Writing
Blog
Notes on engineering, design, and what I'm learning while building.
Only when something is finished. Unsubscribe anytime.
Everything, newest first.
Latest · 3 min read
AI Doesn't Remove the Engineer. It Removes the Drag Around the Engineer.
Giving an AI system a Figma design to extract spacing, sizing, and styling is incredibly useful — but generating code isn't engineering software. AI doesn't make engineers less important; used well, it gives them more leverage.
Read the essay →Structure Makes Speed Shippable: Why I Built LLM Orchestrator
AI coding agents build confidently on assumptions. LLM Orchestrator is an open-source Claude Code plugin that makes research, memory, planning, review, parallel execution, and verification part of the system — trading tokens for correctness.
Good, Great, Brilliant: A Field Guide for the Engineer's Path
Eighteen maxims on what separates good, great, and brilliant engineers — principles that hold constant across mechanical, software, and every discipline I've worked in.
The Code That Sleep Forgot: How Developer Burnout Rewrites Our Neural Architecture
When exhaustion becomes the enemy of innovation—how chronic sleep deprivation fundamentally alters the brain's capacity for the creative thinking that defines great programming.
Every Computer You've Ever Used Is Mathematically Suboptimal
Mathematics proves that base‑3 computing is effectively superior, yet we settled for binary 80 years ago, and then we got stuck.
The Invisible Architecture: How Silence Builds the Walls That Code Cannot Scale
Research shows collaborative teams achieve 20-30% higher productivity, yet many engineers let friction fester into silence—here's what happens when two teammates find the courage to bridge the chasm between them.
Machines and Meaning, Part 2: The Illusion of Understanding — When Mimicry Becomes Mastery
The greatest breakthrough in human communication comes from systems that don't understand what they're saying.
Machines and Meaning, Part 1: The Mathematics of Understanding
How artificial minds learn to navigate the landscape of human meaning—and what this reveals about understanding itself.
The Vibe Coding Crisis: Why AI-Generated Code is Creating a Security and Quality Nightmare
How the rush to 'vibe code' with AI is introducing massive technical debt, security vulnerabilities, and maintenance nightmares—and why AI pair programming is the smarter path forward.
From Russian Feuds to AI Minds: The Probabilistic Path to Machine Intelligence
How a cranky mathematician's chain reaction sparked the algorithms that make ChatGPT seem brilliant
The Paradox of Hallucination: How AI's Greatest Flaw Became Its Most Creative Feature
Explore how AI hallucinations have been creatively leveraged to embark on new discoveries, and whether the quest for perfect accuracy might eliminate the accidents that occasionally lead to breakthroughs.
Performance Detective: Master Chrome DevTools
Turn performance mysteries into solved cases with Chrome DevTools' flame charts, live metrics, and detective-level debugging techniques.
Authoring a Vue ESLint Rule: v-if-else-key
I authored vue/v-if-else-key, now merged into eslint-plugin-vue — a rule that catches a subtle Vue 2 virtual-DOM bug where conditionally rendered, repeated components reuse the wrong instance. Here's the problem it solves and how it works.