Each card below opens into a longer write-up: what I worked on, why it mattered, what I learned, and links to the upstream pull request or merge request. New entries appear as I ship them.
Why I'm beginning to contribute to OpenTelemetry, what I plan to work on first, and a longer reflection on why the second-order cybernetic learning curve is the one that matters most for engineers.
OpenTelemetry · Friction logA detailed friction log documenting where the current OpenTelemetry GenAI semantic conventions fit poorly when instrumenting a deterministic, content-blind workflow enforcement layer.
An essay on a growing class of AI infrastructure that participates in execution without generating, and why current observability vocabularies do not see it.
An essay on how technical standards are shaped not only by the systems they describe but by the commercial and institutional position of the people doing the describing.
An analysis of the .method provenance pattern from the
agentic-authorization semantic conventions — why a producer-emitted
score needs a way to signal when the measurement itself changed, not
just the agent being measured.
A reflection on how an open issue thread quietly turned a friction log into a shared proposal — the story of how a handful of strangers, arriving from different directions, built something none of them would have made alone.