I know I'm going to get a lot of grief over this, but I'm coming to the conclusion Tuples have a place in Kotlin, primarily for quickly improvising key objects in functional operations. https://github.com/thomasnield/kotlin-statistics/issues/23

I missed the conversation but I have also been tempted by tuples. In my scenario it's working with Apache Spark and doing a bunch of small transformations on my way to the actual model object.