Editorial

The rest of this site teaches, and teaching material has no opinion. This is where the opinion goes. Each piece starts from something that actually happened in artificial intelligence or in software architecture — a release, a result, a failure, a shift in how the work is done — and argues its way to a position: what genuinely changed, what is only being marketed as change, and what a working architect should do differently because of it. Positions are arguable on purpose, and everything they rest on is linked and dated.

8 editorials 5 strands 53 cited sources 12,661 words

All Systems & Scale2 Economics & Value1 Platforms & Power2 Evidence & Evaluation1 Reliability & Consequence2

Benchmarks, claims, measurement, and the distance between what is demonstrated and what is sold.

Evidence & Evaluation 7 September 2026 7 min read New

You cannot bill a histogram

OpenTelemetry's GenAI conventions describe token usage fourteen ways on a span and one way in a metric. The fourteen can be priced and get sampled; the one is kept in full and cannot tell a cached token from a full-price one. Cost attribution for agents is being built on the gap between them.

The only complete token signal in the GenAI conventions cannot tell a cached token from a full-price one, and everything that can price a call rides on spans tracing exists to discard, so AI cost attribution runs on a measurement nobody guarantees is complete.

Read it → 8 sources observabilitycost attributionopentelemetrysampling