Writing
Notes from the work
18 years of decisions, research, and design. These are the patterns that keep recurring.
September 2026 · 9 min read
How I Designed a Trade Show Game for Logitech G That Players Line Up For, in Five Working Days
A gamified click speed kiosk for Logitech G, designed and shipped in five working days, inside a fifteen day window, for Kamp Grizzly. The client killed the first direction in one sentence, a playtest bug became the celebration screen, and the whole thing ships as one HTML file.
Read the postFurther reading
April 2026 · 12 min
I Used Claude Code to Build My Entire UX Portfolio. Here Are the Actual Numbers.
47 git commits. 10 build sessions. ~100K lines of code. One designer, no developer. Here is every number, every failure, and the exact workflow behind displayedux.com.
Read12 min readAugust 2026 · 8 min
Claude Design Has Four Input Channels. Most Designers Are Using the Slowest One.
Chat, inline comments, direct editing, and sliders. Claude Design ships four ways to change a design, and the one the cursor starts in is the slowest. Here is the channel rule I use.
Read8 min readApril 2026 · 8 min
Mental Models, User Data, and the Quiet Goal of Less Cognitive Load
The mental model is the hypothesis. The data is the proof. Eighteen years of A/B and multivariate testing have taught me what to do when the two agree, and what to do when they don't.
Read8 min readJanuary 2026 · 9 min
The Onboarding Problem I Keep Solving at Every Company I Have Worked At
Netflix. Appily.com. TeleSign. Three different products, three different user groups, the same root cause every time.
Read9 min readFebruary 2026 · 10 min
Designing for Experts Who Cannot Afford to Be Wrong
Fraud analysts make hundreds of high-stakes decisions per day. Here is what designing for that user taught me about complexity, trust, and speed.
Read10 min readDecember 2025 · 10 min
18 Years In, I Still Think Most Designers Confuse Output with Outcome
Early in my career I measured my effectiveness by what I shipped. It took years to understand why that was the wrong metric.
Read10 min read
