Magic Wand:
One Hour. One Question. One Unexpected Capability.
When a real designer has sixty minutes to build something at a maker event, the first move isn't to code. It's to ask: what's already in my hands that nobody else is thinking about?
The Setup
Code with Claude had a Makers Station. One hour. M5Stack Cardputer-Adv in the corner — a palm-sized device with QWERTY keyboard, accelerometer, Bluetooth, speaker, color screen. Hundreds of people watching what people would ship in 60 minutes.
Most makers would see: "What can I build on this device?"
I asked a different question: "What can I build between this device and everything else?"
The Thinking
Three months earlier, I'd built an integration with the Oura Ring MCP — exploring what happens when you connect biometric data to Claude as a context layer. That project taught me something: the magic isn't in the individual tools. It's in the unexpected combinations.
The Cardputer has accelerometer data. It has Bluetooth. It can send commands. My Mac can receive commands. My Mac controls system settings.
Most people see that and think: "Build a game." "Build a remote control." "Build a calculator."
I thought: What if the gesture itself is the interface? What if instead of pressing buttons, you shake the device and something unexpected happens on your Mac? The moment I asked that, the entire design solved itself in reverse. Because the answer wasn't "what can the Cardputer do?" — it was "what feels like magic to a human watching it happen?"
Here's the constraint that matters: I had one hour. That meant I couldn't build the obvious path. No elaborate UI. No complex interaction design. No time for polish. I had to find the elegant intersection between:
What the hardware could do (shake detection via accelerometer)
What the Mac could do (toggle system appearance via AppleScript)
What would make someone's brain break in a good way (gesture = spell)
The well-travelled path would be: build another app, another dashboard, another control center.
The unexpected path: Make the device a wand.
What That Means in Practice
The UX design was one gesture. You hold the Cardputer. You shake it. Your Mac's dark mode toggles to light mode. The gesture and the outcome are so directly linked that your brain doesn't process "Bluetooth protocol" or "JSON packet" — it processes magic.
When people tried it:
They immediately pointed the device at the screen (like aiming a wand at a spell target)
They didn't read instructions
They laughed when it worked
They tested the limits (how far does the range go? can I break it?)
They wanted to try again
That's not luck. That's the design working. Because the interface dissolved. The gesture became the entire product.
The Technical Feasibility Check
Once the concept was locked, the technical path became clear:
M5Stack Cardputer runs custom MicroPython firmware (not the default UIFlow)
Accelerometer detects the shake (threshold + hit-count to filter false positives)
BLE Nordic UART Service sends a simple JSON command over Bluetooth Low Energy
Mac-side listener (Python + bleak library) receives the command
AppleScript toggles macOS appearance
Sync sound effects on both devices for tactile confirmation
None of this required inventing new technology. All of it was combining existing capabilities in a way nobody else had bothered to try.
Why This Matters for Design
When you have unlimited time, you can afford to be obvious. You can build what everyone expects. You can follow the playbook.
When you have one hour, you have to be dangerous. You have to know your tools well enough to see combinations that seem impossible. You have to trust that simplicity beats complexity, and that a single elegant gesture beats a dashboard full of options.
The real design work wasn't in the code. It was in the question: "What haven't I seen before?"
And the answer was: a device that feels like a wand.
The Event Context
Built at Code with Claude (San Francisco, May 6, 2026) — Anthropic's developer event with hands-on workshops and live demos.
The Makers Station was specifically designed to showcase what Claude Code could enable. Not just software development, but hardware provisioning in one command. No manual setup. No environment configuration. Plug in the device, type m5-onboard go, and everything happens automatically: firmware flash, file transfer, device reboot into custom launcher.
That's Claude Code as a design tool, not just a coding tool.
What People Built After Seeing It
The moment Magic Wand worked, other makers at the station got curious about what else was possible on the Cardputer:
Someone built a game
Someone built a piano simulator (with sound)
Someone built tools
All of them started from the same question I asked: "What's already in my hands that's unexpected?"
That's the design philosophy that matters.
The Takeaway
You don't need more time. You need better questions.
Instead of: "What can I build?"
Ask: "What combination of existing capabilities would surprise someone?"
Instead of: "What's the obvious solution?"
Ask: "What gesture or interaction would feel like magic?"
Instead of: "What features should this have?"
Ask: "What if the entire interface was one unexpected thing?"
One hour. One question. One device that felt like casting a spell.
That's the design move.
Built: Code with Claude, San Francisco (May 6, 2026)
Stack: M5Stack Cardputer-Adv · MicroPython · BLE Nordic UART · Python bleak · AppleScript
Official Event: Code with Claude
Media: LinkedIn coverage