Stanford Researchers Release OpenJarvis | Garry Tan Releases gstack

Qodo Outperforms Claude in Code Review Benchmark

Our latest benchmark reveals Qodo surpasses Claude Code Review in AI code review performance. Read more in the blog.

Many SWE-bench-Passing PRs Would Not Be Merged into Main

Git is the new code

AI is writing more code than ever before (if not all of it). But our most important job as developers hasn’t gone away—it’s simply changed. We spend less time typing code and more time reading, reviewing, and making sure everything works as it should. Here are some quick guidelines and git commands to help you

If code is cheap, intent is the currency — /dev/knill

Generating code is cheap now, so the intent behind it matters more than ever. The commit message is where that intent lives.

How We Do Code Reviews at Cloud Four

If your team members dread the notification that they’ve been added as a reviewer on a pull request, I think the following guidelines can help.

How to Kill the Code Review

Human-written code died in 2025. Code reviews will die in 2026.

The Work Moved: What the AI Coding Debate Actually Agrees On

AI agents generate 98% more PRs but reviews take 91% longer. The work didn't disappear — it moved. A synthesis of eight perspectives on where it actually went.

Building GitHuman: An AI-Coded Tool for Reviewing AI Code

Like everyone else, I had to build something over the end-of-year break. I wanted to fix a burning problem I had: how could I review the code my agents...