Depends on what kind of profs.Those who teach low level systems stuff absolutely need you to be able to program in C or something similar. On the theoretical side you can kinda sorta get away with very strong math skills. You open CLRS - it just gives you a pseudocode, you do efficiency/correctness analysis on it and forward you go. Then open an Automata book and realize it's all math all the way down. Not a lot of math either.

Kernel programming still resists AI so C and assembly arent going anywhere.

I think in the era of AI knowing and understanding concepts of the computer science, know how to properly build an application, understand all supporting harness like logging, devops and other stuff become much more important that knowing some implementation details of each language.

Don't get me wrong, it still important, but I think understanding concepts and principles behind each application become more important

You still need to understand how database works and how to analyze performance, how to avoid N+1 problem. You still need to understand how to run your application in the cloud or some bare-bone vm. You still need to understand security concepts