Why is there still no proper T9 on desktop and what do LLMs have to do with it?

The author analyzes a paradox in modern user experience: why mobile devices are equipped with advanced autocorrect and predictive text systems, while desktop operating systems are limited to basic spell-checking with red underlines. The article examines the historical evolution of input tools and the reasons why desktop solutions have remained static for so long. The focus is on the potential of modern Large Language Models (LLMs). The author discusses whether neural network technologies could be the missing link that enables intelligent predictive input on computers, adapting to the user's context and style. The article raises questions about the technical implementation, performance, and usability of integrating such solutions into daily PC work, offering a look at the future of text input through the lens of generative AI.