Jason Gorman’s article highlights the difficulty of moving from 90% to 99% reliability, noting the journey is as hard as the initial achievement. The author relates this to daily experience with service outages, observing that Git Hub, CI, AI, and Slack are increasingly unreliable. Status pages display numbers like 99.72% or 98.98%, which appear similar to a grade school 'A,' yet the difference between 99.9% and 99.99% represents a tenfold improvement.

Infrastructure teams understand this 'two nines, three nines' shorthand, but the general audience does not. The article argues that uptime metrics are a poor interface for non-technical users. A proposed solution replaces opaque percentages with concrete timeframes, such as '12 hours affected in the last 30 days,' making service reliability accessible to everyone without requiring mathematical interpretation.