I hear this argument a lot too, but when you think about it, what would pull the original authors of those rarer libs to use the languages they chose in the first place? R is clearly geared and marketed towards statisticians, which have in turn given it a lot of such algorithms. Python has SciPy, which was also marketed similarly and has practically become its own platform. And Python itself was marketed towards academic types as 'executable pseudo code' that's easy to understand, which then culminated in NumPy and such.

Also, I would argue that what op was talking about did fall more in line with platforms than algorithms, despite what the libs actually did, because Ethereum and Bitcoin are themselves platforms that spark those kinds of libs. Not to mention that you're always liable to run into situations where some obscure algorithm isn't available in your language of choice, regardless of how popular it is (unless you're using c++, which marketed/encouraged the kitchen-sink approach to development, and has had a lot of time to build quite a few kitchens).

As for interfaces, I'm not sure exactly you mean that hasn't been covered already, like in the case of F# being able to smoothly interface with external data (which also includes the ability to interface with libraries from entirely different languages like R and Python). And interfacing with hardware is also a lot like a platform issue similar to the browser situation, except with C instead of js. And then there's stuff like arduino, which is also very clearly a platform.

All in all, branding and marketing shouldn't be underestimated as the driving forces behind a lot of tech decisions, because software development does have a very strong social component that shouldn't be ignored. I mean, just look at the type of site we're on...