One of my favorite things I've learned from the designers Charles and Ray Eames is a simple principle: Never delegate understanding.

There's an irony in how relevant those words feel today. We're increasingly delegating our understanding to technology built to recognize and generate patterns from what already exists. AI can produce sophisticated, sometimes genuinely novel work. But for humans, arriving at an answer is only part of what matters. There's also what we learn, experience, and become along the way.

AI can generate something new without having experienced anything new. Humans are changed by what they encounter.

It reminds me of one of my favorite lines from the late poet Andrea Gibson: "To become who we are meant to become, the becoming must be ours."

I love those words alongside the Eames principle. Becoming is hard. Not knowing is hard. Being wrong, struggling to understand, having our assumptions challenged, and discovering what we're capable of are uncomfortable experiences.

They're also the point.

I've started noticing what happens when we skip some of that work. A candidate presents an impressive exercise but struggles to explain the thinking behind it. A more junior HR professional I'm mentoring shares pages of research and recommendations, but I can't find where they've wrestled with the options themselves. What do they think? What would they recommend, and why?

I've also seen the other side of this. Some of the best aha moments I've witnessed on teams have come from productive struggle. The slow, sometimes frustrating process of figuring something out together, and reaching an understanding none of us had when we started.

I worry we're creating a world in which having an answer is becoming more important than developing an understanding.

I think about my own kids, and what they might miss if they grow up without having to wrestle with difficult questions or figure things out for themselves. I wouldn't want them, or any of us, to miss the frustration of not knowing, or the satisfaction and confidence that come from finally figuring something out.

Organizations have a responsibility here, too. Beyond establishing rules for AI use, we need to protect what makes a culture of learning: curiosity, experimentation, mentorship, the freedom to make mistakes, and the expectation that people develop their own judgment. Managers can help by making thinking visible: asking why someone made a recommendation, inviting disagreement, sharing their own mistakes, and giving people opportunities to make decisions and learn from the consequences.

I don't think we need to preserve difficulty for its own sake. But not everything that's difficult is something we should try to eliminate.

If we stop doing the work of understanding, who do we become?