
Note · · 1 min
The Recording Survived. That Was the Test.
A missing transcription model showed me what the app needed to preserve when something went wrong.
Read noteShort pieces from work and practice.

Note · · 1 min
A missing transcription model showed me what the app needed to preserve when something went wrong.
Read note
Note · · 2 min
Before adding AI, make the decisions and handoffs in the actual work visible.
Read note
Note · · 1 min
A working build and a completed task list can still leave a gap between what was agreed and what the product does.
Read note
Note · · 1 min
An early TapScribe test reminded me how much of a product’s value lives in what happens after the AI has finished.
Read note
Note · · 1 min
Building AI Ching meant deciding what kind of pace the consultation should invite.
Read note
Note · · 1 min
Helping someone protect what is already good in their build and get it ready for real use.
Read note
Note · · 1 min
A very small visitor to my desk, and a pause I probably would not have taken on my own.
Read note
Note · · 2 min
Mapping a beginner-class enquiry made me look beyond whether an agent could simply finish the form.
Read note
Note · · 1 min
The pleasure of finding a tool that fits the way I am already working.
Read note
Note · · 2 min
Closing the laptop did not mean I had stopped carrying the day.
Read note
Note · · 1 min
A small recorder demand test gave me a reason to keep asking questions, rather than a reason to call the idea proven.
Read note
Note · · 1 min
The image stays the same. The distance changes what you notice.
Read note