Have we reached the capability of a local STT+LLM system being constantly listening for normal speech in a room and being able to understand when the human is a

Have we reached the capability of a local STT+LLM system being constantly listening for normal speech in a room and being able to understand when the human is addressing the system instead of talking to another human?