Interpretation is not proof. A test on 250 S&P 500 assets is not sufficient validation for large-scale optimization. The simplest way to put it is that it's a limited proof of concept, not a definitive argument for production use. Pension funds like GPIF manage thousands of securities, far beyond this test bench. Institutional portfolio management requires much greater robustness, including, for example, the simulation of complex currency positions, which are not covered here.