Articles from substack.com

5 kept

browse.sh
browse.sh is a browser CLI that gives AI agents pre-built skills to automate websites, cutting token costs and enabling complex multi-site workflows.
How Claude Code works in large codebases: Best practices and where to start
Successful Claude Code deployments at enterprise scale rely on a layered harness of CLAUDE.md files, hooks, skills, plugins, LSP, and MCP servers rather than model alone.
https://substack.com/redirect/a04808d0-f7f3-4c7f-8c90-4cf78228bebf?j=eyJ1IjoiNGRza3BlIn0.cWLfcLhrE0spw67UwQFN-a0emx7WI0vcpkS5qD2nMKs
https://substack.com/redirect/4e54edc6-caeb-4d4f-97f4-f89bef50b15c?j=eyJ1IjoiNGRza3BlIn0.cWLfcLhrE0spw67UwQFN-a0emx7WI0vcpkS5qD2nMKs
LLMs as Judges: Measuring Bias, Hinting Effects, and Tier Preferences
LLMs used as judges to evaluate other LLMs exhibit measurable self-preference bias, with GPT showing the strongest self-bias and Claude the weakest, while revealing model identities through hinting changes judge behavior unevenly across vendors and domains.