Jan Schaumann interviewed at Velocity 2012
summary
Argues that adding layers to chase performance increases accidental complexity and makes failures harder to diagnose, cites Sturgeon’s “90% of everything is crud” as context, and pushes for simple base tools, open source with a programmable API and active community so edge cases can be scripted. Recommends measuring total cost including ramp‑up, human and integration cost rather than just server count, and often prefers throwing money at capacity to minimize toil and optimize human time.
AI-written summary · about
