the higher level of coding we need to do and the more existing systems we need to interact with, the more difficult it would be to make a switch, so even if someone does rewrite a competition for assembler that somehow gains an edge, at this point the edge would have to be so great to convince anyone into making the switch that it's unlikely, and getting less likely by the decade. maybe quantum computing or something like it will do that, but i think its more likely we'll figure out a way to port assembler to it, or some hybrid in which we can make requests from an assembler based command to a quantum processing unit, rather then build a whole new basis for coding. to bring down C you might need a computer science apocalypse.
you are right regarding natural science, but technological sciences are actually a lot closer to traditions then they are to natural sciences: in religion two, it depends on the philosophical innovations of the members within it in order to change, and since they are biased towards their traditions, it gets less likely the older the philosophy gets. it is a growing tradition of how to apply a science more then it is a science, and 1's and 0's have so far successfully remained 1's and 0's.