The only way to move embedded development into safer languages is with real products developed in such languages, otherwise the hardcore C guys will never be convinced.
It cannot be generalized for all types of embedded hardware projects for sure, but the use of C, C++ and Assembly should be reduced to the bare minimum type of projects where no other technology is possible.
Are you bumping into some significant performance problems, for instance?
For Mongo: Most people use mgo http://labix.org/mgo
For precog: dont know, but there might be something, but probably not.
I just forgot to say: a FRP*  Go equivalent of Play and that there's nothing like these upcoming Coursera courses   for Go with examples.
Perhaps a better way for me to have put my sentiment would have been, "I'm not tired of this, so I don't mind seeing it", where "this" means both phdtree specifically and hearing about project being written in, moving to, or considering moving to Go in general. Clearly you are tired of it, either from having seen other specific phdtree postings or from having seen more postings about Go in general than you care to. That's fine, but it's not where I'm at.