[codingdave]

> It's no longer good enough to see if someone's good a programing, instead we need to screen if someone is good at engineering.

That has been true for many years. That is why we don't just ask FizzBuzz and hire people who can do it. Your ideas of the additional depth that is needed are 100% correct... but they aren't new since LLMs came out. They express the same depth that we've been interviewing for all along.

[Python3267]

I guess what I'm stabbing at is that the FAANG interviews I've done and my friends who work there operate with that mindset. You do need to ask questions to answer the problem in those interviews but they heavily rely on interviewee's code.