How would dynamic block size economics differ from the economics of dynamic difficulty adjustment?