I'm a bit of an idiot when it comes to stats. Mainly stick to SPSS. I honestly don't even know if ML is up to what I want it to do at this point. This is the talk I went to (or it roughly the same)
Depends on what you're trying to do with it. I've had some success with F#, specifically using a package called Charon to make random forests. It's fun stuff to play around with.