.5x—10k-20k stars6883%0.4x—5k-10k stars7753%0.5x—2k-5k stars261%——<br>A dash means the segment has too few comparable repos for that measure to publish — a median over a handful of ratios is noise wearing a number.
How to read this<br>Attributed vs rest, never AI vs human. The comparison group contains AI-assisted work that leaves no trace, which narrows every gap shown here rather than widening it.<br>The sample self-selects twice. These are repositories someone chose to scan plus a curated backfill of well-known projects — a population of repos people care about, not a random sample of software.<br>Withheld beats invented. Repos below 10 analyzed PRs do not count; segments below 12 comparable repos publish no median.<br>The Index republishes quarterly. Repos already read keep their history, so future editions carry trend lines — not just where the tax lands, but where it is moving.<br>Full method and every limitation: /methodology. Place your own repo against this population: /scan.