Most technical decisions optimize for the next sprint. But some systems run for decades. What changes when you design for thirty years instead of three months? A framework for thinking about longevity, and why it matters more than ever.
The hardest problems in distributed systems aren't about code. They're about consensus, truth in adversarial environments, and reasoning under uncertainty. It turns out epistemology is more practical than anyone in the CS department wants to admit.
Every abstraction you add is a tax on everyone who comes after you. Most systems fail not from lack of capability but from accumulated complexity that nobody can reason about anymore. On the discipline of aggressive simplification.
Most best practices are context-free solutions to context-dependent problems. The real skill isn't knowing the patterns—it's knowing when they don't apply. On developing judgment instead of collecting rules.
Some code I wrote in 2003 is still running. Other code I wrote last year is already gone. What separated them? Reflections on building things that survive the people who made them.