14. Who Can Tell If It's Any Good

Up to now this book has mostly been about you. Your mind, your feelings, your decisions, the people you think with. But most of us don't use AI alone. We use it inside teams and companies and professions, and those places are making choices right now about what AI is for. A lot of those choices don't look like choices. They look like a pilot program, or a hiring freeze, or a new tool rolled out without any notice or instructions.

So this chapter is about what happens to expertise when the machine can produce almost anything. About the kind of uncertainty no amount of intelligence resolves. And about leadership, because whoever leads decides—whether they mean to or not—what kind of people an organization grows.

I'll say up front where I'm coming from because I worry we’re on the wrong track—at work, and in what it means to age and to share your own contribution to this project we call humanity. I think growing a human should cost more than it used to. More bespoke, more personal, harder, more real. AI makes the cheap version of growing people very tempting and I want to use this chapter to explain my argument against taking it.

I've sat through a lot of wild AI demos lately. Some in board meetings, others in management offsights, plus some startups that should probably stay nameless. They go the same way. Someone gives the machine a messy problem, and a few minutes later there's an answer that is honestly remarkable—a competitive analysis, or sometimes a working prototype. People are impressed because it is genuinely impressive. Someone says, "That would have taken us two weeks."

But what I've started to notice is that it is rare for someone to ask the follow up question they should be asking. Namely: “Who here would know if that's wrong?” And I mean wrong in the way that reads well and gets approved, and is still profoundly, embarrassingly WRONG.

I'll start with a confession, because, as a small operation, Dave and I live with this everyday.