I know they do auto-generate some tree information which can be incorrect, but that can largely be ignored or used as hints, etc.
I get the sense that not many people realize just exactly how much stuff they have actually accumulated. It's truly amazing for the areas that I've been looking at which are primarily US, Canada, Britain and Ireland.
I tend to think of it as one of the pillars of online data collections along with things like, archive.org and, say, Google Books. It's incredible the amount of material that is at our fingertips now.
The one biggest missing pieces for me is the staggering number of microfilmed historical newspapers that have yet to be digitized. I know the library of Congress has put a lot on line but there are still 10s of thousands of rolls of microfilm that haven't been processed and I'm not sure if there is any push to do it.