Mistakes I see engineers making in code reviews

Mistakes I see engineers making in their code reviews
11 minutes by Sean Goedecke

Code review has become more critical as AI tools now generate code easily, but reviewing remains just as challenging. Engineers often spend more time reviewing AI-generated code than their colleagues' work. The most effective code reviews look beyond just the code changes to understand how new code fits into the entire system. Sean argues that good reviewers limit comments to five or six meaningful ones rather than overwhelming authors with dozens of minor suggestions.

Stop Bots and Abuse in Real Time with WorkOS Radar
sponsored by WorkOS

Bots exploit trial signups and drain compute. WorkOS Radar uses device fingerprinting, traffic analysis, and behavioral signals to detect and block abuse in real time while integrating cleanly with your existing auth flow through a simple API.

Code like a surgeon
5 minutes by Geoffrey Litt

AI coding tools work best when developers think like surgeons rather than managers. Surgeons do the core work while support teams handle prep and secondary tasks. Geoffrey uses AI to handle background tasks like writing documentation and fixing bugs, while focusing personally on creative design work. This approach works because AI can run secondary tasks overnight without career concerns or status issues that affect human team members.

Designing software for things that rot
9 minutes by Vadim Drobinin

Vadim turned fermentation hobby into a data science problem. After nearly poisoning himself with questionable salami mold, he built an app called Fermento that applies food industry safety standards to home fermentation. The app tracks fermentation phases and automatically generates HACCP compliance documents, making it safer to cure meats and handle risky ferments at home.

How we saved $500,000 per year by rolling our own “S3”
13 minutes by Miedwar Meshbesher

Miedwar shares how Nanit built N3, a custom in-memory landing zone, to replace S3 for their video processing pipeline. S3's per-object fees and 24-hour minimum storage costs were expensive for videos that only needed storage for seconds. N3 handles normal uploads in memory while using S3 as a fallback for overflows or failures. This change saved about $500,000 per year in costs.

The Linux boot process: From power button to kernel
13 minutes by 0xkato

You press the power button. A second later a wall of text scrolls by, or a logo fades in, and eventually Linux appears. What happens in between is not magic. It is a careful handshake between tiny programs and a very literal CPU. In this article 0xkato follows that handshake until the very first line of C code inside the Linux kernel runs.

And the most popular article from the last issue was:

• What actually happens when you press send to ChatGPT