VCs keep saying AI will disrupt everything except venture itself. ADIN was featured in @WIRED last week in an article that poses this question: What parts of VC are actually defensible?

The clearest shift: analysis costs collapsing to near zero. What once took weeks of junior labor now takes minutes of compute. Pattern recognition, regulatory edge cases, market sizing are probabilistic problems machines handle well.

The pushback is predictable: "VC is art." "Taste matters." "Networks can't be automated." All true. But they're last-mile advantages.