Notes from
the field.
AI, interface design, human judgment, and the systems that help people do better work.
Design AI for Correction, Not Just Confirmation
A human review step only helps when people can challenge the evidence, change the recommendation, and stop the work before it causes harm.
What You Should Never Delegate to AI
AI can prepare consequential decisions without owning them. How to retain responsibility for goals, tradeoffs, commitments, and the people affected.
The Interface Is the Accountability Layer
AI interfaces should make evidence, authority, action, and recovery visible so people can take responsibility for the work they approve.
AI Makes Production Cheap. Judgment Becomes the Bottleneck.
When AI makes drafts and options easier to produce, the useful constraint becomes our capacity to evaluate them and take responsibility for the result.
Stop Looking for the Perfect Prompt. Build a Better System.
Better AI work starts with clear purpose, bounded authority, usable evidence, and a review process, not a prompt that tries to cover everything.
Maintain AGENTS.md Without Building a Policy Landfill
A practical method for adding, revising, and deleting agent instructions without turning one file into permanent context debt.
Layering Context for AI Coding Agents
How to separate global preferences, repository architecture, package rules, and task intent without creating conflicting instructions.
Turn Validation into Executable Agent Instructions
How to replace vague definitions of done with focused checks that AI coding agents can run, interpret, and report.
Setting Autonomy Boundaries for AI Coding Agents
How to give coding agents room to work without silently delegating consequential product and engineering decisions.
Reference Codes for Better AI Decisions
A small notation system for keeping options, risks, questions, and decisions stable across long conversations with AI agents.
AGENTS.md Is a Communication Protocol
How I use AGENTS.md to define clear communication, bounded autonomy, and repeatable engineering behavior for AI coding agents.
Making AI Research Source-Conscious
How I built a multi-agent research system that separates sourced facts from hypotheses, preserves the work behind a decision, and makes uncertainty visible.
Building a Multi-Agent Code Review Board
How I turned code review into a routed, source-conscious system with independent reviewers, bounded execution, and one useful report.
Building Breadcrumbs with Remix useMatches Hook
The useMatches hook is a very powerful tool that allows you to access child route data. We'll use this to build a breadcrumb navigation.
Javascript Array Methods
Simple showcase of the most used array methods in Javascript.