Voice interfaces are not that easy (for people) to follow. VUIs can require put more cognitive load on people than well designed GUIs.
An assumption that a voice interface is more straightforward and requires less cognitive load from people comes to mind quickly.
It is a false premise that there are fewer things they need to process if people do not see them.
Spoken language, although more straightforward, requires people to remember a thing, recall information, use short term memory, and lack any visual support. When not used for simple things, voice interfaces can put a much bigger cognitive load on people.
GUIs are far from being dead.