Life comes at you fast

GitHub's CTO recently published a post on availability issues driven by AI-fueled record growth in pull requests, commits, and repos. The author analyzes this through David Woods's 'Messy 9' patterns for complex systems, highlighting how increased tempo leads to saturation risk. GitHub can't simply scale horizontally and must rearchitect systems, but firefighting outages steals engineering cycles needed for long-term fixes — a dangerous feedback loop. Separately, Anthropic's Claude suffered another outage, dropping to one nine of availability over 60 days, also attributed to unexpectedly rapid demand growth. The author argues LLMs are unlikely to improve reliability because they are themselves a primary driver of the saturation problem.