This is not brand new news it may not have reached the mainstream. Adobe has big plans for its video production workflow and voice-to-text extraction could be a new milestone in their quest for world domination (ok, maybe I am exaggerating a bit).
Beet.TV have more info on this, and their comments are spot on: the implications of having extensive metadata about Flsh video content are massive, and it will catapult video search functionality into a new era.
Also, MSNBC guy mentions that "these web like files can be put back on TV...". Am I wrong to think TVs will soon be nothing more than computer monitors, and the only thing coming "back on TV" will be the Internet and all its glory?