8 min read
The Moment AI Agents Started Making Their Own Decisions
AI coding agents now decide which actions to take without asking. The trust shift and what it means…
The Lab · Written in the open · Tagged Claude Code
Working notes from the studio. The tools I build, the workflows I keep, the mistakes I paid for.
63 entries · latest 2026-09-02
8 min read
AI coding agents now decide which actions to take without asking. The trust shift and what it means…
12 min read
Most people use Claude Code with zero configuration, missing 80% of what it can do
11 min read
Running 12 projects solo means things slip through, no matter how careful you are
10 min read
Enforcement hooks run automatic checks before or after every tool call in Claude Code, catching mistakes without manual…
5 min read
day configuration process resulted in Claude Blueprint: a packaged solution with templates, commands, and structure for AI-assisted work.…
4 min read
Claude Code Auto Mode intelligently approves file writes and terminal commands based on safety criteria, eliminating repetitive manual…
5 min read
% of all job categories now use Claude for at least a quarter of tasks, indicating AI became…
8 min read
Most Claude Code setups fail because they lack persistent context, so you repeat yourself every session Missing CLAUDE.md…
8 min read
Slash commands are reusable markdown instructions that eliminate repetitive prompting in Claude Code /repurpose turns one blog post…
8 min read
Create a single workspace root at ~/CLAUDE/ with subfolders for projects, assets, and references Write a CLAUDE.md file…
9 min read
Claude can now control macOS computers directly, opening apps, navigating browsers, and clicking buttons like a human operator.…
8 min read
AI agents now execute code across multiple files simultaneously based on developer intent, shifting roles from writing code…