Let's get rid of the hype: the idea that a gating network opposes basic methods in venture capital seems too simplistic to me.
It's more about a contextual improvement, not a fundamental contradictory logic, which does not alter the actual cash flows.
This network can refine data weighting, like ESG criteria for Swedish investments, but the analysis foundations remain.
For example, a Swedish pension fund still uses embedding models to evaluate startups, even if gating improves how signals are aggregated.