Thilak wrote an article a little earlier today about a company called Podzinger. Not only can you search text, but audio as well! Whether it is in a podcast or a video, speech is indexed just like text is.
Speech to text is something I am a bit on the fence about. There is the camp that says speech recognition is the way of the future. All mobile devices will have it and it is the one thing that stands in the way of truly being “mobile.” After all, who wants to sit in their car and bang away at little keys when you can just say it?
I am tending to lean on the other side though. I think it will be imperative for ad networks to ‘get it.’ We are quickly approaching a point where so much content is tied up in formats other than text; ie. podcasts and video. In order to remain competitive from an ad serving standpoint, there needs to be a way to break down speech. But I don’t buy the fact that all data queries and input will be speech based.
We have the tools necessary to translate language into something understood by computers, but will it happen? Will we be able to talk to the next wave of mainstream applications? My magic ball isn’t showing anything, which for me is a bad sign :0)