Mastodawn

Cursor (@cursor_ai)

Cursor의 코드 리뷰 에이전트가 PR 활동에서 학습해 실시간으로 자기 개선할 수 있게 되었습니다. PR이 머지될 때쯤 발견된 이슈의 78%가 이미 해결된다는 점을 강조해, 개발 워크플로에 실질적인 품질 향상 효과를 보여줍니다.

https://x.com/cursor_ai/status/2041969870234120231

#cursor #codereview #agenticai #developertools #ai

Cursor (@cursor_ai) on X

Cursor's code review agent can now learn from activity on PRs to self-improve in real time. 78% of issues found are resolved by the time the PR is merged.

X (formerly Twitter)

Zenn Trends 2d ago

🕐 2026-04-08 22:18 UTC

📰 全PRの83%をAIレビューだけでマージできるようにした (👍 241)

🇬🇧 Kauche automated 83% of PR merges using AI review plus CI, then improved review quality by nightly tuning the rules behind Claude Code Action.
🇰🇷 카우셰는 AI 리뷰와 CI만으로 PR의 83%를 자동 머지하도록 만들었고, Claude Code Action 규칙을 매일 밤 개선해 정확도를 끌어올렸습니다.

🔗 https://zenn.dev/kauche/articles/e051583461c181

#AI #CodeReview #Zenn

全PRの83%をAIレビューだけでマージできるようにした

Zenn

Steven Carneiro 3d ago

AI code overload:
#aiassistant #genai #codegeneration #codereview #overload #technology #futurism
🤖

https://www.nytimes.com/2026/04/06/technology/ai-code-overload.html

The Big Bang: A.I. Has Created a Code Overload

Companies are scrambling to deal with the glut.

The New York Times

Georg Weissenbacher 3d ago

Coding agents are generating code much faster than humans can review it:

https://www.nytimes.com/2026/04/06/technology/ai-code-overload.html?unlocked_article_code=1.ZFA.5UxQ.-SKLe76L7Fnh&smid=url-share

#llm #security #codereview #nytimes

1/2

Velox Studio 6d ago

AI automation tip: use an LLM for “second-pass code review,” not final judgment. Have it scan diffs for missing edge cases, vague error handling, and migration risk, then let a human approve. You get broader coverage without turning review into roulette. #AIEngineering #CodeReview

Anirudha 6d ago

What if your PRs could tell you how many reviewers they need?

Made a GitHub Action called PR Reviewer Labeler that does exactly this. It analyzes every pull request and recommends the ideal reviewer count based on lines changed and file complexity.

You can even weigh file types differently. Tests at half weight, core code at 1.5x. A 500-line PR that's mostly tests hits very different than one rewriting your auth middleware.

#OpenSource #GitHub #DevTools #CodeReview #FOSS

Michael Dorner Apr 3

Stop doing code review (in the name of science)

https://www.michaeldorner.de/posts/stop-reviewing-code/

#codereview #softwareengineering #research

Stop Doing Code Review

In software engineering, code review has become a pillar of collaborative software engineering and, for many teams, a non-negotiable step in software engineering. But what if we stopped?

Michael Dorner

All Things Open Apr 2

🚀 NEW on We ❤️ Open Source 🚀

Open source maintainers do not always need more free tools. Sometimes they need funding, flexibility, and space to keep critical projects healthy.

Erik Thorelli explains why CodeRabbit is putting cash behind sustainability, how AI can help with code review without replacing developers, and why exploring tools is still one of the best ways to learn.

https://allthingsopen.org/articles/open-source-maintainers-cash-ai-code-review

#WeLoveOpenSource #OpenSource #AI #CodeReview