How I build with AI
Not a resume. Everything here is real and most of it is running right now.
Running right now
wade-wiki
A personal wiki that maintains itself.
I say something on a walk. By the time I'm home it's filed, linked to five related ideas, and committed to git. Fixes the graveyard problem: ideas you capture and never revisit.
Burpee Counter
Voice counting designed around the AI's failure mode.
Speech recognition mishears a man gasping at rep 70. So spoken numbers set the count instead of incrementing it, and a miss heals itself. Designing around the model's weakness is most of the job.
AI Usage Meter
A menubar app that tracks how much AI I have left.
Quota is a household utility now. It sits in the menubar so I stop finding out at 11pm.
The bench — tools in rotation, with verdicts
Opinions I've earned
- Design around the failure mode, not the demo.
- Raw stays raw.
- Files are the treasure. Tools are rentals.
- Autonomy is a dial, not a switch.
- $0 marginal cost is a design constraint that makes you smarter.
- Specificity beats generality.
Didn't survive contact
LifeOS
b. January 2026 † superseded May 2026The processing worked; the ritual didn't. Lesson: start from capture, not reports.
The tutorial-following era
a kanban board and a project-management app that shall live in private reposFollowed other people's prompts. Learned the tools, shipped nothing. Lesson: only your own problems teach the craft.
An AI note-taking app
b. and † 2026, quicklyNotes apps are where notes go to die. Lesson: capture was never the hard part.
Poke it. It's all live and on GitHub. Get in touch for the long version. And yes, this page was drafted by Claude Code inside that wiki. Of course it was.