The Governance Alignment Problem: Why Politicians Optimize the Wrong Thing

AI safety asks how to align a powerful optimizer with the goals we actually want. This episode argues that the same problem already exists in government: the job of politician selects for election-winning, not outcome-producing. Politicians are not mainly misaligned because they are bad people; they are misaligned because the role rewards popularity, short-term promises, plausible deniability, and re-election. The terminal principal — civilizational flourishing over deep time — has no agent. The fix is not “elect better people,” but architecture: hard constraints, public overrides, mechanism audits, automatic triggers, and a Fourth Branch that asks whether laws and institutions actually produce their stated outcomes.The Governance Alignment ProblemThe Alignment Problem in Your Government