Entries for September 2025

Google's Code Review Guidelines (GitHub Adaptation)

/2025/09/29

This is an adaptation of the original Google’s Code Review Guidelines, to use GitHub specific terminology. Google has their own internal tools for version control (Piper) and code review (Critique). They have their own terminology, like “Change List” (CL) instead of “Pull Request” (PR) which most developers are more familiar with. The changes are minimal and the content is kept as close to the original as possible. The hope is to make this gem accessible to a wider audience.

I also combined the whole set of documents into a single file, to make it easier to consume. You can find my fork here. If you notice any mistakes, please feel free to submit a PR to the fork.

@onusoz · 2025-09-13

/2025/09/13

Just downgraded my anthropic sub, 200 usd openai plan is finally justified after 1 year

Enjoying gpt5 codex free lunch before openai inevitably starts cutting corners just like anthropic (I hope to be wrong in 3 months, this model is very good at one shotting things and I don’t want it to be nerfed)

@onusoz · 2025-09-08

/2025/09/08

@thsottiaux 2/ Model just stops working on a task even though I tell it to run something and not stop until it works. I have to frequently say “ok do it then”. Probably a model problem and not harness problem

CLAUDE.md to AGENTS.md Migration Guide

/2025/09/08

This post will age like sour milk, because Anthropic will eventually adopt the company-agnostic AGENTS.md standard.

@onusoz · 2025-09-05

/2025/09/05

So let me get this straight, the main reason for Responses API exists is that OpenAI doesn’t want to show reasoning traces? Therefore the whole world should bend backwards to fit your obscurantist standards? Responses will not get adopted the same reason Windows Server didn’t