• Computer scientist Geoff Hinton's statement that chatbots' predictive text is 'glorified autocomplete' prompts reflections on intellectual processes at work. The author compares their own work experience, where they often provide associative responses similar to a chatbot's autocomplete, with deeper logical thought processes.
  • Differentiating between associative reasoning, which is likened to 'autocomplete', and genuine logical reasoning, the author disagrees with Hinton's assumption that by training a chatbot to predict the next word, it is forced to understand. They argue that the two types of reasoning are distinct both in how they feel and what they do.