I think you’re right about people, but by volume, 99% of all future code will be written by machines. So there’s ample opportunity for other languages to flourish; if humans can’t get past imperative programming, machines can.
In my mind this hinges on whether LLMs become capable of actual comprehension of the code they’re writing. If not, for serious projects (especially those which are mission-critical), it still makes sense to optimize for human understanding when selecting languages so the code can reasonably be verified.