---
format: "aidr-story-markdown/v1"
id: "1ed525b777a52b12b89e104750e4fc3a315188fefe7ccb95fbd7298d6a85927a"
canonical_url: "https://aidr.today/1ed525b7?lang=en"
title: "Harvard researchers find AI coding agents shifted the bottleneck to code review"
lang: "en"
requested_lang: "en"
available_langs: ["en","vi"]
translation_fallback: null
fallback_fields: []
published_at: "2026-11-17T21:16:12.000Z"
category: "Tools"
topics: ["agent","devtools"]
source_urls: ["https://news.ycombinator.com/item?id=50538303"]
summary: "Harvard study: AI coding agents raised code output 30 percent; finished work did not measurably rise, and getting code through review took 49 percent longer."
---

# Harvard researchers find AI coding agents shifted the bottleneck to code review

> [Open the canonical story](<https://aidr.today/1ed525b7?lang=en>)

**Published:** 2026-11-17T21:16:12.000Z
**Category:** Tools
**Topics:** agent, devtools

## Summary

Harvard study: AI coding agents raised code output 30 percent; finished work did not measurably rise, and getting code through review took 49 percent longer\.

## Sources

- [Discussion](<https://news.ycombinator.com/item?id=50538303>)

