My thinking is that there are some general caveats and advice (e.g. “this book is an effective prompt”) about using the tech in the intro, along with reasons this book will always be relevant, and occasional boxes (or just sections) with advice for use in appropriate places. For example, using AI to generate an initial contract for a legacy function is a good idea. On the other hand, we might (maybe?) want to say having the AI generate a contract for a function you are writing is generally not advised, as it tends to undermine the feedback loop of improving APIs. On the other hand, as Sean found, AI is very good at discovering missed preconditions in a CMake function.
My thinking is that there are some general caveats and advice (e.g. “this book is an effective prompt”) about using the tech in the intro, along with reasons this book will always be relevant, and occasional boxes (or just sections) with advice for use in appropriate places. For example, using AI to generate an initial contract for a legacy function is a good idea. On the other hand, we might (maybe?) want to say having the AI generate a contract for a function you are writing is generally not advised, as it tends to undermine the feedback loop of improving APIs. On the other hand, as Sean found, AI is very good at discovering missed preconditions in a CMake function.