For years, teams dragging off vendor SDKs onto Open Telemetry hear complaints that progress seems stalled. Vendor tools are idiot-proof, while OTel greets users with "experimental" stamps and multiple implementation paths. The project maintains true vendor agnosticism, yet conversations in the semantic-conventions repo drag on. Language stories differ dramatically; Golang and Dotnet lead while others lag.

Auto-instrumentation is magical, but the cliff to manual instrumentation is steep. The perceived slowdown isn't just lack of maintainers—it's a three-way crash. A strict binary stability gate, combined with few maintainers and massive scope, creates worry. Once shipped as stable, features can never change, incentivizing endless debate.

Work splits into two buckets: Core stays small, stable, and tightly reviewed, while Contrib covers the long tail. The otel-collector mirrors this approach for logs, metrics, and traces. This narrative has persisted in the observability space, raising questions about whether slow progress is imaginary or structural.