I don't know about tel, but I use both pretty frequently (and usually together with rpy). R is great for some things, but I typically find that I need to use a general purpose language to acquire the data I'm processing. There are some processing steps that I easier in R though, so I offload things to that when it is more efficient. But for simpler things, I find that numpy is more than adequate for processing data that I've already got in a Python session.
I use them for bioinformatics analysis (large scale genomic comparisons and microarray data analysis). So the larger problem isn't the calculations, it's the data management. And for that Python wins over a pure R solution.
I personally love R. More specifically I love how everything is implemented (at least sketchily) in it. Most specifically, I love how its plots look. If it could handle the large data sets I'm working with these days (MRI 3D+time images, sometimes multiple contrasts at once) I would use it a lot more.
So yeah, RPy is another big win for the Python side of things.
Python has NumPy/SciPy/and friends, so I use Python.
This, I feel, is a strong point for most who are using a scripting language to do scientific work.