You don’t really show that trusted is better than trustless — you merely show that trustless is mostly impossible for any real world problem.