Thanks for infusing realism into the Erlang hype. As you point out "There is no free lunch". People also frequently tend to forget Amdahl's law when discussing Erlang and somehow assume that "Erlang = Linear Scaling" regardless of algorithm and problem.