I can realistically do most mediums quite quickly, missing an edge case here and there without a test-fix loop like LC provides. On LC and in Python I find most of the medium problems trivial. This is not my problem.

My problem is I find these interviews dehumanizing, and they reflect on how the work is afterwards. I like my work, and I like working with people who like their work and are there because they're good at doing what the work is, not because they could do what most second year undergrads could easily do. I've been historically targeting positions related to certain field, and I'd want to be hired because of what knowledge and experience I've built up.

Still, my feeling is worsened by how long these interview processes take (multiple sessions of the same thing basically), and how you can fail arbitrarily because of slight mistakes. Which means, even if you know you'll pass 80% (quite good) of your LC medium focused interviews, if you take 5 of them in a row, you have a 32% of not failing one. It's a numbers game, obviously, but that doesn't mean the human playing it doesn't experience feelings of stress and frustration.