And it's been around forever. If you're interested in high-speed processing, there's lots of good papers from that project (e.g. how to avoid creating branches, stuff like that).
I agree, I'm excited. Hopefully the optimizer can take full advantage of these column stores.
However, postgresql isn't the only open source column store. MonetDB is another one. It is an academic project but people do use it commercially. I know some people who get great performance out of MonetDB, even though its optimizer has some gaps.
Basically ParAccel (like used by Amazon RedShift) at reasonable cost.