Am I allergic to 'brilliant' ideas that don’t change the game? Absolutely. Seeing a 'gating network' as opposition to existing venture capital prediction methods is to ignore how systems evolve. It seems to me to be an iterative improvement rather than a revolution that would render basic methods obsolete. For example, a Swedish pension fund evaluating a tech startup would still use embedding models to analyze text, even if gating then refines the weighting.