For centuries, the hard part was getting an answer. You had to find a scholar, a book, a library. The answer was the scarce resource, and the question cost nothing. That economy has flipped: the answer has become abundant and nearly free. What's scarce now is knowing what to ask.
We published an article on how to phrase a request, with concrete techniques. Here's the deeper side: why the question has become the hard part.
What sets a good question apart
A good question isn't a well-turned one. It's a question that opens the right space.
Three qualities keep coming up.
It names the real problem. Many questions target a symptom. "How do I convince my team to adopt this tool" presupposes that adoption is the problem. The real question might be "does this tool solve a problem my team actually has." A poorly framed question yields excellent answers that miss the point.
It doesn't contain its own answer. Asking why one option is better gets you arguments, not an evaluation. That's the mechanism of sycophancy: the system follows the direction you've indicated.
It exposes its assumptions. Every question rests on things taken for granted. Making them explicit lets you discover that one of them is false, which is often the most useful finding of all.
Why it's hard
Here's the paradox: asking a good question already requires understanding the domain.
To ask "what's the difference between these two approaches," you need to know there are two approaches. To ask "does this reasoning account for such-and-such effect," you need to know that effect exists. The precise question is the product of skill, not a substitute for it.
That's why the idea that these tools would make expertise unnecessary doesn't hold. They make expertise more profitable, by removing the execution work that went with it, but they don't replace it: they shift it upstream.
This ties into what we wrote about learning to code: someone who knows nothing doesn't know what to ask, and can't evaluate what they receive either.
A method that works remarkably well. First, ask what you'd need to know to answer your question properly. You often discover dimensions you hadn't considered. Next, reframe your question in light of those elements. Finally, ask what could invalidate the answer you've got. You get a more accurate answer, and above all you learn the structure of the problem, which will serve you next time without a tool.
What this changes at work
This shift explains a visible transformation in many professions, one we described regarding jobs facing automation.
The execution share shrinks, the framing share grows. A consultant spends less time producing slides and more time defining what needs to be analysed. A developer writes less code and decides more about what should be built.
That's good news for those who enjoy thinking, and a real difficulty for those whose work rested on executional excellence. This transition isn't socially neutral, and it doesn't happen without casualties.
What to take away
There's something rather reassuring in this shift. The ability to ask a good question is a deeply human skill, an ancient one, cultivated through curiosity, reading, and experience of the real world. It doesn't automate, because it requires knowing what matters, and knowing what matters requires having something at stake.
Socrates didn't give answers. He asked questions, and that's what has survived twenty-five centuries. In an era where answers are worth nothing, he may never have been more relevant.