...Just started using this last night, went through the tutorial and training and more training. Several hours later pretty sure I've got this thing down pat. Just gonna point out a few things about it that I do and do not like.

 1: interaction with MEDIA PLAYERS ,  not a strong point. It would be nice to use this with my iTunes and not have to use the mousegrid to zero in on the artist I want to listen to after saying the artist/band name repeatedly does nothing but show me numbers. From which I must then say ok, and then say enter (mind you this is only 50% of the time). WMP interaction is better but only by a hair. Overall I think it would be nice if you could use the speech recog while listening to music. Seems that playing music and trying to do this results in epic failure everytime because as far as I can tell the recog is attempting to make commands of words/sounds it hears from the music you are playing. Proving difficult for it to understand anything you try to tell it. I feel like if this were possible to work around it should have been done already, but thats just me. I've found that if you start the song you want to listen to and quicky shout "stop listening" it does turn off before the song kicks in all the way and drowns you out.

 2:  Navigation through folders and files, Superb, couldn't be more fluid (once you fully understand how and when to use the right commands). This is where I believe it shines best.

 3:  an additional set of COMMANDS! for things such as programs already running in the taskbar, instead of having to go find them through start menu. Dedicated commands that allow you to restart, shutdown, lock, or put your computer to sleep. Some may argue this but it would be nice, should you have your hands full and are in a hurry. 

 4: Dictation, much like  looking for files is just as easy to do. The correction methods appear to be on the money. I used this with AOL Instant messenger actually, no noticable difference in speed with replies. 

 ...if anyone else has begun or has been using speech recognition on their tx2's feel free to share what you've found, like, dislike about it here. This is a learning thread. let's spread knowledge here!