Comment 1 for bug 1170784

Ted Gould (ted) wrote :

The only words we know how to detect are the actions, so we'll always guess one of those. Otherwise we're doing generic voice recognition which is "hard."