Felt exaggerated at the time, but it often seems like the truth.
GFS only translated into HDFS after 5 years, same with MapReduce, BigTable etc.
That said, I suspect a large part of the reason is that Google's MapReduce is highly optimized for running in a Google datacenter, and they're generally pretty tight-lipped about the real secret sauce in their infrastructure.
My understanding is that sharing details takes time off very valuable employees, away from fixing important internal issues – a lot more than the company being afraid those details might leak.
For the handful of companies who can and needs to implement equivalent, poaching said employees is simpler than back-engineering from blogposts and it makes implementation easier.
However, Facebook and Amazon almost certainly have similar challenges. I don't think they are necessarily unique so much as so huge the rest of us don't have the money or the problems to solve.