Language Is the New UI (bytesauna.com)

A recent discourse suggests that Large Language Models (LLMs) should be viewed as a new layer of abstraction in human-computer interaction, termed Language User Interfaces (LUI). This perspective contrasts with the common narrative that frames LLMs as a revolutionary, standalone technology. By comparing LUIs to historical interface developments—from command line interfaces (CLI) to graphical user interfaces (GUI)—the argument posits that LLMs can enhance usability by allowing users to communicate in natural language, similar to conversing with a colleague. Significantly, this shift towards LUIs may democratize access to complex functionalities, making them more intuitive for a broader user base, yet it introduces challenges related to precision and efficiency. Unlike CLIs, which cater to users needing exactitude, the inherent “fuzziness” of natural language can lead to ambiguities that may hinder tasks requiring strict specifications. While LUIs have the potential to streamline interactions and could play a crucial role in software development, the expectation is that they won't fully replace existing interface forms. The discussion emphasizes the importance of reevaluating these technologies without preconceived notions, inviting fresh perspectives on their role in the evolving landscape of AI and machine learning.