Agent Harness Lab Retrospective
An experiment on SWE-bench Lite comparing code repair by the same model using One-shot and an Agent Loop, with analysis of repeated rollouts, Pass@k, step limi…
Agent Harness Lab Follow-up Experiment Retrospective
A comparison of one-off generation, tool-free multi-turn revision, and an Agent that can search, edit, and run code on 20 real code-repair tasks from open-sour…
I Built a Video-to-Handout Tool with Codex. It Quietly Left Out Half a Lecture.
A look back at an AI Coding project and what I learned from it.
I Just Wanted a Clean IP for ChatGPT. Somehow I Ended Up Building a Proxy Chain.
From a Dedicated ISP Proxy and HTTP CONNECT failures to SOCKS5 and proxy chaining in Clash / Shadowrocket, a troubleshooting and multi-device setup retrospecti…
Automating My GitHub Profile: From Manual Commits to a README Generator
A record of automating my GitHub Profile: from manually recreating historical commits and troubleshooting Git and network issues to using Python scripts to sta…