Every time there's a voice recognition post here someone comments about acquiring data. Why is this method better than having access to all of the video and podcasts sites on the internet?