Personally, I don't see Julia replacing R for a while. As you note, the R community has much of what it needs already. When I initially switched from R to Julia, it was because R wasn't really usable for the kind of work I do. It's since become clear to me that I work on problems that don't come up for most R users: problems involving large sparse matrices, including large-scale optimization and MCMC. If you don't deal with those kinds of things, R is a good choice.
Granted I don't have much experience with Stan, but from what little I've poked around RStan, the workflow left something to be desired. Setting up a C++ environment, and then embedding Stan code (which appears to be a C++ DSL of sorts) as strings inside your R code seems... unpleasant.
Perhaps if MCMC.jl matures, and if it can offer performance competitive with BUGS, JAGS, Stan, etc., I could see Julia's statistical fortunes rise along with Bayesian methods generally. I'm working on a PhD in a discipline that is just now beginning to dip its toes in Bayesian waters. I get the feeling that the adoption surge is yet to come.
Interesting that you're implementing your own MCMC methods mainly. As part of my coursework I did a little bit of that, but perpetually felt like I wasn't smart enough to anything sophisticated--either in terms of exotic sampling methods, or working with very complex posterior distributions. It may just be my frequentist R toolbox experience messing with me, but I know I feel more comfortable with the safety blanket of a framework.
Second that. The purpose of Julia is not to replace every other language that people use for technical computing. I personally use Python for other projects (www.circuitscape.org), and only recommend using Julia for new classes of problems that need Julia's performance and capabilities. That class of problems keeps expanding over time.
