First, yes there faster versions of base like https://mran.microsoft.com depending on what you are doing.
But I would like to respond with a question in return- Because using R is never going to be the fastest way of doing a task (for instance to writing something in pure C/C++ to do something), the question becomes "What tasks are you trying to do that the speed of R is a problem?"
Because depending on the task, it may be something R will never be the fastest at (and so would be better to use other tools) or it may be something where the manner in which you approach the problem is not an R way of doing so in the first place, so the approach is creating slowdowns.
Certainly, when I was at the stage of using datasets large enough to notice, that was when I found that the naïve code I was writing could be sped up by hundreds of times. Not be underlying computer changes, but by thinking about what I needed to do with the data in the course of the analysis and how I was managing that.