preface

 

Using a large language model for the first time is an almost magical experience. I still remember my first chat with GPT-3 (nowadays an outdated model). For the first time, it seemed to me that my computer actually understood me and could react appropriately to a wide range of complex inputs. What’s more, I gave it various tasks, ranging from text analysis to coding, and the model was able to solve them based on my instructions alone! I was used to a world in which neural networks had to be trained for highly specialized tasks using large amounts of task-specific training data that had to be labeled tediously by hand, so this was an absolute game-changer that opened a world of new and exciting possibilities.