The Java interop compromise is probably the biggest weakness of the proposal - it works beautifully within Kotlin but degrades at boundaries. This is similar to how Kotlin's nullable types (String?) become platform types in Java.
I think it's good nonetheless to add stuff to Kotlin that won't translate 1:1 to Java, both because Java is evolving but also because Kotlin is used in "Native" (non-JVM) contexts as well (not extensively, but hopefully that'll change).