This simulation puts a single, all-powerful AI in charge of an entire planet's population, resource stock and climate stability, then lets you decide what it optimizes for. Set the AI's objective weights across growth, conservation and stabilization and watch it reallocate its intervention effort in real time on a live 3D globe, where population is rendered as a swarm of instanced points that thrives green or reddens under stress. Because the AI's own growth lever quietly erodes the resources and stability its carrying capacity depends on, pushing any single objective too far produces the same cautionary pattern real optimization systems show: short-term gains followed by a collapse in the very metric being chased — a concrete, interactive illustration of Goodhart's law at planetary scale.