AI agents removed the friction from writing telemetry
I used to avoid adding telemetry because it felt like tedious busywork. Now with Claude handling the OpenTelemetry boilerplate, I'm instrumenting everything.
Happy Friday!
Over the years I've read a lot of great essays about design systems, and want to share them with you.
This week's long-read is Measuring the Health of a Design System by Chase McCloy. Chase defines what it means to have a healthy design system and shares examples of how his team tracks their system's health.
Feel free to send me job descriptions you come across as well (you can DM me on Twitter @MCapoz or email me at mae@maecapozzi.com).
Design
Engineering
Talk soon,
Mae
I used to avoid adding telemetry because it felt like tedious busywork. Now with Claude handling the OpenTelemetry boilerplate, I'm instrumenting everything.
AI agents work better when given appropriate context and guardrails.
I've been using Claude Code to offload tedious parts of platform engineering: dependency reviews, generating test PRs, dependency migrations, and project cleanup.