AI chatbots are banned from our docs… for now

The text discusses the author's experience with two different AI applications - GitHub Copilot and an AI chatbot for their docs. While they found GitHub Copilot to be extremely helpful, they had to abandon the use of the AI chatbot due to its inaccuracies. They suggest that the key difference between the two is supervision; GitHub Copilot's output is supervised by a trained human, while the docs bot was unsupervised. The author believes that supervised services are likely to become more prevalent and change how we work, allowing us to focus on higher-level tasks. However, it remains uncertain whether unsupervised services will eventually close the gap in accuracy.