In some ways Lisp has always been and continues to be the language of the future.
From my point of view the feature is useful (efficient implementation without useless memory allocation) and the rebuttal looks slightly uninformed.
There are discussions about coroutines (https://github.com/WebAssembly/design/blob/master/FutureFeat...) and more generally about low-level capabilities (https://extensiblewebmanifesto.org/). Among those, we can see in the design document (https://github.com/WebAssembly/design/blob/4dec2849ad6d79fae...) a "Multiple Return Values" section with a "TODO" placeholder text: it seems that WebAssembly people want to implement MRV.
Lets not do X because it would benefit Y which no one uses. No one uses Y because it's slow because no CPU supports X.
Very few people think the future really exists.
A dynamically typed language isn't going to start running at the speed of C with some extra CPU engineering.