I've been involved in hiring panels across several companies with very different interview strategies. Before going into this I leaned heavily toward your style: Easy questions that act as a simple filter for people who can't program.

Looking back, I have to admit the simple question approach wasn't as good as I thought it would be. The companies who gave more difficult questions and pushed candidates toward harder and harder interview problems with the expectation that they would hit a wall at some point really did sort candidates better. I know there's a widespread belief that difficult interview questions are prone to too many false negatives due to anxiety in interviews and people who freeze up. However, when we tried interview do-overs or alternate take-home problems as a fallback for these candidates we didn't really see any improvement. I can think of a few candidates who we talked ourselves into giving a pass because we assumed interview anxiety was obscuring some real skill, but we got it wrong in both cases.

I still don't know the correct answer, but these days I'm leaning more toward the interview formats where the questions get hard enough that the candidate is really challenged as opposed to the simple ones like FizzBuzz where it's just a quick filter. Note that I've worked in domains where algorithmic problem solving is something we actually do, so finding interview problems that match problems we've actually solved in production is possible, not just rote LeetCode stuff.