Humans are not aligned. Neither is life.

Absolute safety is a bad idea. Agent populations run under real start systems. First run: Germany, scaled. The target is relative safety, not a righteous soul.