The simulation manifesto: The limits of brute‐force empiricism in geopolitical forecasting