The idea that a designed machine could be competitive with living things at self-replication (the sole target of a multi-millenia optimisation process) seems… pretty unlikely to me.