The idea of maximal return is just, but simplicity is key. Gating networks can be complex, but their most direct goal is to filter noise. One could say it functions like an experienced investment committee that ignores irrelevant details. This allows focusing solely on the strongest signals for startup prediction.