This paper describes an emerging project that aims at better understanding how speakers of different languages, with diverse needs, produce texts today, and rethinking how they should do so in the future. Text entry (TE), the process of inputting text into a computer, is a crucial part of our interaction with technology. However, research has shown that the conventional TE method--typing on a keyboard--is not the most efficient or appropriate way of inputting text in various contexts, for diverse users. Now, how can we design more effective, accessible and inclusive TE methods that consider the uniqueness of different languages and diverse users and use cases? This fascinating question is likely to motivate speech technology, writing-process and human-computer interaction researchers alike, particularly in the generative AI age. But we are also living in the age of speech and multimodal interactions, which offer unprecedented opportunities to reinvent how we learn, write and communicate.