Saddest part of speech recognition is with all the gigahertz and storage we have, we send it out to the internet to be processed by something else then get the result.