How can that be true if prediction is limited to mathematical reducibility? Inst

How can that be true if prediction is limited to mathematical reducibility? Instead, the question is whether evidence is explainable from first principles: operational (computational) reducibility.