In addition, it would be prohibitively difficult for those media content services not associated with any VAS (such as I HEART RADIO, PANDORA, TUNEIN, etc.) and those media playback systems not associated with a VAS to develop voice-processing technology that could be even moderately competitive with that of the already-existing VAS(es). This is because NLU processing is computationally intensive, and providers of VAS(es) must maintain and continually develop processing algorithms and deploy an increasing number of resources, such as additional cloud servers, to process and learn from the myriad voice inputs that are received from users all over the world. Specifically with respect to media playback systems, inclusion of a sophisticated VAS would add significant cost, and also cause the system to consume considerably more energy, which of course is undesirable.