Today, technology is changing so quickly that it can be hard to follow how tools evolve and understand how they actually work. As a result, we often use complex technologies such as AI without really knowing what goes on behind the scenes.
The Large Language Writer project responds to this challenge by making the process of AI text generation more open and understandable. Instead of offering a seamless and seemingly magical experience, it invites users to engage in a transparent and hands-on interaction. The system is built around three core elements: Dataset, Emphasis, and Probability. These help break down how language models like GPT (generative pre-trained transformers) create text. A custom-built machine with a voxel-based body, keyboard, e-ink screen, and camera allows users to physically explore and test this process. Through this approach, the project aims to shift our view of AI from passive use to active understanding and help uncover what is really behind the words we read.
