Simply delaying the voice init until the first voice query would also be bad - it would take > 10 seconds for HUD to even start listening to the speaker.
What we really want is to have a separate daemon that starts after the main boot thrashing is over.
Simply delaying the voice init until the first voice query would also be bad - it would take > 10 seconds for HUD to even start listening to the speaker.
What we really want is to have a separate daemon that starts after the main boot thrashing is over.