How to Use AI Without Lying to Yourself
Author's Note
A note before the book starts.
I am an engineering manager. I have been pushing AI into my team's workflow since before it was expected and after it became required. I am not a skeptic watching from a safe distance. I am inside the experiment, running it, defending it in planning meetings, and dealing with what it sends back.
What it sends back is not what the slides say it sends back.
The productivity numbers are real. So is everything else. The tiredness that doesn't match the hours. The speed that doesn't match the output. The way people are working differently without being able to say how. None of this shows up on any dashboard I have seen. I started writing it down because I could not find anyone else who had. Then I started writing this.
I wrote it with AI assistance. Grammar, wording, the parts of editing that feel like homework. It told me when I repeated myself, which happened more than I would like to report. I reviewed everything. I kept what sounded like me and removed what did not, which is the same thing any writer does with any editor, except this one was available at 2 AM and did not have opinions about the title.
Not mentioning this in a book about AI honesty would have been a very confident opening move. I decided against it.
On to the book.