Should coding interviews just become vibe coding interviews at this point?

I don’t really get why interviews are still so focused on obscure data structures, algorithms, or complex SQL and pandas problems. At this point, most of us are using AI in some capacity to write or assist with code anyway.

Why does it still matter if I can invert a binary tree in 10 minutes from memory? Wouldn’t it make more sense to talk about actual experience, ML concepts, or even do a coding exercise where AI is allowed, like how people actually work on the job?

Why do you think companies are still stuck using these older methods to evaluate candidates?