The cocktail napkin is older than design.

It’s where the idea is allowed to be wrong — before the deck, the Figma file, the Jira ticket.

Mine used to be a Moleskine. Then a whiteboard. Lately it’s been a Claude chat session open in Plan Mode.

See the work or Read the writing

  1. 01

    Designs should be simple. Interactions within design should be ridiculously simple. A decade of design systems taught me that complexity is almost always a sign that someone upstream didn’t make a decision.

  2. 02

    Figma is the right place to start. It’s almost never the long-term source of truth. The moment you’re building a real design system, the npm package starts pulling ahead — and knowing when to go back to Figma to reconcile, versus when to keep building forward, is the judgment most teams never develop.

  3. 03

    AI in design is a knowledge problem before it’s a prompt problem. Most teams skip straight to prompting, then wonder why the output sounds like no one who works there.

  4. 04

    The napkin sketch is the spec. Figma is a translation, code is a translation, and the meeting where someone asks “but what did you mean by this?” is what happens when the intent didn’t survive the trip.

  5. 05

    Simple, yet strong. Knots and components work the same way — they hold until the requirements change, and then you see what you actually built.

Jeremy’s Principles of Design A line drawing of a clove hitch tied around a vertical post.

“The clove hitch is an incredibly simple knot. Once under tension, it holds until the rope breaks.”

A room next to this one

The Workshop.

Bear Paw is a control system for Uniden scanners. Kerchunk is a Pi Zero 2W and a stick SDR I’m trying to get into something pocket-sized. I’ve been running a Meshtastic mesh network across Kansas City. I spent three months trying to break my Apple Music algorithm on purpose.

The Workshop is where I keep all of this. Hiring managers can skip it. Designers who keep reading will probably stay.

Enter the Workshop →