Can you explain what you actually want?
“The way you build an agent is you ask an agent to build it for you.”
When Nvidia CEO Jensen Huang said this during a recent keynote, it struck me as one of those statements whose implications are far greater than the words themselves. If you can ask AI to build the very tools you need, the ability to explain what you want becomes extraordinarily powerful. More people can take something they have imagined and begin turning it into something real.
I think this deserves much more attention. So much of the conversation about AI revolves around what the technology can do, how quickly it is improving, and which jobs it might change. I find myself thinking about the person using it. What do they want to accomplish? Can they explain why it matters? Can they ask questions that help them see beyond their first idea? These are familiar human abilities, and their value grows as we gain access to technology that can act on them.
Most of us have abandoned an idea because we couldn’t see a practical way to pursue it. Perhaps you understood a problem well enough to imagine a solution, but you didn’t know how to build it or who could help. Finding out would have required time, money, or expertise you didn’t have. Eventually, you put the idea aside. I wonder how many useful things have never existed simply because the person who saw the possibility had no affordable way to explore it.
AI is changing that calculation. You can explain an idea, ask how it might work, and begin developing a first version through conversation. There are still obstacles, and some problems require considerable expertise and effort. But you can explore further before those obstacles force a decision. You have a way to discover whether an idea has potential while learning what it would take to make it work.
That brings us to a surprisingly difficult question: can you explain what you actually want?
Imagine sitting down with an architect to describe your dream home. You might begin by saying you want somewhere beautiful, comfortable, and spacious. Those words sound meaningful, but they leave almost everything open to interpretation. The architect could design a wonderful house that feels completely wrong for you because you haven’t mentioned that you spend Sunday afternoons cooking with friends, that your father struggles with stairs, or that morning sunlight in the kitchen matters more to you than an extra bedroom. The details that make the home yours are still in your head.
We do this with AI all the time. We have a feeling about the result we want, give a brief description, and hope the system will fill in the gaps as we would. When the answer feels generic, part of the problem may be that we haven’t communicated the things that would make it specific. We have left out the purpose, the circumstances, or the tradeoffs we care about. Learning to recognize those omissions—and express what is missing—is becoming a valuable skill.
What makes this especially interesting is that we often discover what we want while trying to explain it. You might begin by asking for help organizing your week and realize, through a few questions, that your real problem is agreeing to work you don’t have time to do. You could spend hours building a more elaborate calendar without addressing that. A good conversation helps you examine the problem before committing yourself to a solution.
This is why I think asking better questions matters so much. You can ask AI to explore your assumptions, explain alternatives, or question you about something you haven’t thought through. You can tell it that you’re struggling to express an idea and ask it to help you find the words. There is no requirement to arrive with a perfect instruction. Being willing to stay in the conversation, think about the answers, and clarify your intentions can take you much further than trying to get everything right in the first request.
There is also something powerful about describing the future in concrete terms. “I want to spend less time on administration” is an understandable ambition, but imagine explaining that you want to finish your working day without another hour of sorting notes, updating records, and remembering who needs a response. Now you have something you can examine. You can describe where information comes from, what decisions require your attention, and what you would trust a system to handle. As you explain that future, you begin to understand how to create it.
For me, this is the connection between communication and imagination. Putting an idea into words forces you to give it substance. You have to consider what would actually change, whose life would improve, and how you would know whether you had succeeded. Questions expose the parts you’ve left vague. A first attempt gives you something to react to. Gradually, the future you pictured becomes clear enough to work toward.
Your experience becomes especially valuable once there is something to evaluate. A teacher can recognize when an explanation will confuse a child, even if every sentence is factually correct. A restaurant owner can see why a proposed booking process would cause trouble during a busy service. Someone caring for an aging parent can identify difficulties that another person might never think to ask about. That knowledge helps you judge what AI produces and explain how it needs to improve. The better you understand a situation, the more useful you can make the conversation.
This is why I’m encouraging people to spend time practicing communication. Explain an idea you’ve been carrying around. Ask questions about something you don’t understand. When an answer disappoints you, work out why and put that into words. The effort helps you become clearer about your own thinking, and it improves your ability to work with other people as well. Everyone benefits when they understand what you are trying to achieve and why.
I believe this ability can give people the keys to the kingdom. There is enormous human potential tied up in ideas that have never been expressed clearly enough to pursue. As AI makes more of the work of building accessible, more of us will have the opportunity to act on what we know, what frustrates us, and what we can imagine.
Think about the idea you keep returning to, or the problem you understand so well that you’ve stopped noticing how much it bothers you. Try explaining what you would like to change and what life would look like if you succeeded. You may discover that the conversation takes you somewhere you’ve wanted to go for years.
