The original post makes a good point that, paraphrased, it's about improving a flawed information architecture instead of trying to fix symptoms, like the hamburger menu. And though it's tempting to use esoteric, hard-to-use/understand gestures and interaction models, it's almost never the right way. It's that subtle difference between an interface that feels like it's been designed to 'feel smart', instead of something that's easy to use.