I think the best interview question, and really the only one you need to determine technical ability is ask someone to describe a http request in as much detail as possible.
To write code (even with the benefit of AI) effectively you need a mental model of the systems you work with, reading the chatGPT response doesn't prove you have that.
Yes, technical interview questions should be relevant to the job field. What's your point?
The hard part is selecting good questions that act as reliable predictors of actual job performance. Very few hiring managers can do that reliably, although many fool themselves into believing that they can.
The point is that someone gave a specific example of the much more general concept of probing for mental model by way of detailed explanation of a process he ought to be familiar with. You objected to the specific details - knowledge of HTTP. That's not an indictment of the general approach.
That said ML models have gotten to the point where I'd have to disagree with OP that this approach will necessarily filter their use. However there are plenty of available mitigations, from latency of response to requiring a video feed that fully covers the candidate, his screen, and his keyboard.
The "what happens when I enter facebook.com in the browser and hit enter" is/was a well-known FAANG question a few years back so I would expect that all the LLMs are well versed in it, as will be NK infiltrators
To write code (even with the benefit of AI) effectively you need a mental model of the systems you work with, reading the chatGPT response doesn't prove you have that.